I bought Olimex MSP430F1611 prototype board and HMC5843 breakout board from Sparkfun a few days ago. I am trying to UART0 I2C module of F1611 to read 6 bytes XYZ sensor raw data continuously from the HMC5843 chip through I2C interface.

Now I tied SDA and SCL of HMC5843 to P3.1 and P3.3 respectively with two 10K Ohms pull-up resistors. HMC5843 is powered by 3.3V single power supply from the F1611 prototype board and I added a 100uF Electrolytic capacitor in parellel to the 4.7uF Tantulum cap on the breakout baord.

I am new to I2C protocol. I tried some example C codes from TI. But the HMC5843 refused to enter RXRDYIE interrupt to output any data.

Can anyone give me a simple piece of C code so that MSP430F1611 can talk to HMC5843? Thanks.

------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Well lots have tried, and some have come close, but for a long time the
MSP has ruled the roost when it comes to low power, but finally I think
it may have met its match, and not because anyone else has done anything
spectacular, but because Ti has basically failed to do anything. WEll,
that's not entirely true. They've taken a micro whose greatest asset,
and major marketing point, was it's low power performance, and then
totally neglected any further low power development for the last 10
years or more, developing it along the lines of every other micro
family, going for big memory, big RAM, big pin counts, damn for those in
Aus they could almost have thrown in the big PRUNE while they were at
it. In fact if you examine the data sheets over the years you find a
gradual creep up in uA/MHz on some parts.

I've been hunting for a lower power part for years, a useful low power
part, one that doesn't need resetting to come back to life again, or at
least one that can run an RTC and precision timing while in LP modes.
The MSP was always expensive. The newer parts sorted that slightly.

Now it looks like Energy Micro is a real contender, true low power sleep
with RTC and some neat timing, plus dedicated low power peripherals like
an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
but so far the specs look very good. Only major downer is the lack of
any small parts, everything starts at 32K flash and 8kRAM, and just
grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
somewhat bigger than the smaller MSPs, and for me small is good.

For those who don't know, ENERGY MICROS founders include some of the ex
CHIPCON people responsible for the CC11xx etc.

Cheers

Al
------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

On Sun, 25 Oct 2009 06:35:12 +1030, Al wrote:

>Well lots have tried, and some have come close, but for a long time the
>MSP has ruled the roost when it comes to low power, but finally I think
>it may have met its match, and not because anyone else has done anything
>spectacular, but because Ti has basically failed to do anything. WEll,
>that's not entirely true. They've taken a micro whose greatest asset,
>and major marketing point, was it's low power performance, and then
>totally neglected any further low power development for the last 10
>years or more, developing it along the lines of every other micro
>family, going for big memory, big RAM, big pin counts, damn for those in
>Aus they could almost have thrown in the big PRUNE while they were at
>it. In fact if you examine the data sheets over the years you find a
>gradual creep up in uA/MHz on some parts.
>
>I've been hunting for a lower power part for years, a useful low power
>part, one that doesn't need resetting to come back to life again, or at
>least one that can run an RTC and precision timing while in LP modes.
>The MSP was always expensive. The newer parts sorted that slightly.
>
>Now it looks like Energy Micro is a real contender, true low power sleep
>with RTC and some neat timing, plus dedicated low power peripherals like
>an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
>but so far the specs look very good. Only major downer is the lack of
>any small parts, everything starts at 32K flash and 8kRAM, and just
>grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
>somewhat bigger than the smaller MSPs, and for me small is good.
>
>For those who don't know, ENERGY MICROS founders include some of the ex
>CHIPCON people responsible for the CC11xx etc.

Looking at them, now. I see earliest availability as 2/2010 and the
24-IO parts as 3/2010? The 12-bit 1MHz won't cut it for me, right
now. Current application requires >= 1MHz 16-bit SAR ADC and uses
short bursts of 3ms data gathering [for 12-bit, would need to increase
amplifier noise and be 32 such bursts, at least, but in that case I'd
rather keep lower noise in the amplifier and use the 16-bit ADC the
same number of times and get more precision in the output.])

I really don't like their video -- nice sizzle, but no meat -- which
is great if I want to be entertained for a moment but terrible if I
want to be informed. And they have no data sheets yet, either.

Operates down to 1.8V, they say (I assume that is where they get their
180uW/MHz.) What does it do when running on 3.3V, though? No way to
know, unless you can point to a data sheet on the part. But thanks
for the heads up. ARM based compilers are pretty mature and I'd like
to see something there with 2us startup times and extreme low power as
a goal. I'll keep an eye out, maybe get a demo board anyway when it
becomes available.

So I guess this explains your "haven't fully checked it out yet."
Doesn't look like a lot is available, yet.

Did the team for this company peel away from Atmel? Who is funding
it? What foundary are they using?

Jon
------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Hi Jon, DEv kits available from Digikey in mid-November, I have data
sheets for all parts. Also reference manuals are available too:-

http://www.energymicro.com/index.php?option=com_content&task=view&id=212&Itemid=232
http://www.energymicro.com/index.php?option=com_content&task=view&id=213&Itemid=231

You have to look under the downloads tab on their website. The info on
parts, dev kits etc came from being subscribed for a while and talking
to the people there. Cost of the dev kit is pretty high in my opinion at
US$299. The first two parts available are the smaller 32 pin parts.

Although this won't meet your ADC needs, obviously the MSP doesn't
either, and this is both marginally faster, and consumes slightly less
current, but that's on paper, it will be interesting to see what it can
really do.

Cheers

Al

Jon Kirwan wrote:
> On Sun, 25 Oct 2009 06:35:12 +1030, Al wrote:
>
>> Well lots have tried, and some have come close, but for a long time the
>> MSP has ruled the roost when it comes to low power, but finally I think
>> it may have met its match, and not because anyone else has done anything
>> spectacular, but because Ti has basically failed to do anything. WEll,
>> that's not entirely true. They've taken a micro whose greatest asset,
>> and major marketing point, was it's low power performance, and then
>> totally neglected any further low power development for the last 10
>> years or more, developing it along the lines of every other micro
>> family, going for big memory, big RAM, big pin counts, damn for those in
>> Aus they could almost have thrown in the big PRUNE while they were at
>> it. In fact if you examine the data sheets over the years you find a
>> gradual creep up in uA/MHz on some parts.
>>
>> I've been hunting for a lower power part for years, a useful low power
>> part, one that doesn't need resetting to come back to life again, or at
>> least one that can run an RTC and precision timing while in LP modes.
>> The MSP was always expensive. The newer parts sorted that slightly.
>>
>> Now it looks like Energy Micro is a real contender, true low power sleep
>> with RTC and some neat timing, plus dedicated low power peripherals like
>> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
>> but so far the specs look very good. Only major downer is the lack of
>> any small parts, everything starts at 32K flash and 8kRAM, and just
>> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
>> somewhat bigger than the smaller MSPs, and for me small is good.
>>
>> For those who don't know, ENERGY MICROS founders include some of the ex
>> CHIPCON people responsible for the CC11xx etc.
>
> Looking at them, now. I see earliest availability as 2/2010 and the
> 24-IO parts as 3/2010? The 12-bit 1MHz won't cut it for me, right
> now. Current application requires >= 1MHz 16-bit SAR ADC and uses
> short bursts of 3ms data gathering [for 12-bit, would need to increase
> amplifier noise and be 32 such bursts, at least, but in that case I'd
> rather keep lower noise in the amplifier and use the 16-bit ADC the
> same number of times and get more precision in the output.])
>
> I really don't like their video -- nice sizzle, but no meat -- which
> is great if I want to be entertained for a moment but terrible if I
> want to be informed. And they have no data sheets yet, either.
>
> Operates down to 1.8V, they say (I assume that is where they get their
> 180uW/MHz.) What does it do when running on 3.3V, though? No way to
> know, unless you can point to a data sheet on the part. But thanks
> for the heads up. ARM based compilers are pretty mature and I'd like
> to see something there with 2us startup times and extreme low power as
> a goal. I'll keep an eye out, maybe get a demo board anyway when it
> becomes available.
>
> So I guess this explains your "haven't fully checked it out yet."
> Doesn't look like a lot is available, yet.
>
> Did the team for this company peel away from Atmel? Who is funding
> it? What foundary are they using?
>
> Jon
> ------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

JON, BY THE WAY THE 180uA /MHz is quoted at 3.0V using the external high
frequency crystal oscillator, it is 220uA/MHz using the internal RC osc,
so no hiding behind low voltages

Al

Jon Kirwan wrote:
> On Sun, 25 Oct 2009 06:35:12 +1030, Al wrote:
>
>> Well lots have tried, and some have come close, but for a long time the
>> MSP has ruled the roost when it comes to low power, but finally I think
>> it may have met its match, and not because anyone else has done anything
>> spectacular, but because Ti has basically failed to do anything. WEll,
>> that's not entirely true. They've taken a micro whose greatest asset,
>> and major marketing point, was it's low power performance, and then
>> totally neglected any further low power development for the last 10
>> years or more, developing it along the lines of every other micro
>> family, going for big memory, big RAM, big pin counts, damn for those in
>> Aus they could almost have thrown in the big PRUNE while they were at
>> it. In fact if you examine the data sheets over the years you find a
>> gradual creep up in uA/MHz on some parts.
>>
>> I've been hunting for a lower power part for years, a useful low power
>> part, one that doesn't need resetting to come back to life again, or at
>> least one that can run an RTC and precision timing while in LP modes.
>> The MSP was always expensive. The newer parts sorted that slightly.
>>
>> Now it looks like Energy Micro is a real contender, true low power sleep
>> with RTC and some neat timing, plus dedicated low power peripherals like
>> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
>> but so far the specs look very good. Only major downer is the lack of
>> any small parts, everything starts at 32K flash and 8kRAM, and just
>> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
>> somewhat bigger than the smaller MSPs, and for me small is good.
>>
>> For those who don't know, ENERGY MICROS founders include some of the ex
>> CHIPCON people responsible for the CC11xx etc.
>
> Looking at them, now. I see earliest availability as 2/2010 and the
> 24-IO parts as 3/2010? The 12-bit 1MHz won't cut it for me, right
> now. Current application requires >= 1MHz 16-bit SAR ADC and uses
> short bursts of 3ms data gathering [for 12-bit, would need to increase
> amplifier noise and be 32 such bursts, at least, but in that case I'd
> rather keep lower noise in the amplifier and use the 16-bit ADC the
> same number of times and get more precision in the output.])
>
> I really don't like their video -- nice sizzle, but no meat -- which
> is great if I want to be entertained for a moment but terrible if I
> want to be informed. And they have no data sheets yet, either.
>
> Operates down to 1.8V, they say (I assume that is where they get their
> 180uW/MHz.) What does it do when running on 3.3V, though? No way to
> know, unless you can point to a data sheet on the part. But thanks
> for the heads up. ARM based compilers are pretty mature and I'd like
> to see something there with 2us startup times and extreme low power as
> a goal. I'll keep an eye out, maybe get a demo board anyway when it
> becomes available.
>
> So I guess this explains your "haven't fully checked it out yet."
> Doesn't look like a lot is available, yet.
>
> Did the team for this company peel away from Atmel? Who is funding
> it? What foundary are they using?
>
> Jon
> ------------------------------------

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

On Sun, 25 Oct 2009 12:24:08 +1030, you wrote:

>Hi Jon, DEv kits available from Digikey in mid-November, I have data
>sheets for all parts. Also reference manuals are available too:-
>
>http://www.energymicro.com/index.php?option=com_content&task=view&id=212&Itemid=232
>http://www.energymicro.com/index.php?option=com_content&task=view&id=213&Itemid=231
>
>You have to look under the downloads tab on their website. The info on
>parts, dev kits etc came from being subscribed for a while and talking
>to the people there. Cost of the dev kit is pretty high in my opinion at
>US$299. The first two parts available are the smaller 32 pin parts.

Yes, US$299 is too high. Unless it's a pretty serious kit. One can
buy that new Kindle Amazon just announced, at list, for that much
money and still have enough for a pretty decent dinner, left over.

>Although this won't meet your ADC needs, obviously the MSP doesn't
>either, and this is both marginally faster, and consumes slightly less
>current, but that's on paper, it will be interesting to see what it can
>really do.

I don't want to make too much of the converter, really. One can buy
it externally and there are lots of good parts, then. Trying to merge
the two is very hard and I'm still surprised at how well Cygnal/SiLabs
managed to get that thing shoe-horned together. The 16-bit ADC
includes an appropriately fast sample and hold and draws 1.5mA,
besides working well, so I'm suitably impressed. I use it for short
bursts and can turn it off. And I don't expect to find anything like
it elsewhere. It's a one of a kind unicorn, of sorts. The only
reason for playing with it, at all, is that the entire cpu and the
converter costs me the same or less than a decent converter alone
costs me... so why not? Turns out, after rewriting that ln(x)
function and getting it to produce a flawless 16 bits (4.12 format, in
my case) on wider data inputs at 18 microseconds, that I no longer
have any doubts and their 8051 is just fine. I had worried, going
into it, since this has to process tens of thousands of ADC values
each second and perform lsq fits with larger than 50 bit accumulators
to yield precise 24 bit results with low noise. There's time to keep
software truncation errors to nil, so I'm happy.

An ARM would be a fair bit faster than an MSP or the 8051 for this. I
could use that. But there's no particular reason to select such a new
part, either. I have good experience with quite a number of other
parts that would do as well and for which I have some long experience
with the companies and know what to expect from them in terms of
support, etc. The extreme low power is, frankly, less an issue....
though more power adds cost and size and is generally important.

