Hi there - we are working on a single board computer and having some
strange issues. The single board computer is a very strange form
factor (not any of the PC-104 variants or anything like that), and we
have it installed on a board that we built many years ago (and the
person who built it is not around). This board has a mini-PCI slot as
well as an FPGA that talks to the CPU over PCI. We are trying to
upgrade from an older version of this SBC to a newer version. The CPUs
are supposedly pin for pin compatible - but the new one refuses to
boot from our board. We found, however, that if you remove the mini-
PCI card it will boot without any troubles.

Strange, right?

Now I must admit I'm not very experienced with PCI. Everything I know
about it I've learned over the last couple days trying to figure out
what is going on with this problem. Anyways - the IDSEL pins on the
FPGA and the mini-PCI card are connected to address/data bus pins
(AD25 and AD25, respectively) through 33.2 ohm resistors. It seems
that the SBC ties certain address/data pins to specific IRQs. So you
have to use very specific AD pins for the IDSELs. It also seems that
the pins you can use changed from the old SBC to the new "pin for pin
compatible" SBC. On the new SBC you can use AD19 and AD20. I decided
it was time to do some kludging. So now I have the 33.2 ohm resistors
still connected to one of the pads they were originally connected to
(the mini-PCI and FPGA IDSEL pins) while their other pads (going to
AD25/AD26) are left open. The other sides of the resistors are wired
with little teeny kludge wires to the AD19 and AD20 pins on the mini-
PCI connector.

Unfortunately, this setup has not changed things. The board still
refuses to boot with the mini-PCI card installed, but boots just fine
without it installed. I am not sure if the FPGA is coming up on PCI in
any of these situations - it's very difficult to check without having
the mini-PCI card functional as well.

Does anybody have any idea what could be going on here? I'm wondering
if my kludge wires could be causing some sort of signal integrity
issues for the PCI lines? (or maybe the floating AD25/AD26 pads that
used to be connected through resistors to the FPGA/mini-PCI pins are
causing issues?)

Thanks!

-Michael

On Sep 19, 12:00=A0pm, Michael <nleah...@gmail.com> wrote:
> Hi there - we are working on a single board computer and having some
> strange issues. The single board computer is a very strange form
> factor (not any of the PC-104 variants or anything like that), and we
> have it installed on a board that we built many years ago (and the
> person who built it is not around). This board has a mini-PCI slot as
> well as an FPGA that talks to the CPU over PCI. We are trying to
> upgrade from an older version of this SBC to a newer version. The CPUs
> are supposedly pin for pin compatible - but the new one refuses to
> boot from our board. We found, however, that if you remove the mini-
> PCI card it will boot without any troubles.
>
> Strange, right?
>
> Now I must admit I'm not very experienced with PCI. Everything I know
> about it I've learned over the last couple days trying to figure out
> what is going on with this problem. Anyways - the IDSEL pins on the
> FPGA and the mini-PCI card are connected to address/data bus pins
> (AD25 and AD25, respectively) through 33.2 ohm resistors. It seems
> that the SBC ties certain address/data pins to specific IRQs. So you
> have to use very specific AD pins for the IDSELs. It also seems that
> the pins you can use changed from the old SBC to the new "pin for pin
> compatible" SBC. On the new SBC you can use AD19 and AD20. I decided
> it was time to do some kludging. So now I have the 33.2 ohm resistors
> still connected to one of the pads they were originally connected to
> (the mini-PCI and FPGA IDSEL pins) while their other pads (going to
> AD25/AD26) are left open. The other sides of the resistors are wired
> with little teeny kludge wires to the AD19 and AD20 pins on the mini-
> PCI connector.
>
> Unfortunately, this setup has not changed things. The board still
> refuses to boot with the mini-PCI card installed, but boots just fine
> without it installed. I am not sure if the FPGA is coming up on PCI in
> any of these situations - it's very difficult to check without having
> the mini-PCI card functional as well.
>
> Does anybody have any idea what could be going on here? I'm wondering
> if my kludge wires could be causing some sort of signal integrity
> issues for the PCI lines? (or maybe the floating AD25/AD26 pads that
> used to be connected through resistors to the FPGA/mini-PCI pins are
> causing issues?)
>
> Thanks!
>
> -Michael

On Sep 19, 12:00 pm, Michael <nleah...@gmail.com> wrote:
> Hi there - we are working on a single board computer and having some
> strange issues. The single board computer is a very strange form
> factor (not any of the PC-104 variants or anything like that), and we
> have it installed on a board that we built many years ago (and the
> person who built it is not around). This board has a mini-PCI slot as
> well as an FPGA that talks to the CPU over PCI. We are trying to
> upgrade from an older version of this SBC to a newer version. The CPUs
> are supposedly pin for pin compatible - but the new one refuses to
> boot from our board. We found, however, that if you remove the mini-
> PCI card it will boot without any troubles.
>
> Strange, right?
>
> Now I must admit I'm not very experienced with PCI. Everything I know
> about it I've learned over the last couple days trying to figure out
> what is going on with this problem. Anyways - the IDSEL pins on the
> FPGA and the mini-PCI card are connected to address/data bus pins
> (AD25 and AD25, respectively) through 33.2 ohm resistors. It seems
> that the SBC ties certain address/data pins to specific IRQs. So you
> have to use very specific AD pins for the IDSELs. It also seems that
> the pins you can use changed from the old SBC to the new "pin for pin
> compatible" SBC. On the new SBC you can use AD19 and AD20. I decided
> it was time to do some kludging. So now I have the 33.2 ohm resistors
> still connected to one of the pads they were originally connected to
> (the mini-PCI and FPGA IDSEL pins) while their other pads (going to
> AD25/AD26) are left open. The other sides of the resistors are wired
> with little teeny kludge wires to the AD19 and AD20 pins on the mini-
> PCI connector.
>
> Unfortunately, this setup has not changed things. The board still
> refuses to boot with the