Jon
>Cheers
>
>Al
>
>Jon Kirwan wrote:
>> On Sun, 25 Oct 2009 06:35:12 +1030, Al wrote:
>>
>>> Well lots have tried, and some have come close, but for a long time the
>>> MSP has ruled the roost when it comes to low power, but finally I think
>>> it may have met its match, and not because anyone else has done anything
>>> spectacular, but because Ti has basically failed to do anything. WEll,
>>> that's not entirely true. They've taken a micro whose greatest asset,
>>> and major marketing point, was it's low power performance, and then
>>> totally neglected any further low power development for the last 10
>>> years or more, developing it along the lines of every other micro
>>> family, going for big memory, big RAM, big pin counts, damn for those in
>>> Aus they could almost have thrown in the big PRUNE while they were at
>>> it. In fact if you examine the data sheets over the years you find a
>>> gradual creep up in uA/MHz on some parts.
>>>
>>> I've been hunting for a lower power part for years, a useful low power
>>> part, one that doesn't need resetting to come back to life again, or at
>>> least one that can run an RTC and precision timing while in LP modes.
>>> The MSP was always expensive. The newer parts sorted that slightly.
>>>
>>> Now it looks like Energy Micro is a real contender, true low power sleep
>>> with RTC and some neat timing, plus dedicated low power peripherals like
>>> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
>>> but so far the specs look very good. Only major downer is the lack of
>>> any small parts, everything starts at 32K flash and 8kRAM, and just
>>> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
>>> somewhat bigger than the smaller MSPs, and for me small is good.
>>>
>>> For those who don't know, ENERGY MICROS founders include some of the ex
>>> CHIPCON people responsible for the CC11xx etc.
>>
>> Looking at them, now. I see earliest availability as 2/2010 and the
>> 24-IO parts as 3/2010? The 12-bit 1MHz won't cut it for me, right
>> now. Current application requires >= 1MHz 16-bit SAR ADC and uses
>> short bursts of 3ms data gathering [for 12-bit, would need to increase
>> amplifier noise and be 32 such bursts, at least, but in that case I'd
>> rather keep lower noise in the amplifier and use the 16-bit ADC the
>> same number of times and get more precision in the output.])
>>
>> I really don't like their video -- nice sizzle, but no meat -- which
>> is great if I want to be entertained for a moment but terrible if I
>> want to be informed. And they have no data sheets yet, either.
>>
>> Operates down to 1.8V, they say (I assume that is where they get their
>> 180uW/MHz.) What does it do when running on 3.3V, though? No way to
>> know, unless you can point to a data sheet on the part. But thanks
>> for the heads up. ARM based compilers are pretty mature and I'd like
>> to see something there with 2us startup times and extreme low power as
>> a goal. I'll keep an eye out, maybe get a demo board anyway when it
>> becomes available.
>>
>> So I guess this explains your "haven't fully checked it out yet."
>> Doesn't look like a lot is available, yet.
>>
>> Did the team for this company peel away from Atmel? Who is funding
>> it? What foundary are they using?
>>
>> Jon
------------------------------------

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Jon,

> On Sun, 25 Oct 2009 12:24:08 +1030, you wrote:
>=20
> >Hi Jon, DEv kits available from Digikey in mid-November, I have data
> >sheets for all parts. Also reference manuals are available too:-
> >http://www.energymicro.com/index.php?option=3Dcom_content&task=3Dview&id=
=3D212&I
> temid=3D232
>
>http://www.energymicro.com/index.php?option=3Dcom_content&task=3Dview&id=
=3D213&I
> temid=3D231
> >
> >You have to look under the downloads tab on their website. The info on
> >parts, dev kits etc came from being subscribed for a while and talking
> >to the people there. Cost of the dev kit is pretty high in my opinion at
> >US$299. The first two parts available are the smaller 32 pin parts.
>=20
> Yes, US$299 is too high. Unless it's a pretty serious kit. One can
> buy that new Kindle Amazon just announced, at list, for that much
> money and still have enough for a pretty decent dinner, left over.

It's a nice kit; $299 is a fair price for it. Comparing against Kindle, as
you know, is just plain crazy. :-)

The $299 keeps amateurs away and probably will not put off businesses that
are considering EFM32.

--
Paul Curtis, Rowley Associates Ltd=A0=A0 http://www.rowley.co.uk
CrossWorks V2 is out for LPC1700, LPC3100, LPC3200, SAM9, and more!

------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

On Mon, 26 Oct 2009 08:31:59 -0000, Paul wrote:

>Jon,
>
>> On Sun, 25 Oct 2009 12:24:08 +1030, you wrote:
>>
>> >Hi Jon, DEv kits available from Digikey in mid-November, I have data
>> >sheets for all parts. Also reference manuals are available too:-
>> >
>>
>>http://www.energymicro.com/index.php?option=com_content&task=view&id=212&I
>> temid=232
>>
>>http://www.energymicro.com/index.php?option=com_content&task=view&id=213&I
>> temid=231
>> >
>> >You have to look under the downloads tab on their website. The info on
>> >parts, dev kits etc came from being subscribed for a while and talking
>> >to the people there. Cost of the dev kit is pretty high in my opinion at
>> >US$299. The first two parts available are the smaller 32 pin parts.
>>
>> Yes, US$299 is too high. Unless it's a pretty serious kit. One can
>> buy that new Kindle Amazon just announced, at list, for that much
>> money and still have enough for a pretty decent dinner, left over.
>
>It's a nice kit; $299 is a fair price for it. Comparing against Kindle, as
>you know, is just plain crazy. :-)

There are hundreds of ways to compare and none of them necessarily
accurate for anyone in particular. I allowed myself some latitude.
Oh, well.

>The $299 keeps amateurs away and probably will not put off businesses that
>are considering EFM32.

That's one good argument, granted. There are others, of course.

Very nice to see hear from you again, Paul,
Jon
------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

I haven't seen it yet, so will reserve judgement until then. Personally
I don't like too many bells and whistles, however it would be hard
pressed to go past the Primer2 from ST at just $60 as a value for money
evaluation tool/dev kit, or the PIC32, at $50, or the explorer 16, which
is an excellent kit, at $130. In this era $300 is very expensive for an
evaluation kit, so I hope it's worth the effort.

Al

Paul Curtis wrote:
> Jon,
>
>> On Sun, 25 Oct 2009 12:24:08 +1030, you wrote:
>>
>>> Hi Jon, DEv kits available from Digikey in mid-November, I have data
>>> sheets for all parts. Also reference manuals are available too:-
>>>
>> http://www.energymicro.com/index.php?option=com_content&task=view&id=212&I
>> temid=232
>>
>> http://www.energymicro.com/index.php?option=com_content&task=view&id=213&I
>> temid=231
>>> You have to look under the downloads tab on their website. The info on
>>> parts, dev kits etc came from being subscribed for a while and talking
>>> to the people there. Cost of the dev kit is pretty high in my opinion at
>>> US$299. The first two parts available are the smaller 32 pin parts.
>> Yes, US$299 is too high. Unless it's a pretty serious kit. One can
>> buy that new Kindle Amazon just announced, at list, for that much
>> money and still have enough for a pretty decent dinner, left over.
>
> It's a nice kit; $299 is a fair price for it. Comparing against Kindle, as
> you know, is just plain crazy. :-)
>
> The $299 keeps amateurs away and probably will not put off businesses that
> are considering EFM32.
>
> --
> Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk
> CrossWorks V2 is out for LPC1700, LPC3100, LPC3200, SAM9, and more!
>
> ------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

The document of this kit is a pdf of more than 4 M-bytes.=20
It consists of almost two pages. Very efficient!

--- In m...@yahoogroups.com, "Paul Curtis" wrote:
>
> Jon,
>=20
> > On Sun, 25 Oct 2009 12:24:08 +1030, you wrote:
> >=20
> > >Hi Jon, DEv kits available from Digikey in mid-November, I have data
> > >sheets for all parts. Also reference manuals are available too:-
> > >
> >
> >http://www.energymicro.com/index.php?option=3Dcom_content&task=3Dview&id=
=3D212&I
> > temid=3D232
> >
> >http://www.energymicro.com/index.php?option=3Dcom_content&task=3Dview&id=
=3D213&I
> > temid=3D231
> > >
> > >You have to look under the downloads tab on their website. The info on
> > >parts, dev kits etc came from being subscribed for a while and talking
> > >to the people there. Cost of the dev kit is pretty high in my opinion =
at
> > >US$299. The first two parts available are the smaller 32 pin parts.
> >=20
> > Yes, US$299 is too high. Unless it's a pretty serious kit. One can
> > buy that new Kindle Amazon just announced, at list, for that much
> > money and still have enough for a pretty decent dinner, left over.
>=20
> It's a nice kit; $299 is a fair price for it. Comparing against Kindle, =
as
> you know, is just plain crazy. :-)
>=20
> The $299 keeps amateurs away and probably will not put off businesses tha=
t
> are considering EFM32.
>=20
> --
> Paul Curtis, Rowley Associates Ltd=A0=A0 http://www.rowley.co.uk
> CrossWorks V2 is out for LPC1700, LPC3100, LPC3200, SAM9, and more!
>

------------------------------------

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

On Mon, 26 Oct 2009 08:31:59 -0000, "Paul Curtis"
wrote:
>> >You have to look under the downloads tab on their website. The info on
>> >parts, dev kits etc came from being subscribed for a while and talking
>> >to the people there. Cost of the dev kit is pretty high in my opinion
at
>> >US$299. The first two parts available are the smaller 32 pin parts.
>>
>> Yes, US$299 is too high. Unless it's a pretty serious kit. One can
>> buy that new Kindle Amazon just announced, at list, for that much
>> money and still have enough for a pretty decent dinner, left over.
>
> It's a nice kit; $299 is a fair price for it. Comparing against Kindle,
as
> you know, is just plain crazy. :-)
>
> The $299 keeps amateurs away and probably will not put off businesses
that
> are considering EFM32.

No matter how you slice it $300 for a micro dev kit is on the high end. I
think other companies have it right, the dev kit should be priced at or,
even better, slightly below cost. See xilinx Spartan-3A DSP dev kit (which
sells for $299, the FPGA goes for $40-50, and the dev board is a 12 layer
board). The idea of a dev kit is to get your products into the hands of
design engineers to play with and win design-ins. Make money off the
product not the dev kit.

az

------------------------------------

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Most manufacturers put out wasteful glossies like this, however of
better value is their EFm32g reference manual at 6.74Mb and 445 pages.
here in Australia we get ripped off on Book prices, a 200 page book in
large type to stretch it out can cost up to $35, so a free down load
with 445 pages of hopefully interesting reading is always welcome.

Al

old_cow_yellow wrote:
> The document of this kit is a pdf of more than 4 M-bytes.
> It consists of almost two pages. Very efficient!
>
> --- In m...@yahoogroups.com, "Paul Curtis" wrote:
>> Jon,
>>
>>> On Sun, 25 Oct 2009 12:24:08 +1030, you wrote:
>>>
>>>> Hi Jon, DEv kits available from Digikey in mid-November, I have data
>>>> sheets for all parts. Also reference manuals are available too:-
>>>>
>>> http://www.energymicro.com/index.php?option=com_content&task=view&id=212&I
>>> temid=232
>>>
>>> http://www.energymicro.com/index.php?option=com_content&task=view&id=213&I
>>> temid=231
>>>> You have to look under the downloads tab on their website. The info on
>>>> parts, dev kits etc came from being subscribed for a while and talking
>>>> to the people there. Cost of the dev kit is pretty high in my opinion at
>>>> US$299. The first two parts available are the smaller 32 pin parts.
>>> Yes, US$299 is too high. Unless it's a pretty serious kit. One can
>>> buy that new Kindle Amazon just announced, at list, for that much
>>> money and still have enough for a pretty decent dinner, left over.
>> It's a nice kit; $299 is a fair price for it. Comparing against Kindle, as
>> you know, is just plain crazy. :-)
>>
>> The $299 keeps amateurs away and probably will not put off businesses that
>> are considering EFM32.
>>
>> --
>> Paul Curtis, Rowley Associates Ltd http://www.rowley.co.uk
>> CrossWorks V2 is out for LPC1700, LPC3100, LPC3200, SAM9, and more!
>>
> ------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Does Energy Micro have any released products - and if so, has anyone used
them?

I would not expect TI to sit on their hands idly and watch this market
vaporize.
CC2531 is a decent part - albeit I wish it had the '430 core instead of the
'8051.

Regards,
John W.

On Sat, Oct 24, 2009 at 1:05 PM, OneStone wrote:

> Well lots have tried, and some have come close, but for a long time the
> MSP has ruled the roost when it comes to low power, but finally I think
> it may have met its match, and not because anyone else has done anything
> spectacular, but because Ti has basically failed to do anything. WEll,
> that's not entirely true. They've taken a micro whose greatest asset,
> and major marketing point, was it's low power performance, and then
> totally neglected any further low power development for the last 10
> years or more, developing it along the lines of every other micro
> family, going for big memory, big RAM, big pin counts, damn for those in
> Aus they could almost have thrown in the big PRUNE while they were at
> it. In fact if you examine the data sheets over the years you find a
> gradual creep up in uA/MHz on some parts.
>
> I've been hunting for a lower power part for years, a useful low power
> part, one that doesn't need resetting to come back to life again, or at
> least one that can run an RTC and precision timing while in LP modes.
> The MSP was always expensive. The newer parts sorted that slightly.
>
> Now it looks like Energy Micro is a real contender, true low power sleep
> with RTC and some neat timing, plus dedicated low power peripherals like
> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
> but so far the specs look very good. Only major downer is the lack of
> any small parts, everything starts at 32K flash and 8kRAM, and just
> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
> somewhat bigger than the smaller MSPs, and for me small is good.
>
> For those who don't know, ENERGY MICROS founders include some of the ex
> CHIPCON people responsible for the CC11xx etc.
>
> Cheers
>
> Al
>
>
[Non-text portions of this message have been removed]

------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

EM have started releasing to their financial partners, and say they will
sampled to the rest of us in December with product being available in
early 2010.

The point of my original email is that Ti have sat on their hands, as
far as low power development is concerned, since the release of the
MSP430F series. Instead they have concentrated on the typical
'development path' taken by other microcontroller suppliers:- more
memory, more pins, more peripherals, more speed. In the process they
seem to have forgotten what originally made their market niche for them.

Over the last 10 years or so lots of other manufacturers have been
creeping up on Ti's low power claims, some of them slowly taking bites
out of it, while Ti seem to have abandoned any efforts at further
developments along the low power path and thrown their future into the
wireless basket. Ironically this is based on a product originally
developed by the people who are possibly going to prove their most
serious competition in the low power market.

I hope I'm wrong and that Ti have secretly been working on some new kick
ass 50uA/MHz technology with even lower LP modes, but somehow I think not.

Cheers

Al

John Westmoreland wrote:
> Does Energy Micro have any released products - and if so, has anyone used
> them?
>
> I would not expect TI to sit on their hands idly and watch this market
> vaporize.
> CC2531 is a decent part - albeit I wish it had the '430 core instead of the
> '8051.
>
> Regards,
> John W.
>
> On Sat, Oct 24, 2009 at 1:05 PM, OneStone wrote:
>
>>
>> Well lots have tried, and some have come close, but for a long time the
>> MSP has ruled the roost when it comes to low power, but finally I think
>> it may have met its match, and not because anyone else has done anything
>> spectacular, but because Ti has basically failed to do anything. WEll,
>> that's not entirely true. They've taken a micro whose greatest asset,
>> and major marketing point, was it's low power performance, and then
>> totally neglected any further low power development for the last 10
>> years or more, developing it along the lines of every other micro
>> family, going for big memory, big RAM, big pin counts, damn for those in
>> Aus they could almost have thrown in the big PRUNE while they were at
>> it. In fact if you examine the data sheets over the years you find a
>> gradual creep up in uA/MHz on some parts.
>>
>> I've been hunting for a lower power part for years, a useful low power
>> part, one that doesn't need resetting to come back to life again, or at
>> least one that can run an RTC and precision timing while in LP modes.
>> The MSP was always expensive. The newer parts sorted that slightly.
>>
>> Now it looks like Energy Micro is a real contender, true low power sleep
>> with RTC and some neat timing, plus dedicated low power peripherals like
>> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
>> but so far the specs look very good. Only major downer is the lack of
>> any small parts, everything starts at 32K flash and 8kRAM, and just
>> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
>> somewhat bigger than the smaller MSPs, and for me small is good.
>>
>> For those who don't know, ENERGY MICROS founders include some of the ex
>> CHIPCON people responsible for the CC11xx etc.
>>
>> Cheers
>>
>> Al
>>
>>
> [Non-text portions of this message have been removed]
>
> ------------------------------------

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Al,

We all know that TI 'monitors' (reads) the posts in this group - and I am
sure they have taken some of the issues discussed to heart - and I hope also
they have been working on an ultra-low power design. Would seem that is an
obvious area of research.

I still think the '1611 was (is) a great part - especially when it was made.

Maybe a TIer will chime in.

Regards,
John W.

On Mon, Oct 26, 2009 at 8:06 PM, OneStone wrote:

> EM have started releasing to their financial partners, and say they will
> sampled to the rest of us in December with product being available in
> early 2010.
>
> The point of my original email is that Ti have sat on their hands, as
> far as low power development is concerned, since the release of the
> MSP430F series. Instead they have concentrated on the typical
> 'development path' taken by other microcontroller suppliers:- more
> memory, more pins, more peripherals, more speed. In the process they
> seem to have forgotten what originally made their market niche for them.
>
> Over the last 10 years or so lots of other manufacturers have been
> creeping up on Ti's low power claims, some of them slowly taking bites
> out of it, while Ti seem to have abandoned any efforts at further
> developments along the low power path and thrown their future into the
> wireless basket. Ironically this is based on a product originally
> developed by the people who are possibly going to prove their most
> serious competition in the low power market.
>
> I hope I'm wrong and that Ti have secretly been working on some new kick
> ass 50uA/MHz technology with even lower LP modes, but somehow I think not.
>
> Cheers
>
> Al
> John Westmoreland wrote:
> > Does Energy Micro have any released products - and if so, has anyone used
> > them?
> >
> > I would not expect TI to sit on their hands idly and watch this market
> > vaporize.
> > CC2531 is a decent part - albeit I wish it had the '430 core instead of
> the
> > '8051.
> >
> > Regards,
> > John W.
> >
> > On Sat, Oct 24, 2009 at 1:05 PM, OneStone >
> wrote:
> >
> >>
> >> Well lots have tried, and some have come close, but for a long time the
> >> MSP has ruled the roost when it comes to low power, but finally I think
> >> it may have met its match, and not because anyone else has done anything
> >> spectacular, but because Ti has basically failed to do anything. WEll,
> >> that's not entirely true. They've taken a micro whose greatest asset,
> >> and major marketing point, was it's low power performance, and then
> >> totally neglected any further low power development for the last 10
> >> years or more, developing it along the lines of every other micro
> >> family, going for big memory, big RAM, big pin counts, damn for those in
> >> Aus they could almost have thrown in the big PRUNE while they were at
> >> it. In fact if you examine the data sheets over the years you find a
> >> gradual creep up in uA/MHz on some parts.
> >>
> >> I've been hunting for a lower power part for years, a useful low power
> >> part, one that doesn't need resetting to come back to life again, or at
> >> least one that can run an RTC and precision timing while in LP modes.
> >> The MSP was always expensive. The newer parts sorted that slightly.
> >>
> >> Now it looks like Energy Micro is a real contender, true low power sleep
> >> with RTC and some neat timing, plus dedicated low power peripherals like
> >> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
> >> but so far the specs look very good. Only major downer is the lack of
> >> any small parts, everything starts at 32K flash and 8kRAM, and just
> >> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
> >> somewhat bigger than the smaller MSPs, and for me small is good.
> >>
> >> For those who don't know, ENERGY MICROS founders include some of the ex
> >> CHIPCON people responsible for the CC11xx etc.
> >>
> >> Cheers
> >>
> >> Al
> >>
> >>
> >
> >
> > [Non-text portions of this message have been removed]
> >
> >
> >
> > ------------------------------------
>
> >
> >
> >
> >



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Whether they do or not seems irrelevant. If they do they don't seem to
have paid much attention to the low power interests over time. Anybody
and everybody can make quick cheap micros with lots of memory, they're
all out there doing it, many of them quicker and cheaper than Ti, so if
they were paying attention they wouldn't have designed not so quick not
so cheap, slightly higher current than before micros, whose only big
difference was a stable DCO. I think that is probably because control of
the MSP was wrested from the original design group and given to the US,
where BIG IRON rules, big cars, big engines, big micros, big everything,
the TRUE TEXAN MENTALITY, everything bigger is better. 6000 SERIES dsp
RULES THE ROOST.

Don't get me wrong, I was a huge fan of the MSP430, still am, still
designing it in. In the early days I was one of its few fans, probably a
lone voice in the wilderness offering support on the net long before Ti
had their documentation sorted out I was providing support for Ti's own
customers, I even had Ti FAE's call me on several occasions for support.
I still have some of the early dev tools, the serial port JTAG, some of
the software that only worked with specific chip batches, and certain
versions of the tools, and original 1993+ manuals for the C series.

It just galls me that not much has moved in terms of low power in the
intervening years. I want them to do it, and do it well, it's like a
pair of old socks, I want to put them on tomorrow, I don't want to have
to support and learn lots of new families of micros, I'm getting too old
for that.

Al

John Westmoreland wrote:
> Al,
>
> We all know that TI 'monitors' (reads) the posts in this group - and I am
> sure they have taken some of the issues discussed to heart - and I hope also
> they have been working on an ultra-low power design. Would seem that is an
> obvious area of research.
>
> I still think the '1611 was (is) a great part - especially when it was made.
>
> Maybe a TIer will chime in.
>
> Regards,
> John W.
>
> On Mon, Oct 26, 2009 at 8:06 PM, OneStone wrote:
>
>>
>> EM have started releasing to their financial partners, and say they will
>> sampled to the rest of us in December with product being available in
>> early 2010.
>>
>> The point of my original email is that Ti have sat on their hands, as
>> far as low power development is concerned, since the release of the
>> MSP430F series. Instead they have concentrated on the typical
>> 'development path' taken by other microcontroller suppliers:- more
>> memory, more pins, more peripherals, more speed. In the process they
>> seem to have forgotten what originally made their market niche for them.
>>
>> Over the last 10 years or so lots of other manufacturers have been
>> creeping up on Ti's low power claims, some of them slowly taking bites
>> out of it, while Ti seem to have abandoned any efforts at further
>> developments along the low power path and thrown their future into the
>> wireless basket. Ironically this is based on a product originally
>> developed by the people who are possibly going to prove their most
>> serious competition in the low power market.
>>
>> I hope I'm wrong and that Ti have secretly been working on some new kick
>> ass 50uA/MHz technology with even lower LP modes, but somehow I think not.
>>
>> Cheers
>>
>> Al
>> John Westmoreland wrote:
>>> Does Energy Micro have any released products - and if so, has anyone used
>>> them?
>>>
>>> I would not expect TI to sit on their hands idly and watch this market
>>> vaporize.
>>> CC2531 is a decent part - albeit I wish it had the '430 core instead of
>> the
>>> '8051.
>>>
>>> Regards,
>>> John W.
>>>
>>> On Sat, Oct 24, 2009 at 1:05 PM, OneStone >
>> wrote:
>>>> Well lots have tried, and some have come close, but for a long time the
>>>> MSP has ruled the roost when it comes to low power, but finally I think
>>>> it may have met its match, and not because anyone else has done anything
>>>> spectacular, but because Ti has basically failed to do anything. WEll,
>>>> that's not entirely true. They've taken a micro whose greatest asset,
>>>> and major marketing point, was it's low power performance, and then
>>>> totally neglected any further low power development for the last 10
>>>> years or more, developing it along the lines of every other micro
>>>> family, going for big memory, big RAM, big pin counts, damn for those in
>>>> Aus they could almost have thrown in the big PRUNE while they were at
>>>> it. In fact if you examine the data sheets over the years you find a
>>>> gradual creep up in uA/MHz on some parts.
>>>>
>>>> I've been hunting for a lower power part for years, a useful low power
>>>> part, one that doesn't need resetting to come back to life again, or at
>>>> least one that can run an RTC and precision timing while in LP modes.
>>>> The MSP was always expensive. The newer parts sorted that slightly.
>>>>
>>>> Now it looks like Energy Micro is a real contender, true low power sleep
>>>> with RTC and some neat timing, plus dedicated low power peripherals like
>>>> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
>>>> but so far the specs look very good. Only major downer is the lack of
>>>> any small parts, everything starts at 32K flash and 8kRAM, and just
>>>> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
>>>> somewhat bigger than the smaller MSPs, and for me small is good.
>>>>
>>>> For those who don't know, ENERGY MICROS founders include some of the ex
>>>> CHIPCON people responsible for the CC11xx etc.
>>>>
>>>> Cheers
>>>>
>>>> Al
>>>>
>>>>
>>>
>>> [Non-text portions of this message have been removed]
>>>
>>>
>>>
>>> ------------------------------------
>>>
>>>
>>>



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Al, you're right, TI hasn't much really done in the last few years. But the question is, what have the other manufacturer done in the last 10 years, especially regarding the ultra-low power sector? Are there any comparable micros available on the market? The answer is 'No'.

Stefan

--- In m...@yahoogroups.com, OneStone wrote:
>
> Whether they do or not seems irrelevant. If they do they don't seem to
> have paid much attention to the low power interests over time. Anybody
> and everybody can make quick cheap micros with lots of memory, they're
> all out there doing it, many of them quicker and cheaper than Ti, so if
> they were paying attention they wouldn't have designed not so quick not
> so cheap, slightly higher current than before micros, whose only big
> difference was a stable DCO. I think that is probably because control of
> the MSP was wrested from the original design group and given to the US,
> where BIG IRON rules, big cars, big engines, big micros, big everything,
> the TRUE TEXAN MENTALITY, everything bigger is better. 6000 SERIES dsp
> RULES THE ROOST.
>
> Don't get me wrong, I was a huge fan of the MSP430, still am, still
> designing it in. In the early days I was one of its few fans, probably a
> lone voice in the wilderness offering support on the net long before Ti
> had their documentation sorted out I was providing support for Ti's own
> customers, I even had Ti FAE's call me on several occasions for support.
> I still have some of the early dev tools, the serial port JTAG, some of
> the software that only worked with specific chip batches, and certain
> versions of the tools, and original 1993+ manuals for the C series.
>
> It just galls me that not much has moved in terms of low power in the
> intervening years. I want them to do it, and do it well, it's like a
> pair of old socks, I want to put them on tomorrow, I don't want to have
> to support and learn lots of new families of micros, I'm getting too old
> for that.
>
> Al
>
> John Westmoreland wrote:
> > Al,
> >
> > We all know that TI 'monitors' (reads) the posts in this group - and I am
> > sure they have taken some of the issues discussed to heart - and I hope also
> > they have been working on an ultra-low power design. Would seem that is an
> > obvious area of research.
> >
> > I still think the '1611 was (is) a great part - especially when it was made.
> >
> > Maybe a TIer will chime in.
> >
> > Regards,
> > John W.
> >
> > On Mon, Oct 26, 2009 at 8:06 PM, OneStone wrote:
> >
> >>
> >> EM have started releasing to their financial partners, and say they will
> >> sampled to the rest of us in December with product being available in
> >> early 2010.
> >>
> >> The point of my original email is that Ti have sat on their hands, as
> >> far as low power development is concerned, since the release of the
> >> MSP430F series. Instead they have concentrated on the typical
> >> 'development path' taken by other microcontroller suppliers:- more
> >> memory, more pins, more peripherals, more speed. In the process they
> >> seem to have forgotten what originally made their market niche for them.
> >>
> >> Over the last 10 years or so lots of other manufacturers have been
> >> creeping up on Ti's low power claims, some of them slowly taking bites
> >> out of it, while Ti seem to have abandoned any efforts at further
> >> developments along the low power path and thrown their future into the
> >> wireless basket. Ironically this is based on a product originally
> >> developed by the people who are possibly going to prove their most
> >> serious competition in the low power market.
> >>
> >> I hope I'm wrong and that Ti have secretly been working on some new kick
> >> ass 50uA/MHz technology with even lower LP modes, but somehow I think not.
> >>
> >> Cheers
> >>
> >> Al
> >>
> >>
> >> John Westmoreland wrote:
> >>> Does Energy Micro have any released products - and if so, has anyone used
> >>> them?
> >>>
> >>> I would not expect TI to sit on their hands idly and watch this market
> >>> vaporize.
> >>> CC2531 is a decent part - albeit I wish it had the '430 core instead of
> >> the
> >>> '8051.
> >>>
> >>> Regards,
> >>> John W.
> >>>
> >>> On Sat, Oct 24, 2009 at 1:05 PM, OneStone >
> >> wrote:
> >>>> Well lots have tried, and some have come close, but for a long time the
> >>>> MSP has ruled the roost when it comes to low power, but finally I think
> >>>> it may have met its match, and not because anyone else has done anything
> >>>> spectacular, but because Ti has basically failed to do anything. WEll,
> >>>> that's not entirely true. They've taken a micro whose greatest asset,
> >>>> and major marketing point, was it's low power performance, and then
> >>>> totally neglected any further low power development for the last 10
> >>>> years or more, developing it along the lines of every other micro
> >>>> family, going for big memory, big RAM, big pin counts, damn for those in
> >>>> Aus they could almost have thrown in the big PRUNE while they were at
> >>>> it. In fact if you examine the data sheets over the years you find a
> >>>> gradual creep up in uA/MHz on some parts.
> >>>>
> >>>> I've been hunting for a lower power part for years, a useful low power
> >>>> part, one that doesn't need resetting to come back to life again, or at
> >>>> least one that can run an RTC and precision timing while in LP modes.
> >>>> The MSP was always expensive. The newer parts sorted that slightly.
> >>>>
> >>>> Now it looks like Energy Micro is a real contender, true low power sleep
> >>>> with RTC and some neat timing, plus dedicated low power peripherals like
> >>>> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
> >>>> but so far the specs look very good. Only major downer is the lack of
> >>>> any small parts, everything starts at 32K flash and 8kRAM, and just
> >>>> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
> >>>> somewhat bigger than the smaller MSPs, and for me small is good.
> >>>>
> >>>> For those who don't know, ENERGY MICROS founders include some of the ex
> >>>> CHIPCON people responsible for the CC11xx etc.
> >>>>
> >>>> Cheers
> >>>>
> >>>> Al
> >>>>
> >>>>
> >>>
> >>> [Non-text portions of this message have been removed]
> >>>
> >>>
> >>>
> >>> ------------------------------------
> >>>
> >>>
> >>>



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Hi,

> Al, you're right, TI hasn't much really done in the last few years. But
the
> question is, what have the other manufacturer done in the last 10 years,
> especially regarding the ultra-low power sector? Are there any comparable
> micros available on the market? The answer is 'No'.

I believe the correct answer is "There will be". The Cortex-M3 is not a
low-power core, and ARM do not market it as such, it is a core that deliver=
s
more performance than an ARM7 at the same clock frequency. ARM's low-power
core is the Cortex-M0 as seen in LPC1100 devices. Perhaps Energy Micro wil=
l
look at using the Cortex-M0 at some point? Once can only conjecture...

--
Paul Curtis, Rowley Associates Ltd=A0=A0 http://www.rowley.co.uk
CrossWorks V2 is out for LPC1700, LPC3100, LPC3200, SAM9, and more!

------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

If you think that $300 is a lot for development tools, then clearly you don't value your time very much. It costs a company about $100-150 USD per hour to have an Engineer working (I wish that that was takehome pay...). That $300 is minuscule compared to overall costs.

Sure, $300 may be a lot for a "hobby" project, but the vendors don't really care about that market. I wouldn't either, except in the context that it "might" win them a real design win in the future.

For a new company, the burden of supporting many small customers is huge. As documentation becomes better, example code is written, more support tools are available, etc., they need fewer specialists to support more customers, and they can gradually bring the price down. In the last few years of MSP430, "startup" costs went from a few thousand dollars (compiler, FET/JTAG, eval boards, etc.) to free (Kickstart + USB dongle boards).

Stuart

--- In m...@yahoogroups.com, Alan Zubatch wrote:
>
> On Mon, 26 Oct 2009 08:31:59 -0000, "Paul Curtis"
> wrote:
> >> >You have to look under the downloads tab on their website. The info on
> >> >parts, dev kits etc came from being subscribed for a while and talking
> >> >to the people there. Cost of the dev kit is pretty high in my opinion
> at
> >> >US$299. The first two parts available are the smaller 32 pin parts.
> >>
> >> Yes, US$299 is too high. Unless it's a pretty serious kit. One can
> >> buy that new Kindle Amazon just announced, at list, for that much
> >> money and still have enough for a pretty decent dinner, left over.
> >
> > It's a nice kit; $299 is a fair price for it. Comparing against Kindle,
> as
> > you know, is just plain crazy. :-)
> >
> > The $299 keeps amateurs away and probably will not put off businesses
> that
> > are considering EFM32.
>
> No matter how you slice it $300 for a micro dev kit is on the high end. I
> think other companies have it right, the dev kit should be priced at or,
> even better, slightly below cost. See xilinx Spartan-3A DSP dev kit (which
> sells for $299, the FPGA goes for $40-50, and the dev board is a 12 layer
> board). The idea of a dev kit is to get your products into the hands of
> design engineers to play with and win design-ins. Make money off the
> product not the dev kit.
>
> az
>

------------------------------------

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Well they've brought out parts to try and compete. parts with low power
deep sleep modes, and dead modes, like Microchip and Atmel, but they
lose out on real world apps that require an RTC or other precision
timing during sleep. That is one reason I think the EM part might be
getting close. If you don't need an RTC or precision timed sampling a
Microchip XLP part comes damned close to the MSP430 in low power low
duty cycle apps.

Al

stefan.hauenstein wrote:
> Al, you're right, TI hasn't much really done in the last few years. But the question is, what have the other manufacturer done in the last 10 years, especially regarding the ultra-low power sector? Are there any comparable micros available on the market? The answer is 'No'.
>
> Stefan
>
>
> --- In m...@yahoogroups.com, OneStone wrote:
>> Whether they do or not seems irrelevant. If they do they don't seem to
>> have paid much attention to the low power interests over time. Anybody
>> and everybody can make quick cheap micros with lots of memory, they're
>> all out there doing it, many of them quicker and cheaper than Ti, so if
>> they were paying attention they wouldn't have designed not so quick not
>> so cheap, slightly higher current than before micros, whose only big
>> difference was a stable DCO. I think that is probably because control of
>> the MSP was wrested from the original design group and given to the US,
>> where BIG IRON rules, big cars, big engines, big micros, big everything,
>> the TRUE TEXAN MENTALITY, everything bigger is better. 6000 SERIES dsp
>> RULES THE ROOST.
>>
>> Don't get me wrong, I was a huge fan of the MSP430, still am, still
>> designing it in. In the early days I was one of its few fans, probably a
>> lone voice in the wilderness offering support on the net long before Ti
>> had their documentation sorted out I was providing support for Ti's own
>> customers, I even had Ti FAE's call me on several occasions for support.
>> I still have some of the early dev tools, the serial port JTAG, some of
>> the software that only worked with specific chip batches, and certain
>> versions of the tools, and original 1993+ manuals for the C series.
>>
>> It just galls me that not much has moved in terms of low power in the
>> intervening years. I want them to do it, and do it well, it's like a
>> pair of old socks, I want to put them on tomorrow, I don't want to have
>> to support and learn lots of new families of micros, I'm getting too old
>> for that.
>>
>> Al
>>
>> John Westmoreland wrote:
>>> Al,
>>>
>>> We all know that TI 'monitors' (reads) the posts in this group - and I am
>>> sure they have taken some of the issues discussed to heart - and I hope also
>>> they have been working on an ultra-low power design. Would seem that is an
>>> obvious area of research.
>>>
>>> I still think the '1611 was (is) a great part - especially when it was made.
>>>
>>> Maybe a TIer will chime in.
>>>
>>> Regards,
>>> John W.
>>>
>>> On Mon, Oct 26, 2009 at 8:06 PM, OneStone wrote:
>>>
>>>> EM have started releasing to their financial partners, and say they will
>>>> sampled to the rest of us in December with product being available in
>>>> early 2010.
>>>>
>>>> The point of my original email is that Ti have sat on their hands, as
>>>> far as low power development is concerned, since the release of the
>>>> MSP430F series. Instead they have concentrated on the typical
>>>> 'development path' taken by other microcontroller suppliers:- more
>>>> memory, more pins, more peripherals, more speed. In the process they
>>>> seem to have forgotten what originally made their market niche for them.
>>>>
>>>> Over the last 10 years or so lots of other manufacturers have been
>>>> creeping up on Ti's low power claims, some of them slowly taking bites
>>>> out of it, while Ti seem to have abandoned any efforts at further
>>>> developments along the low power path and thrown their future into the
>>>> wireless basket. Ironically this is based on a product originally
>>>> developed by the people who are possibly going to prove their most
>>>> serious competition in the low power market.
>>>>
>>>> I hope I'm wrong and that Ti have secretly been working on some new kick
>>>> ass 50uA/MHz technology with even lower LP modes, but somehow I think not.
>>>>
>>>> Cheers
>>>>
>>>> Al
>>>>
>>>>
>>>> John Westmoreland wrote:
>>>>> Does Energy Micro have any released products - and if so, has anyone used
>>>>> them?
>>>>>
>>>>> I would not expect TI to sit on their hands idly and watch this market
>>>>> vaporize.
>>>>> CC2531 is a decent part - albeit I wish it had the '430 core instead of
>>>> the
>>>>> '8051.
>>>>>
>>>>> Regards,
>>>>> John W.
>>>>>
>>>>> On Sat, Oct 24, 2009 at 1:05 PM, OneStone >
>>>> wrote:
>>>>>> Well lots have tried, and some have come close, but for a long time the
>>>>>> MSP has ruled the roost when it comes to low power, but finally I think
>>>>>> it may have met its match, and not because anyone else has done anything
>>>>>> spectacular, but because Ti has basically failed to do anything. WEll,
>>>>>> that's not entirely true. They've taken a micro whose greatest asset,
>>>>>> and major marketing point, was it's low power performance, and then
>>>>>> totally neglected any further low power development for the last 10
>>>>>> years or more, developing it along the lines of every other micro
>>>>>> family, going for big memory, big RAM, big pin counts, damn for those in
>>>>>> Aus they could almost have thrown in the big PRUNE while they were at
>>>>>> it. In fact if you examine the data sheets over the years you find a
>>>>>> gradual creep up in uA/MHz on some parts.
>>>>>>
>>>>>> I've been hunting for a lower power part for years, a useful low power
>>>>>> part, one that doesn't need resetting to come back to life again, or at
>>>>>> least one that can run an RTC and precision timing while in LP modes.
>>>>>> The MSP was always expensive. The newer parts sorted that slightly.
>>>>>>
>>>>>> Now it looks like Energy Micro is a real contender, true low power sleep
>>>>>> with RTC and some neat timing, plus dedicated low power peripherals like
>>>>>> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
>>>>>> but so far the specs look very good. Only major downer is the lack of
>>>>>> any small parts, everything starts at 32K flash and 8kRAM, and just
>>>>>> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
>>>>>> somewhat bigger than the smaller MSPs, and for me small is good.
>>>>>>
>>>>>> For those who don't know, ENERGY MICROS founders include some of the ex
>>>>>> CHIPCON people responsible for the CC11xx etc.
>>>>>>
>>>>>> Cheers
>>>>>>
>>>>>> Al
>>>>>>
>>>>>>
>>>>> [Non-text portions of this message have been removed]
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------
>>>>>
>>>>>
>>>>>



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

When I started on the MSP430F series in 1998/9 the tools were around the
$200 mark. These compared reasonably favourably to Motorola, whose tools
were crap anyway, but badly to Microchip, whose tools, generally were
good, and whose documentation was excellent. the support cost comes
through predominantly through poor documentation, not cost of
development tools, or necessarily their quality. This isn't 1980 when an
ICE cost $16000. It is a far more competitive era, where the micro sales
dominate. Microchip is in the position its in now because it hit the
market with an absolutely shit product, but a free IDE, which was
probably the best around at that time, inexpensive tools, costing as low
as $28, and great documentation, oh, and a far better philosophy when it
came to supporting smaller players than the then market leader,
Motorola. Funny, I don't see moto in the micro game any more, and
microchip are doing pretty nicely. Same goes for Atmel, similar policy,
cheap, quality tools, same for the MSP430 eventually with the advent of
the MSP430F. I was interested in the MSP430C, the architecture looked
interesting, but frankly at that time the low power wasn't of interest,
and the tools were ludicrously expensive.

This is primarily a tool to allow people to evaluate a product, and one
of its' main aims is to convince the purchaser to switch from their
current favoured device to your product. they already have time and
money invested in the competitors product so it makes absolutely no
sense to make it any less attract for them to at least try your product out.

The argument that it has taken time to develop the tool is true, but
then again not really, because without some kind of evaluation tool they
themselves have no in house test and development facility, nor do they
have any means of selling it. It is therefore an essential part of the
product, which could not be sold without such tools, and therefore the
development effort should be considered and recovered from the sales of
the silicon, not the sales of the tool.

I value my time because my task is to develop designs based upon a
collection of several hardware components designed and built by others,
and then to design software that will make that design perform some
hopefully interesting function. Thus my time and effort is solely to
produce composite systems. However the task of an IC supply company is
primarily to design and sell IC's. To do so they must supply the
infrastructure necessary to make those IC's useful. This includes data
sheets, application notes, errata, programming tools, evaluation tools,
advertising etc. Just as I don't expect to have to pay for the
documentation necessary to to use their IC in my design, as I expect
them to have accounted for the time to produce that documentation, when
they also accounted for the time to design and develop the device, and
the time to develop the rest of the supporting materiele, including such
mundane things as company infrastructure, I expect the development time
for the tools to have been a necessary part of the entire design process.

Final note on your obvious disdain for 'hobby' projects. Again Microchip
supported hobbyists and built a muylti-milliuon dollar market from it.
Parallax ain't doing so bad either last time I looked.

Al

Stuart_Rubin wrote:
> If you think that $300 is a lot for development tools, then clearly
> you don't value your time very much. It costs a company about
> $100-150 USD per hour to have an Engineer working (I wish that that
> was takehome pay...). That $300 is minuscule compared to overall
> costs.
>
> Sure, $300 may be a lot for a "hobby" project, but the vendors don't
> really care about that market. I wouldn't either, except in the
> context that it "might" win them a real design win in the future.
>
> For a new company, the burden of supporting many small customers is
> huge. As documentation becomes better, example code is written, more
> support tools are available, etc., they need fewer specialists to
> support more customers, and they can gradually bring the price down.
> In the last few years of MSP430, "startup" costs went from a few
> thousand dollars (compiler, FET/JTAG, eval boards, etc.) to free
> (Kickstart + USB dongle boards).
>
> Stuart
>
> --- In m...@yahoogroups.com, Alan Zubatch wrote:
>> On Mon, 26 Oct 2009 08:31:59 -0000, "Paul Curtis" wrote:
>>>>> You have to look under the downloads tab on their website.
>>>>> The info on parts, dev kits etc came from being subscribed
>>>>> for a while and talking to the people there. Cost of the dev
>>>>> kit is pretty high in my opinion
>> at
>>>>> US$299. The first two parts available are the smaller 32 pin
>>>>> parts.
>>>> Yes, US$299 is too high. Unless it's a pretty serious kit.
>>>> One can buy that new Kindle Amazon just announced, at list, for
>>>> that much money and still have enough for a pretty decent
>>>> dinner, left over.
>>> It's a nice kit; $299 is a fair price for it. Comparing against
>>> Kindle,
>> as
>>> you know, is just plain crazy. :-)
>>>
>>> The $299 keeps amateurs away and probably will not put off
>>> businesses
>> that
>>> are considering EFM32.
>> No matter how you slice it $300 for a micro dev kit is on the high
>> end. I think other companies have it right, the dev kit should be
>> priced at or, even better, slightly below cost. See xilinx
>> Spartan-3A DSP dev kit (which sells for $299, the FPGA goes for
>> $40-50, and the dev board is a 12 layer board). The idea of a dev
>> kit is to get your products into the hands of design engineers to
>> play with and win design-ins. Make money off the product not the
>> dev kit.
>>
>> az
>>
> ------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Ultra-low power is what has made MSP430 great over the years, and rest assured, TI is hard at work right now on a number of advances with MSP430 that will allow you to operate at power levels never thought possible. It's a bit early to talk about them in a public forum like this, but stay tuned for more details.

As an aside, I do take time to browse the posts to this forum each week. It's a great resource for engineers designing with the MSP430 and it's also gives me a window directly into the challenges you're running into when designing with MSP430. I've been paying close attention to this string in particular over the last couple of days, and while I typically don't like to intrude, I felt compelled to respond to this one. I hope you understand.

If you have some thoughts on how to make MSP430 better from an ultra-low power perspective, I'd love to hear about them. You can email me directly at s...@ti.com or call my offfice at +1.214.212.0288.

Thanks,
Scott Roller
General Manager
MSP430 Product Line
--- In m...@yahoogroups.com, OneStone wrote:
>
> Well lots have tried, and some have come close, but for a long time the
> MSP has ruled the roost when it comes to low power, but finally I think
> it may have met its match, and not because anyone else has done anything
> spectacular, but because Ti has basically failed to do anything. WEll,
> that's not entirely true. They've taken a micro whose greatest asset,
> and major marketing point, was it's low power performance, and then
> totally neglected any further low power development for the last 10
> years or more, developing it along the lines of every other micro
> family, going for big memory, big RAM, big pin counts, damn for those in
> Aus they could almost have thrown in the big PRUNE while they were at
> it. In fact if you examine the data sheets over the years you find a
> gradual creep up in uA/MHz on some parts.
>
> I've been hunting for a lower power part for years, a useful low power
> part, one that doesn't need resetting to come back to life again, or at
> least one that can run an RTC and precision timing while in LP modes.
> The MSP was always expensive. The newer parts sorted that slightly.
>
> Now it looks like Energy Micro is a real contender, true low power sleep
> with RTC and some neat timing, plus dedicated low power peripherals like
> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
> but so far the specs look very good. Only major downer is the lack of
> any small parts, everything starts at 32K flash and 8kRAM, and just
> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
> somewhat bigger than the smaller MSPs, and for me small is good.
>
> For those who don't know, ENERGY MICROS founders include some of the ex
> CHIPCON people responsible for the CC11xx etc.
>
> Cheers
>
> Al
>

------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

On Tue, 27 Oct 2009 12:52:37 -0000, you wrote:

>If you think that $300 is a lot for development tools, then clearly
>you don't value your time very much. It costs a company about
>$100-150 USD per hour to have an Engineer working (I wish that that
>was takehome pay...). That $300 is minuscule compared to overall
>costs.
>

I buy many more kits of families than I ever design into products. (I
do give away some of the kits to schools, after.) Those purchases are
to become familiar enough of those families I think _may_ some day be
useful to my work or to inform me about possibilities I may use in
instrument design that I hadn't closely considered before (educate
myself.) More often than not, I don't have a specific project or
client in mind.

Now, if I already _knew_ in advance, as your point above seems to
suggest as an axiom going into the question, that the part was going
into a product, I'd completely agree. There would be nothing to say,
really.

Just by way of a germane example for this group, I have done the same
with the MSP430 tools and I've spent time there (division routines
posted here and, perhaps, used in at least one compiler tool now.)
None of it with pay and, as of today, no MSP430 has gone into any
product I've designed or worked on as a commercial venture. Not one.
If that happens, though, I will be very very glad for the experiences
I've had beforehand which will inform my planning very well.

$300 may be a lot to some to spend without knowing in advance. Your
comment seems to rest on knowing, first.

Jon
------------------------------------

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

On Tue, 27 Oct 2009 14:01:19 -0000, Scott wrote:

>Ultra-low power is what has made MSP430 great over the years, and rest
>assured, TI is hard at work right now on a number of advances with
>MSP430 that will allow you to operate at power levels never thought
>possible. It's a bit early to talk about them in a public forum like
>this, but stay tuned for more details.
>
>As an aside, I do take time to browse the posts to this forum each
>week. It's a great resource for engineers designing with the MSP430
>and it's also gives me a window directly into the challenges you're
>running into when designing with MSP430. I've been paying close
>attention to this string in particular over the last couple of days,
>and while I typically don't like to intrude, I felt compelled to
>respond to this one. I hope you understand.
>
>If you have some thoughts on how to make MSP430 better from an
>ultra-low power perspective, I'd love to hear about them. You can
>email me directly at s...@ti.com or call my offfice at
>+1.214.212.0288.
>
>Thanks,
>Scott Roller
>General Manager
>MSP430 Product Line

It's been a mark of this group that there are some really good posters
here and contributions from all of the excellent compiler tool vendors
for the MSP430. I've been made to think more closely about details
and learned from all sides here. [I think almost every compiler tool
vendor I've had the pleasure to talk with are exceptional folks. (The
sole general exception in my life has been Green Hills.) But
contributing as much as they have here is wonderful for me to see.]

I hadn't expected this post, but it must also be a mark of excellence
that a TI product line general manager would provide just a little
insight and open his door to informed comment, as well. I consider
this kind of direct customer contact by the inner cicle as the "cost
of doing business," but far too few companies act as though they
understand that as well.

As a cofounder and director of a company selling software we developed
nationally, many years ago, we had an after-sale support staff. But I
left the first two hours of every day available for our support staff
to pass along, directly, customer contacts and problem issues which
they felt exceeded their abilities. Not only did this provide me with
a gut, visceral understanding of who our customers actually were and
what they were dealing with -- that I could not get any other way --
but it also greatly improved the overall success our customers
experienced with our product (because I guided and contributed to its
design and development.) And I think we received very high marks in
some large part to this, as a result.

There were many other activities I could have filled those two hours
with. Activities that would arguably have been better spent actually
focusing on direct development efforts. But it turns out that there
are plenty of other work for a director that is quite interruptable
and I merely moved those activities to the first part of the day where
support staff had access to me. I can't say enough about how
important that exposure, directly to customers and their reactions and
feelings and thoughts, had in improving our product and its success.

Anyway. Thanks. It goes a long way to see this kind of direct
connection taking place.

Jon
------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Although Scott did not directly address this, it reminded me of this news bit from about 20 months ago: http://www.eetimes.com/news/latest/showArticle.jhtml?articleID=206104308 .

That article may show TI not resting on its laurels.

Emmett Redd Ph.D. mailto:E...@missouristate.edu
Professor (417)836-5221
Department of Physics, Astronomy, and Materials Science
Missouri State University Fax (417)836-6226
901 SOUTH NATIONAL Lab (417)836-3770
SPRINGFIELD, MO 65897 USA Dept (417)836-5131

"In theory there is no difference between theory and practice. In practice there is." -- Yogi Berra or Jan van de Snepscheut

________________________________________
From: m...@yahoogroups.com [m...@yahoogroups.com] On Behalf Of scottdroller [s...@ti.com]
Sent: Tuesday, October 27, 2009 9:01 AM
To: m...@yahoogroups.com
Subject: [msp430] Re: IS THIS FINALLY THE START OF THE END?

Ultra-low power is what has made MSP430 great over the years, and rest assured, TI is hard at work right now on a number of advances with MSP430 that will allow you to operate at power levels never thought possible. It's a bit early to talk about them in a public forum like this, but stay tuned for more details.

As an aside, I do take time to browse the posts to this forum each week. It's a great resource for engineers designing with the MSP430 and it's also gives me a window directly into the challenges you're running into when designing with MSP430. I've been paying close attention to this string in particular over the last couple of days, and while I typically don't like to intrude, I felt compelled to respond to this one. I hope you understand.

If you have some thoughts on how to make MSP430 better from an ultra-low power perspective, I'd love to hear about them. You can email me directly at s...@ti.com or call my offfice at +1.214.212.0288.

Thanks,
Scott Roller
General Manager
MSP430 Product Line
--- In m...@yahoogroups.com, OneStone wrote:
>
> Well lots have tried, and some have come close, but for a long time the
> MSP has ruled the roost when it comes to low power, but finally I think
> it may have met its match, and not because anyone else has done anything
> spectacular, but because Ti has basically failed to do anything. WEll,
> that's not entirely true. They've taken a micro whose greatest asset,
> and major marketing point, was it's low power performance, and then
> totally neglected any further low power development for the last 10
> years or more, developing it along the lines of every other micro
> family, going for big memory, big RAM, big pin counts, damn for those in
> Aus they could almost have thrown in the big PRUNE while they were at
> it. In fact if you examine the data sheets over the years you find a
> gradual creep up in uA/MHz on some parts.
>
> I've been hunting for a lower power part for years, a useful low power
> part, one that doesn't need resetting to come back to life again, or at
> least one that can run an RTC and precision timing while in LP modes.
> The MSP was always expensive. The newer parts sorted that slightly.
>
> Now it looks like Energy Micro is a real contender, true low power sleep
> with RTC and some neat timing, plus dedicated low power peripherals like
> an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
> but so far the specs look very good. Only major downer is the lack of
> any small parts, everything starts at 32K flash and 8kRAM, and just
> grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
> somewhat bigger than the smaller MSPs, and for me small is good.
>
> For those who don't know, ENERGY MICROS founders include some of the ex
> CHIPCON people responsible for the CC11xx etc.
>
> Cheers
>
> Al
>

------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

I did not mean to leave the impression that i "disdain" hobby projects. Quite the opposite. Especially when I was younger and had no money for tools, I used whatever free tools I could find, and learned a lot which is directly applicable to my professional work now. I'm just saying that it's reasonable for a chip manufacturer to set the bar slightly higher on a new part ($300) to get started as a way to keep themselves from being flooded with support questions from hobbyists.

As an aside, I think it's wise for vendors to give away (or steeply discount) tools to Universities. Everyone wins. I know i designed-in a chip in a "real" product on my first job after using a kit given to me by an FAE while I was in school (a National Semiconductor COP8). But this probably doesn't make good business sense for a new, small company like EM.

Stuart

> Final note on your obvious disdain for 'hobby' projects. Again Microchip
> supported hobbyists and built a muylti-milliuon dollar market from it.
> Parallax ain't doing so bad either last time I looked.
>
> Al

------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Scott,

I also want to say thank you with this post.

I know you guys are working on some similar ideas - what I'd like to see:

Integrated USB controller with MSP430 Core and CC2500/CC2431/CC2531 Radio.
Please lose the 8051!
Keep up the great work!

This bleeds over into the LPRF stuff but that is where this is headed.

Does any MSP430 based DMA controller not steal clock cycles from the CPU?
If that is true - is it possible to design a fully independent DMA system?

Thanks!
John W.

On Tue, Oct 27, 2009 at 7:01 AM, scottdroller wrote:

> Ultra-low power is what has made MSP430 great over the years, and rest
> assured, TI is hard at work right now on a number of advances with MSP430
> that will allow you to operate at power levels never thought possible. It's
> a bit early to talk about them in a public forum like this, but stay tuned
> for more details.
>
> As an aside, I do take time to browse the posts to this forum each week.
> It's a great resource for engineers designing with the MSP430 and it's also
> gives me a window directly into the challenges you're running into when
> designing with MSP430. I've been paying close attention to this string in
> particular over the last couple of days, and while I typically don't like to
> intrude, I felt compelled to respond to this one. I hope you understand.
>
> If you have some thoughts on how to make MSP430 better from an ultra-low
> power perspective, I'd love to hear about them. You can email me directly at
> s...@ti.com or call my offfice at +1.214.212.0288.
>
> Thanks,
> Scott Roller
> General Manager
> MSP430 Product Line
> --- In m...@yahoogroups.com , OneStone
> wrote:
> >
> > Well lots have tried, and some have come close, but for a long time the
> > MSP has ruled the roost when it comes to low power, but finally I think
> > it may have met its match, and not because anyone else has done anything
> > spectacular, but because Ti has basically failed to do anything. WEll,
> > that's not entirely true. They've taken a micro whose greatest asset,
> > and major marketing point, was it's low power performance, and then
> > totally neglected any further low power development for the last 10
> > years or more, developing it along the lines of every other micro
> > family, going for big memory, big RAM, big pin counts, damn for those in
> > Aus they could almost have thrown in the big PRUNE while they were at
> > it. In fact if you examine the data sheets over the years you find a
> > gradual creep up in uA/MHz on some parts.
> >
> > I've been hunting for a lower power part for years, a useful low power
> > part, one that doesn't need resetting to come back to life again, or at
> > least one that can run an RTC and precision timing while in LP modes.
> > The MSP was always expensive. The newer parts sorted that slightly.
> >
> > Now it looks like Energy Micro is a real contender, true low power sleep
> > with RTC and some neat timing, plus dedicated low power peripherals like
> > an LP UART and DMA that can run in LP. Haven't fully checked it out yet,
> > but so far the specs look very good. Only major downer is the lack of
> > any small parts, everything starts at 32K flash and 8kRAM, and just
> > grows. Their smallest package is a 32pin QFN but it's 7mm on a side,
> > somewhat bigger than the smaller MSPs, and for me small is good.
> >
> > For those who don't know, ENERGY MICROS founders include some of the ex
> > CHIPCON people responsible for the CC11xx etc.
> >
> > Cheers
> >
> > Al
> >
>
[Non-text portions of this message have been removed]

------------------------------------

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

On Mon, Oct 26, 2009 at 11:58 PM, stefan.hauenstein <
s...@gmx.de> wrote:

TI hasn't much really done in the last few years. But the question is, what
> have the other manufacturer done in the last 10 years, especially regarding
> the ultra-low power sector?
>

Huh? Are you crazy? The MSP430's have gotten bigger and faster, with more
onboard peripherals, without significantly increasing power consumption.
Compare an F149 (60k flash, 2k ram, 280uA/MHz, 8MHz) with the F543x TI was
talking about at MCU day (256k flash, 16k ram, 160uA/MHz, 25MHz) That may
not be as dramatic as the difference between an 8051 and the early MSP430s,
but it's not THAT insignificant! (TI's estimate (rather
marketing-originated, but probably not TOO awful) was that a F5xxx system
would consume about a third of the power (for a given task) of some of the
earlier 430s...)

Meanwhile, Microchip did "nanowatt", Atmel did "picopower", Microchip did
"nanowatt xlp", and other vendors have been busy chasing TI's "lowest power
microcontroller" title (not to mention each other) with pretty substantial
enthusiasm.

The whole "very low power" microcontroller landscape is a LOT broader than
it was a decade ago!

(I wonder where it stops. When your micro runs as long on a battery as the
shelf life of the battery? Except batteries improve too.)
[Non-text portions of this message have been removed]

------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

That is basically the whole point, mostly what they've done is gotten
huge, bloated in their old age (a bit like me I guess! but I have a real
excuse...), and still they're quoting xxxuA/MHz at 2.2V. Not quite as
useful as it might be since most peripherals I typically want to use
don't run down that low, 3V being typically the ball park.

256k flash? sounds horrendous to me, unless I'm doing a lot of graphics,
have large wave tables, or am data logging on board I rarely use more
than about 8k of code space. Why would I in a low power application?
lots of code requires lots of time to munch through, thus the need for a
bucket load of flash suggests that it isn't going to be a particularly
low power application anyway.

AL
Bill Westfield wrote:
> On Mon, Oct 26, 2009 at 11:58 PM, stefan.hauenstein <
> s...@gmx.de> wrote:
>
> TI hasn't much really done in the last few years. But the question is, what
>> have the other manufacturer done in the last 10 years, especially regarding
>> the ultra-low power sector?
>> Huh? Are you crazy? The MSP430's have gotten bigger and faster, with more
> onboard peripherals, without significantly increasing power consumption.
> Compare an F149 (60k flash, 2k ram, 280uA/MHz, 8MHz) with the F543x TI was
> talking about at MCU day (256k flash, 16k ram, 160uA/MHz, 25MHz) That may
> not be as dramatic as the difference between an 8051 and the early MSP430s,
> but it's not THAT insignificant! (TI's estimate (rather
> marketing-originated, but probably not TOO awful) was that a F5xxx system
> would consume about a third of the power (for a given task) of some of the
> earlier 430s...)
>
> Meanwhile, Microchip did "nanowatt", Atmel did "picopower", Microchip did
> "nanowatt xlp", and other vendors have been busy chasing TI's "lowest power
> microcontroller" title (not to mention each other) with pretty substantial
> enthusiasm.
>
> The whole "very low power" microcontroller landscape is a LOT broader than
> it was a decade ago!
>
> (I wonder where it stops. When your micro runs as long on a battery as the
> shelf life of the battery? Except batteries improve too.)
> [Non-text portions of this message have been removed]
>
> ------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

As I 've told you in another post, back in 2006 I started the design of
a very low power system that maily do a lot of floating point math. In
that time I chose F169 because it had the geatest flash capabilities. No
my customer needs more calculations, and I ran out of memory. That could
be an example of why could I need mor flash space. Oh, I forgot
something. My system has to have 3 different languages, so all the text
tables consumes a lot of space.
B

OneStone escribió:
>
>
> That is basically the whole point, mostly what they've done is gotten
> huge, bloated in their old age (a bit like me I guess! but I have a real
> excuse...), and still they're quoting xxxuA/MHz at 2.2V. Not quite as
> useful as it might be since most peripherals I typically want to use
> don't run down that low, 3V being typically the ball park.
>
> 256k flash? sounds horrendous to me, unless I'm doing a lot of graphics,
> have large wave tables, or am data logging on board I rarely use more
> than about 8k of code space. Why would I in a low power application?
> lots of code requires lots of time to munch through, thus the need for a
> bucket load of flash suggests that it isn't going to be a particularly
> low power application anyway.
>
> AL
>
> Bill Westfield wrote:
> > On Mon, Oct 26, 2009 at 11:58 PM, stefan.hauenstein <
> > s...@gmx.de > wrote:
> >
> > TI hasn't much really done in the last few years. But the question
> is, what
> >> have the other manufacturer done in the last 10 years, especially
> regarding
> >> the ultra-low power sector?
> >>
> >
> > Huh? Are you crazy? The MSP430's have gotten bigger and faster, with
> more
> > onboard peripherals, without significantly increasing power consumption.
> > Compare an F149 (60k flash, 2k ram, 280uA/MHz, 8MHz) with the F543x
> TI was
> > talking about at MCU day (256k flash, 16k ram, 160uA/MHz, 25MHz)
> That may
> > not be as dramatic as the difference between an 8051 and the early
> MSP430s,
> > but it's not THAT insignificant! (TI's estimate (rather
> > marketing-originated, but probably not TOO awful) was that a F5xxx
> system
> > would consume about a third of the power (for a given task) of some
> of the
> > earlier 430s...)
> >
> > Meanwhile, Microchip did "nanowatt", Atmel did "picopower",
> Microchip did
> > "nanowatt xlp", and other vendors have been busy chasing TI's
> "lowest power
> > microcontroller" title (not to mention each other) with pretty
> substantial
> > enthusiasm.
> >
> > The whole "very low power" microcontroller landscape is a LOT
> broader than
> > it was a decade ago!
> >
> > (I wonder where it stops. When your micro runs as long on a battery
> as the
> > shelf life of the battery? Except batteries improve too.)
> >
> >
> > [Non-text portions of this message have been removed]
> >
> >
> >
> > ------------------------------------
> >
> >
> >
> >



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Right, lots of tables, one of the possible reasons I mentioned for using
more flash, but did it really need floating point math? That discussion
has gone on here many times. Floating point is almost always avoidable.

Al

b wrote:
> As I 've told you in another post, back in 2006 I started the design of
> a very low power system that maily do a lot of floating point math. In
> that time I chose F169 because it had the geatest flash capabilities. No
> my customer needs more calculations, and I ran out of memory. That could
> be an example of why could I need mor flash space. Oh, I forgot
> something. My system has to have 3 different languages, so all the text
> tables consumes a lot of space.
> B
>
> OneStone escribió:
>>
>>
>> That is basically the whole point, mostly what they've done is gotten
>> huge, bloated in their old age (a bit like me I guess! but I have a real
>> excuse...), and still they're quoting xxxuA/MHz at 2.2V. Not quite as
>> useful as it might be since most peripherals I typically want to use
>> don't run down that low, 3V being typically the ball park.
>>
>> 256k flash? sounds horrendous to me, unless I'm doing a lot of graphics,
>> have large wave tables, or am data logging on board I rarely use more
>> than about 8k of code space. Why would I in a low power application?
>> lots of code requires lots of time to munch through, thus the need for a
>> bucket load of flash suggests that it isn't going to be a particularly
>> low power application anyway.
>>
>> AL
>>
>> Bill Westfield wrote:
>>> On Mon, Oct 26, 2009 at 11:58 PM, stefan.hauenstein <
>>> s...@gmx.de > wrote:
>>>
>>> TI hasn't much really done in the last few years. But the question
>> is, what
>>>> have the other manufacturer done in the last 10 years, especially
>> regarding
>>>> the ultra-low power sector?
>>>>
>>> Huh? Are you crazy? The MSP430's have gotten bigger and faster, with
>> more
>>> onboard peripherals, without significantly increasing power consumption.
>>> Compare an F149 (60k flash, 2k ram, 280uA/MHz, 8MHz) with the F543x
>> TI was
>>> talking about at MCU day (256k flash, 16k ram, 160uA/MHz, 25MHz)
>> That may
>>> not be as dramatic as the difference between an 8051 and the early
>> MSP430s,
>>> but it's not THAT insignificant! (TI's estimate (rather
>>> marketing-originated, but probably not TOO awful) was that a F5xxx
>> system
>>> would consume about a third of the power (for a given task) of some
>> of the
>>> earlier 430s...)
>>>
>>> Meanwhile, Microchip did "nanowatt", Atmel did "picopower",
>> Microchip did
>>> "nanowatt xlp", and other vendors have been busy chasing TI's
>> "lowest power
>>> microcontroller" title (not to mention each other) with pretty
>> substantial
>>> enthusiasm.
>>>
>>> The whole "very low power" microcontroller landscape is a LOT
>> broader than
>>> it was a decade ago!
>>>
>>> (I wonder where it stops. When your micro runs as long on a battery
>> as the
>>> shelf life of the battery? Except batteries improve too.)
>>>
>>>
>>> [Non-text portions of this message have been removed]
>>>
>>>
>>>
>>> ------------------------------------
>>>
>>>
>>>
>>>



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Yes, Floating point math was one of my system's requirements.

OneStone escribió:
>
>
> Right, lots of tables, one of the possible reasons I mentioned for using
> more flash, but did it really need floating point math? That discussion
> has gone on here many times. Floating point is almost always avoidable.
>
> Al
>
> b wrote:
> > As I 've told you in another post, back in 2006 I started the design of
> > a very low power system that maily do a lot of floating point math. In
> > that time I chose F169 because it had the geatest flash
> capabilities. No
> > my customer needs more calculations, and I ran out of memory. That
> could
> > be an example of why could I need mor flash space. Oh, I forgot
> > something. My system has to have 3 different languages, so all the text
> > tables consumes a lot of space.
> > B
> >
> > OneStone escribió:
> >>
> >>
> >> That is basically the whole point, mostly what they've done is gotten
> >> huge, bloated in their old age (a bit like me I guess! but I have a
> real
> >> excuse...), and still they're quoting xxxuA/MHz at 2.2V. Not quite as
> >> useful as it might be since most peripherals I typically want to use
> >> don't run down that low, 3V being typically the ball park.
> >>
> >> 256k flash? sounds horrendous to me, unless I'm doing a lot of
> graphics,
> >> have large wave tables, or am data logging on board I rarely use more
> >> than about 8k of code space. Why would I in a low power application?
> >> lots of code requires lots of time to munch through, thus the need
> for a
> >> bucket load of flash suggests that it isn't going to be a particularly
> >> low power application anyway.
> >>
> >> AL
> >>
> >> Bill Westfield wrote:
> >>> On Mon, Oct 26, 2009 at 11:58 PM, stefan.hauenstein <
> >>> s...@gmx.de
> > wrote:
> >>>
> >>> TI hasn't much really done in the last few years. But the question
> >> is, what
> >>>> have the other manufacturer done in the last 10 years, especially
> >> regarding
> >>>> the ultra-low power sector?
> >>>>
> >>> Huh? Are you crazy? The MSP430's have gotten bigger and faster, with
> >> more
> >>> onboard peripherals, without significantly increasing power
> consumption.
> >>> Compare an F149 (60k flash, 2k ram, 280uA/MHz, 8MHz) with the F543x
> >> TI was
> >>> talking about at MCU day (256k flash, 16k ram, 160uA/MHz, 25MHz)
> >> That may
> >>> not be as dramatic as the difference between an 8051 and the early
> >> MSP430s,
> >>> but it's not THAT insignificant! (TI's estimate (rather
> >>> marketing-originated, but probably not TOO awful) was that a F5xxx
> >> system
> >>> would consume about a third of the power (for a given task) of some
> >> of the
> >>> earlier 430s...)
> >>>
> >>> Meanwhile, Microchip did "nanowatt", Atmel did "picopower",
> >> Microchip did
> >>> "nanowatt xlp", and other vendors have been busy chasing TI's
> >> "lowest power
> >>> microcontroller" title (not to mention each other) with pretty
> >> substantial
> >>> enthusiasm.
> >>>
> >>> The whole "very low power" microcontroller landscape is a LOT
> >> broader than
> >>> it was a decade ago!
> >>>
> >>> (I wonder where it stops. When your micro runs as long on a battery
> >> as the
> >>> shelf life of the battery? Except batteries improve too.)
> >>>
> >>>
> >>> [Non-text portions of this message have been removed]
> >>>
> >>>
> >>>
> >>> ------------------------------------
> >>>
> >>>
>
>
> >>>
> >>>



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

This is going slightly off the original topic, but, and please don't
take this the wrong way, was it an absolute requirement of the SYSTEM or
was it a requirement that the customer stated?

I question this mainly because your first sentence describes the system
as "very low power", and I don't see floating point and "very low power"
as synonymous with one another.

AL

b wrote:
> Yes, Floating point math was one of my system's requirements.
>
> OneStone escribió:
>>
>>
>> Right, lots of tables, one of the possible reasons I mentioned for using
>> more flash, but did it really need floating point math? That discussion
>> has gone on here many times. Floating point is almost always avoidable.
>>
>> Al
>>
>> b wrote:
>>> As I 've told you in another post, back in 2006 I started the design of
>>> a very low power system that maily do a lot of floating point math. In
>>> that time I chose F169 because it had the geatest flash
>> capabilities. No
>>> my customer needs more calculations, and I ran out of memory. That
>> could
>>> be an example of why could I need mor flash space. Oh, I forgot
>>> something. My system has to have 3 different languages, so all the text
>>> tables consumes a lot of space.
>>> B
>>>
>>> OneStone escribió:
>>>>
>>>> That is basically the whole point, mostly what they've done is gotten
>>>> huge, bloated in their old age (a bit like me I guess! but I have a
>> real
>>>> excuse...), and still they're quoting xxxuA/MHz at 2.2V. Not quite as
>>>> useful as it might be since most peripherals I typically want to use
>>>> don't run down that low, 3V being typically the ball park.
>>>>
>>>> 256k flash? sounds horrendous to me, unless I'm doing a lot of
>> graphics,
>>>> have large wave tables, or am data logging on board I rarely use more
>>>> than about 8k of code space. Why would I in a low power application?
>>>> lots of code requires lots of time to munch through, thus the need
>> for a
>>>> bucket load of flash suggests that it isn't going to be a particularly
>>>> low power application anyway.
>>>>
>>>> AL
>>>>
>>>> Bill Westfield wrote:
>>>>> On Mon, Oct 26, 2009 at 11:58 PM, stefan.hauenstein <
>>>>> s...@gmx.de
>> > wrote:
>>>>> TI hasn't much really done in the last few years. But the question
>>>> is, what
>>>>>> have the other manufacturer done in the last 10 years, especially
>>>> regarding
>>>>>> the ultra-low power sector?
>>>>>>
>>>>> Huh? Are you crazy? The MSP430's have gotten bigger and faster, with
>>>> more
>>>>> onboard peripherals, without significantly increasing power
>> consumption.
>>>>> Compare an F149 (60k flash, 2k ram, 280uA/MHz, 8MHz) with the F543x
>>>> TI was
>>>>> talking about at MCU day (256k flash, 16k ram, 160uA/MHz, 25MHz)
>>>> That may
>>>>> not be as dramatic as the difference between an 8051 and the early
>>>> MSP430s,
>>>>> but it's not THAT insignificant! (TI's estimate (rather
>>>>> marketing-originated, but probably not TOO awful) was that a F5xxx
>>>> system
>>>>> would consume about a third of the power (for a given task) of some
>>>> of the
>>>>> earlier 430s...)
>>>>>
>>>>> Meanwhile, Microchip did "nanowatt", Atmel did "picopower",
>>>> Microchip did
>>>>> "nanowatt xlp", and other vendors have been busy chasing TI's
>>>> "lowest power
>>>>> microcontroller" title (not to mention each other) with pretty
>>>> substantial
>>>>> enthusiasm.
>>>>>
>>>>> The whole "very low power" microcontroller landscape is a LOT
>>>> broader than
>>>>> it was a decade ago!
>>>>>
>>>>> (I wonder where it stops. When your micro runs as long on a battery
>>>> as the
>>>>> shelf life of the battery? Except batteries improve too.)
>>>>>
>>>>>
>>>>> [Non-text portions of this message have been removed]
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------
>>>>>
>>>>>
>>
>>
>>>>>

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Yeah, they are uncompatible. But floating point calculations were done
at a very low frequcney (maximum frequency, 2hz.... 1 day on the most of
the cases).

OneStone escribió:
>
>
> This is going slightly off the original topic, but, and please don't
> take this the wrong way, was it an absolute requirement of the SYSTEM or
> was it a requirement that the customer stated?
>
> I question this mainly because your first sentence describes the system
> as "very low power", and I don't see floating point and "very low power"
> as synonymous with one another.
>
> AL
>
> b wrote:
> > Yes, Floating point math was one of my system's requirements.
> >
> > OneStone escribió:
> >>
> >>
> >> Right, lots of tables, one of the possible reasons I mentioned for
> using
> >> more flash, but did it really need floating point math? That discussion
> >> has gone on here many times. Floating point is almost always avoidable.
> >>
> >> Al
> >>
> >> b wrote:
> >>> As I 've told you in another post, back in 2006 I started the
> design of
> >>> a very low power system that maily do a lot of floating point math. In
> >>> that time I chose F169 because it had the geatest flash
> >> capabilities. No
> >>> my customer needs more calculations, and I ran out of memory. That
> >> could
> >>> be an example of why could I need mor flash space. Oh, I forgot
> >>> something. My system has to have 3 different languages, so all the
> text
> >>> tables consumes a lot of space.
> >>> B
> >>>
> >>> OneStone escribió:
> >>>>
> >>>> That is basically the whole point, mostly what they've done is gotten
> >>>> huge, bloated in their old age (a bit like me I guess! but I have a
> >> real
> >>>> excuse...), and still they're quoting xxxuA/MHz at 2.2V. Not quite as
> >>>> useful as it might be since most peripherals I typically want to use
> >>>> don't run down that low, 3V being typically the ball park.
> >>>>
> >>>> 256k flash? sounds horrendous to me, unless I'm doing a lot of
> >> graphics,
> >>>> have large wave tables, or am data logging on board I rarely use more
> >>>> than about 8k of code space. Why would I in a low power application?
> >>>> lots of code requires lots of time to munch through, thus the need
> >> for a
> >>>> bucket load of flash suggests that it isn't going to be a
> particularly
> >>>> low power application anyway.
> >>>>
> >>>> AL
> >>>>
> >>>> Bill Westfield wrote:
> >>>>> On Mon, Oct 26, 2009 at 11:58 PM, stefan.hauenstein <
> >>>>> s...@gmx.de
>
> >> > wrote:
> >>>>> TI hasn't much really done in the last few years. But the question
> >>>> is, what
> >>>>>> have the other manufacturer done in the last 10 years, especially
> >>>> regarding
> >>>>>> the ultra-low power sector?
> >>>>>>
> >>>>> Huh? Are you crazy? The MSP430's have gotten bigger and faster, with
> >>>> more
> >>>>> onboard peripherals, without significantly increasing power
> >> consumption.
> >>>>> Compare an F149 (60k flash, 2k ram, 280uA/MHz, 8MHz) with the F543x
> >>>> TI was
> >>>>> talking about at MCU day (256k flash, 16k ram, 160uA/MHz, 25MHz)
> >>>> That may
> >>>>> not be as dramatic as the difference between an 8051 and the early
> >>>> MSP430s,
> >>>>> but it's not THAT insignificant! (TI's estimate (rather
> >>>>> marketing-originated, but probably not TOO awful) was that a F5xxx
> >>>> system
> >>>>> would consume about a third of the power (for a given task) of some
> >>>> of the
> >>>>> earlier 430s...)
> >>>>>
> >>>>> Meanwhile, Microchip did "nanowatt", Atmel did "picopower",
> >>>> Microchip did
> >>>>> "nanowatt xlp", and other vendors have been busy chasing TI's
> >>>> "lowest power
> >>>>> microcontroller" title (not to mention each other) with pretty
> >>>> substantial
> >>>>> enthusiasm.
> >>>>>
> >>>>> The whole "very low power" microcontroller landscape is a LOT
> >>>> broader than
> >>>>> it was a decade ago!
> >>>>>
> >>>>> (I wonder where it stops. When your micro runs as long on a battery
> >>>> as the
> >>>>> shelf life of the battery? Except batteries improve too.)
> >>>>>
> >>>>>
> >>>>> [Non-text portions of this message have been removed]
> >>>>>
> >>>>>
> >>>>>
> >>>>> ------------------------------------
> >>>>>
> >>>>>
>
> >>
> >>
> >>>>>

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

You're obviously much closer to the design than I, so better placed to
make the decisions. It's just that when I think low power, and I've been
trying to get quite extreme on that front lately, it means nothing
gets left out, even going so far as looking at custom batteries because
nobody makes what I want, either nothing small enough, or nothing with
the pulse current in a small package, or low leakage or whatever. I Now
have a nice collection of odd shaped and sized Li-poly and Li-SOCl cells
to play with!

Al
b wrote:
> Yeah, they are uncompatible. But floating point calculations were done
> at a very low frequcney (maximum frequency, 2hz.... 1 day on the most of
> the cases).
>
> OneStone escribió:
>>
>>
>> This is going slightly off the original topic, but, and please don't
>> take this the wrong way, was it an absolute requirement of the SYSTEM or
>> was it a requirement that the customer stated?
>>
>> I question this mainly because your first sentence describes the system
>> as "very low power", and I don't see floating point and "very low power"
>> as synonymous with one another.
>>
>> AL
>>
>> b wrote:
>>> Yes, Floating point math was one of my system's requirements.
>>>
>>> OneStone escribió:
>>>>
>>>> Right, lots of tables, one of the possible reasons I mentioned for
>> using
>>>> more flash, but did it really need floating point math? That discussion
>>>> has gone on here many times. Floating point is almost always avoidable.
>>>>
>>>> Al
>>>>
>>>> b wrote:
>>>>> As I 've told you in another post, back in 2006 I started the
>> design of
>>>>> a very low power system that maily do a lot of floating point math. In
>>>>> that time I chose F169 because it had the geatest flash
>>>> capabilities. No
>>>>> my customer needs more calculations, and I ran out of memory. That
>>>> could
>>>>> be an example of why could I need mor flash space. Oh, I forgot
>>>>> something. My system has to have 3 different languages, so all the
>> text
>>>>> tables consumes a lot of space.
>>>>> B
>>>>>
>>>>> OneStone escribió:
>>>>>> That is basically the whole point, mostly what they've done is gotten
>>>>>> huge, bloated in their old age (a bit like me I guess! but I have a
>>>> real
>>>>>> excuse...), and still they're quoting xxxuA/MHz at 2.2V. Not quite as
>>>>>> useful as it might be since most peripherals I typically want to use
>>>>>> don't run down that low, 3V being typically the ball park.
>>>>>>
>>>>>> 256k flash? sounds horrendous to me, unless I'm doing a lot of
>>>> graphics,
>>>>>> have large wave tables, or am data logging on board I rarely use more
>>>>>> than about 8k of code space. Why would I in a low power application?
>>>>>> lots of code requires lots of time to munch through, thus the need
>>>> for a
>>>>>> bucket load of flash suggests that it isn't going to be a
>> particularly
>>>>>> low power application anyway.
>>>>>>
>>>>>> AL
>>>>>>
>>>>>> Bill Westfield wrote:
>>>>>>> On Mon, Oct 26, 2009 at 11:58 PM, stefan.hauenstein <
>>>>>>> s...@gmx.de
>>
>>>> > wrote:
>>>>>>> TI hasn't much really done in the last few years. But the question
>>>>>> is, what
>>>>>>>> have the other manufacturer done in the last 10 years, especially
>>>>>> regarding
>>>>>>>> the ultra-low power sector?
>>>>>>>>
>>>>>>> Huh? Are you crazy? The MSP430's have gotten bigger and faster, with
>>>>>> more
>>>>>>> onboard peripherals, without significantly increasing power
>>>> consumption.
>>>>>>> Compare an F149 (60k flash, 2k ram, 280uA/MHz, 8MHz) with the F543x
>>>>>> TI was
>>>>>>> talking about at MCU day (256k flash, 16k ram, 160uA/MHz, 25MHz)
>>>>>> That may
>>>>>>> not be as dramatic as the difference between an 8051 and the early
>>>>>> MSP430s,
>>>>>>> but it's not THAT insignificant! (TI's estimate (rather
>>>>>>> marketing-originated, but probably not TOO awful) was that a F5xxx
>>>>>> system
>>>>>>> would consume about a third of the power (for a given task) of some
>>>>>> of the
>>>>>>> earlier 430s...)
>>>>>>>
>>>>>>> Meanwhile, Microchip did "nanowatt", Atmel did "picopower",
>>>>>> Microchip did
>>>>>>> "nanowatt xlp", and other vendors have been busy chasing TI's
>>>>>> "lowest power
>>>>>>> microcontroller" title (not to mention each other) with pretty
>>>>>> substantial
>>>>>>> enthusiasm.
>>>>>>>
>>>>>>> The whole "very low power" microcontroller landscape is a LOT
>>>>>> broader than
>>>>>>> it was a decade ago!
>>>>>>>
>>>>>>> (I wonder where it stops. When your micro runs as long on a battery
>>>>>> as the
>>>>>>> shelf life of the battery? Except batteries improve too.)
>>>>>>>
>>>>>>>
>>>>>>> [Non-text portions of this message have been removed]
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------
>>>>>>>
>>>>>>>
>>
>>>>
>>>>
>>>>>>>

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

>
> mostly what they've done is gotten huge, bloated in their old age
Heh. Software and software requirements have bloated too. As the ensuing
discussion has pointed out.
It might have been nicer to keep the same memory and clock frequency and
lower power consumption by similar factors, but that tends to be more
difficult. (in fact, since density improvements tend to come from smaller
feature sizes, and leakage currents go UP with smaller feature sizes,
increasing memory size fourfold while not incurring a penalty in static
power consumption is a pretty nice feat.)
xxxuA/MHz at 2.2V. ... most peripherals I typically want to use don't run
> down that low, 3V being typical
>

Peripherals on the MSP430? Or external peripherals? A lot of the lower
power in todays chips comes from lowering the supply voltage, and if
external peripherals haven't kept up, I'm not sure you can blame that on the
micro. And this is another reason for "bloating" the MSP430 with on-chip
peripherals; you know they'll work over the quoted operational range (for
instance, TI's brownout detector seems to get high marks compared to other
low-power CPUs whose BOD consumes a lot of power, or doesn't work (requiring
an external BOD that consumes extra power.))

I'll repeat my basic statement - the msp430 and other vendors' low power
offerings may not have advanced in exactly the directions that will make
EVERYONE happy, but it's not like there hasn't been any progress at all...

BillW
[Non-text portions of this message have been removed]

------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Why has software bloated? Why have the requirements bloated? In 2002 I
was working on a navigational system based on the 149, due to the wide
variety of different interfaces required for the different external
peripherals that were available at that time it required more code space
than my current design, which has more sensors, more features and needs
less code, because the peripherals are more integrated, share common
interfaces, and are include more on board processing.

External peripherals. It doesn't make sense to blame them for not
keeping up, in fact many peripherals are still stuck in 5V days. If you
want to talk to external components then you have to deal with what the
real world offers. Ti have been quoting clock speeds at 2.2V for the
last 10 years at least, well before there were many parts available at
3V. Therefore quoting current consumption at 2.2V is simply a marketing
ploy not a real world reflection of the state of the technology.

The MSP isn't being bloated so much with on chip peripherals as with
bucket loads of memory. The peripherals that are being loaded in aren't
exactly awe inspiring either. The original timers and ADC were
excellent, but the UART was pretty basic, and not too clever, the
current one isn't a great improvement either. The DMA is basic and
hasn't improved with age, the pull-up implementation even is crappy,
being able to write its own flash is good, but far too power hungry for
a low power part, and what about a decent crossbar so that peripherals
can be allocated as and where needed. I don't think the peripherals have
moved on that much in the last 10 years.

There's always progress, good, bad or indifferent. For a product that is
hyped entirely on its low power capabilities I would expect progress to
have meant significantly lower power not significantly larger parts. In
the last 10 years the current consumption has actually not gone far at
all. In many cases it has increased significantly, in some it has
reduced marginally, but taking the mean of the current parts vs the mean
of the original 2 parts the trend is increase not decrease.

In fact if you look purely at the quoted current draw at 2.2V it is all
over the place, which to me is a clear indicator that Ti's eye has not
been on the low power 'prize' when it comes to development.

F1121 160uA
F149,13x,15x,16x,43x,41x 280uA
F11x2 200uA
F1611 330uA
F20xx 220uA
F21x1 200uA later revised to 250uA
F22x2,22x4 250uA later revised to 270uA
f2410 355uA later revised to 270uA
F2419,261x 365uA
FG461x 400uA
F55xx 200uA
F5438 165uA

The very first of the F devices was actually the lowest power part, yet
it had more pins and more memory than the F20xx series, which drew
significantly more current.

Al

Bill Westfield wrote:
>> mostly what they've done is gotten huge, bloated in their old age
> Heh. Software and software requirements have bloated too. As the ensuing
> discussion has pointed out.
> It might have been nicer to keep the same memory and clock frequency and
> lower power consumption by similar factors, but that tends to be more
> difficult. (in fact, since density improvements tend to come from smaller
> feature sizes, and leakage currents go UP with smaller feature sizes,
> increasing memory size fourfold while not incurring a penalty in static
> power consumption is a pretty nice feat.)
> xxxuA/MHz at 2.2V. ... most peripherals I typically want to use don't run
>> down that low, 3V being typical
>> Peripherals on the MSP430? Or external peripherals? A lot of the lower
> power in todays chips comes from lowering the supply voltage, and if
> external peripherals haven't kept up, I'm not sure you can blame that on the
> micro. And this is another reason for "bloating" the MSP430 with on-chip
> peripherals; you know they'll work over the quoted operational range (for
> instance, TI's brownout detector seems to get high marks compared to other
> low-power CPUs whose BOD consumes a lot of power, or doesn't work (requiring
> an external BOD that consumes extra power.))
>
> I'll repeat my basic statement - the msp430 and other vendors' low power
> offerings may not have advanced in exactly the directions that will make
> EVERYONE happy, but it's not like there hasn't been any progress at all...
>
> BillW
> [Non-text portions of this message have been removed]
>
> ------------------------------------



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Guys,

Time for me to chime in here. I have been reading this fatalist rant for a while and here is my view.

Like anything in technology, things are changing. As a professional you need to be an expert in something and have unique to scarce knowledge and experience to become valuable in the technical marketplace.

If you look at some MSP430-based products that were designed five year ago, some of these product lines have evolved to needing more memory, CPU speed, storage or IO than before. Has the MSP430 line kept up with the needs of some of these products? Some have and some have not. When you find yourself at a crossroads to where the technology you are experienced in no longer fits specs of a product you design, you have a choice to make. You either dump the technology and follow the product evolution adopting new technologies or you dump the product line and move to other markets that still consider your technology valuable.

Trends have shown that many MSP430-based products are evolving to need lower power and faster performance. This is why TI bought Luminary Micro this year. They have a wonderful ARM Cortex-M3 semiconductor line. For whatever reason, TI did not embrace this technology and subsequently there is a lot of ex-TI that started Luminary Micro and now got "outside promoted" back into TI as executives from doing so.

If you want to stick with TI and evolve your product, look at the Luminary line long and hard. If you want to stick with the MSP430, there are new markets popping up where this type of CPU is still desired.

The choice is yours. The worst thing to do now is the same thing you have been doing for the last five to ten years thinking your mortgage will still get paid somehow.

Steve
--- In m...@yahoogroups.com, OneStone wrote:
>
> Why has software bloated? Why have the requirements bloated? In 2002 I
> was working on a navigational system based on the 149, due to the wide
> variety of different interfaces required for the different external
> peripherals that were available at that time it required more code space
> than my current design, which has more sensors, more features and needs
> less code, because the peripherals are more integrated, share common
> interfaces, and are include more on board processing.
>
> External peripherals. It doesn't make sense to blame them for not
> keeping up, in fact many peripherals are still stuck in 5V days. If you
> want to talk to external components then you have to deal with what the
> real world offers. Ti have been quoting clock speeds at 2.2V for the
> last 10 years at least, well before there were many parts available at
> 3V. Therefore quoting current consumption at 2.2V is simply a marketing
> ploy not a real world reflection of the state of the technology.
>
> The MSP isn't being bloated so much with on chip peripherals as with
> bucket loads of memory. The peripherals that are being loaded in aren't
> exactly awe inspiring either. The original timers and ADC were
> excellent, but the UART was pretty basic, and not too clever, the
> current one isn't a great improvement either. The DMA is basic and
> hasn't improved with age, the pull-up implementation even is crappy,
> being able to write its own flash is good, but far too power hungry for
> a low power part, and what about a decent crossbar so that peripherals
> can be allocated as and where needed. I don't think the peripherals have
> moved on that much in the last 10 years.
>
> There's always progress, good, bad or indifferent. For a product that is
> hyped entirely on its low power capabilities I would expect progress to
> have meant significantly lower power not significantly larger parts. In
> the last 10 years the current consumption has actually not gone far at
> all. In many cases it has increased significantly, in some it has
> reduced marginally, but taking the mean of the current parts vs the mean
> of the original 2 parts the trend is increase not decrease.
>
> In fact if you look purely at the quoted current draw at 2.2V it is all
> over the place, which to me is a clear indicator that Ti's eye has not
> been on the low power 'prize' when it comes to development.
>
> F1121 160uA
> F149,13x,15x,16x,43x,41x 280uA
> F11x2 200uA
> F1611 330uA
> F20xx 220uA
> F21x1 200uA later revised to 250uA
> F22x2,22x4 250uA later revised to 270uA
> f2410 355uA later revised to 270uA
> F2419,261x 365uA
> FG461x 400uA
> F55xx 200uA
> F5438 165uA
>
> The very first of the F devices was actually the lowest power part, yet
> it had more pins and more memory than the F20xx series, which drew
> significantly more current.
>
> Al
>
> Bill Westfield wrote:
> >> mostly what they've done is gotten huge, bloated in their old age
> >
> >
> > Heh. Software and software requirements have bloated too. As the ensuing
> > discussion has pointed out.
> > It might have been nicer to keep the same memory and clock frequency and
> > lower power consumption by similar factors, but that tends to be more
> > difficult. (in fact, since density improvements tend to come from smaller
> > feature sizes, and leakage currents go UP with smaller feature sizes,
> > increasing memory size fourfold while not incurring a penalty in static
> > power consumption is a pretty nice feat.)
> >
> >
> > xxxuA/MHz at 2.2V. ... most peripherals I typically want to use don't run
> >> down that low, 3V being typical
> >>
> >
> > Peripherals on the MSP430? Or external peripherals? A lot of the lower
> > power in todays chips comes from lowering the supply voltage, and if
> > external peripherals haven't kept up, I'm not sure you can blame that on the
> > micro. And this is another reason for "bloating" the MSP430 with on-chip
> > peripherals; you know they'll work over the quoted operational range (for
> > instance, TI's brownout detector seems to get high marks compared to other
> > low-power CPUs whose BOD consumes a lot of power, or doesn't work (requiring
> > an external BOD that consumes extra power.))
> >
> > I'll repeat my basic statement - the msp430 and other vendors' low power
> > offerings may not have advanced in exactly the directions that will make
> > EVERYONE happy, but it's not like there hasn't been any progress at all...
> >
> > BillW
> >
> >
> > [Non-text portions of this message have been removed]
> >
> >
> >
> > ------------------------------------
> >
> >
> >
> >

______________________________
Stellaris® MCU Family: New Parts, New Package, New Price.


(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )

Hi Steve, well it isn't really a fatalist rant, more an attempt to try
and shake Ti off their bums because I kind of like Ti, like dealing with
them mostly, like the fact that they do listen to the guys in the
trenches by way of monitoring these groups, and sometimes even act on
what they hear (witness the CC430, even though I think they could have
done it a hell of a lot better at least they have done it, an integrated
transceiver without an 8051!).

I am fortunate/unfortunate in that I've had a lot of time to spend
evaluating the opposition over the last 10 years, due to long periods of
no work while I recovered from regular spinal surgery. I have also spent
a lot of that time really trying to push the boundaries on size, current
consumption etc, and have been very close to a lot of the new trends in
power technologies. If you look back on similar threads you will see a
commonality in the past when others have raised the issue of low power
competitors. Normally I have been one of the vocal ones pointing out the
flaws in the competitors offerings, but I do see a really quite serious
lack of true progress in low power development from the micro
perspective, others have been trying to claw into MSP territory, and
yes, there have been some changes to the MSP, but if you look at the
MSP430 development path the thrust of the developments have not been
further improvements to low power, the fluctuations in current
consumption have almost been a side effect.

Along the way significant efforts have been going on with the sensors
that many designs need, for example accelerometers have dropped from
several mA to a few uA, and can include sleep type modes, or some
limited autonomy, or processing capacity that greatly reduces CPU
intervention, for example the highly specialised FC30 3 axis device from
ST, which is reduced to reporting 6 way orientation and/or click/double
click action, rather than actual acceleration. Then there are
temperature sensors that can schedule their own sampling and wake the
micro when their result fifo is full, GPS modules with quite
sophisticated power management built in, such that 500mAh gets you 6
months operational life or more from a wireless GPS tracker data logger.

But one stumbling block I constantly hit is that of peak current vs
size. It's easy enough to design very tiny systems, but powering them
takes a lot of work, small batteries, and other power sources simply
cannot deliver more than a few uA of current. The best small batteries
are KOH electrolyte silver oxides, but at only 1.55v they aren't able to
directly power the MSP430, and the size and weight of a switcher circuit
defeats the objective.

My issue is that I'm not trying to do the same old stuff, I'm trying to
push the envelope with respect to low power, and, quite often, though
not always, with small size. Quite frankly Ti is the only viable game in
town in most cases, I just need them to get much better than they
already are.

Al

Stealth wrote:
> Guys,
>
> Time for me to chime in here. I have been reading this fatalist rant
> for a while and here is my view.
>
> Like anything in technology, things are changing. As a professional
> you need to be an expert in something and have unique to scarce
> knowledge and experience to become valuable in the technical
> marketplace.
>
> If you look at some MSP430-based products that were designed five
> year ago, some of these product lines have evolved to needing more
> memory, CPU speed, storage or IO than before. Has the MSP430 line
> kept up with the needs of some of these products? Some have and some
> have not. When you find yourself at a crossroads to where the
> technology you are experienced in no longer fits specs of a product
> you design, you have a choice to make. You either dump the
> technology and follow the product evolution adopting new technologies
> or you dump the product line and move to other markets that still
> consider your technology valuable.
>
> Trends have shown that many MSP430-based products are evolving to
> need lower power and faster performance. This is why TI bought
> Luminary Micro this year. They have a wonderful ARM Cortex-M3
> semiconductor line. For whatever reason, TI did not embrace this
> technology and subsequently there is a lot of ex-TI that started
> Luminary Micro and now got "outside promoted" back into TI as
> executives from doing so.
>
> If you want to stick with TI and evolve your product, look at the
> Luminary line long and hard. If you want to stick with the MSP430,
> there are new markets popping up where this type of CPU is still
> desired.
>
> The choice is yours. The worst thing to do now is the same thing you
> have been doing for the last five to ten years thinking your mortgage
> will still get paid somehow.
>
> Steve
> --- In m...@yahoogroups.com, OneStone wrote:
>> Why has software bloated? Why have the requirements bloated? In
>> 2002 I was working on a navigational system based on the 149, due
>> to the wide variety of different interfaces required for the
>> different external peripherals that were available at that time it
>> required more code space than my current design, which has more
>> sensors, more features and needs less code, because the peripherals
>> are more integrated, share common interfaces, and are include more
>> on board processing.
>>
>> External peripherals. It doesn't make sense to blame them for not
>> keeping up, in fact many peripherals are still stuck in 5V days. If
>> you want to talk to external components then you have to deal with
>> what the real world offers. Ti have been quoting clock speeds at
>> 2.2V for the last 10 years at least, well before there were many
>> parts available at 3V. Therefore quoting current consumption at
>> 2.2V is simply a marketing ploy not a real world reflection of the
>> state of the technology.
>>
>> The MSP isn't being bloated so much with on chip peripherals as
>> with bucket loads of memory. The peripherals that are being loaded
>> in aren't exactly awe inspiring either. The original timers and ADC
>> were excellent, but the UART was pretty basic, and not too clever,
>> the current one isn't a great improvement either. The DMA is basic
>> and hasn't improved with age, the pull-up implementation even is
>> crappy, being able to write its own flash is good, but far too
>> power hungry for a low power part, and what about a decent crossbar
>> so that peripherals can be allocated as and where needed. I don't
>> think the peripherals have moved on that much in the last 10 years.
>> There's always progress, good, bad or indifferent. For a product
>> that is hyped entirely on its low power capabilities I would expect
>> progress to have meant significantly lower power not significantly
>> larger parts. In the last 10 years the current consumption has
>> actually not gone far at all. In many cases it has increased
>> significantly, in some it has reduced marginally, but taking the
>> mean of the current parts vs the mean of the original 2 parts the
>> trend is increase not decrease.
>>
>> In fact if you look purely at the quoted current draw at 2.2V it is
>> all over the place, which to me is a clear indicator that Ti's eye
>> has not been on the low power 'prize' when it comes to development.
>> F1121 160uA F149,13x,15x,16x,43x,41x 280uA F11x2 200uA
>> F1611 330uA F20xx 220uA F21x1 200uA later revised to
>> 250uA F22x2,22x4 250uA later revised to 270uA f2410 355uA
>> later revised to 270uA F2419,261x 365uA FG461x 400uA F55xx
>> 200uA F5438 165uA
>>
>> The very first of the F devices was actually the lowest power part,
>> yet it had more pins and more memory than the F20xx series, which
>> drew significantly more current.
>>
>> Al
>>
>> Bill Westfield wrote:
>>>> mostly what they've done is gotten huge, bloated in their old
>>>> age
>>>
>>> Heh. Software and software requirements have bloated too. As
>>> the ensuing discussion has pointed out. It might have been nicer
>>> to keep the same memory and clock frequency and lower power
>>> consumption by similar factors, but that tends to be more
>>> difficult. (in fact, since density improvements tend to come
>>> from smaller feature sizes, and leakage currents go UP with
>>> smaller feature sizes, increasing memory size fourfold while not
>>> incurring a penalty in static power consumption is a pretty nice
>>> feat.)
>>>
>>>
>>> xxxuA/MHz at 2.2V. ... most peripherals I typically want to use
>>> don't run
>>>> down that low, 3V being typical
>>>>
>>> Peripherals on the MSP430? Or external peripherals? A lot of
>>> the lower power in todays chips comes from lowering the supply
>>> voltage, and if external peripherals haven't kept up, I'm not
>>> sure you can blame that on the micro. And this is another reason
>>> for "bloating" the MSP430 with on-chip peripherals; you know
>>> they'll work over the quoted operational range (for instance,
>>> TI's brownout detector seems to get high marks compared to other
>>> low-power CPUs whose BOD consumes a lot of power, or doesn't work
>>> (requiring an external BOD that consumes extra power.))
>>>
>>> I'll repeat my basic statement - the msp430 and other vendors'
>>> low power offerings may not have advanced in exactly the
>>> directions that will make EVERYONE happy, but it's not like there
>>> hasn't been any progress at all...
>>>
>>> BillW
>>>
>>>
>>> [Non-text portions of this message have been removed]
>>>
>>>
>>>
>>> ------------------------------------
>>>
>>>
>>>
>>>



(You need to be a member of msp430 -- send a blank email to msp430-subscribe@yahoogroups.com )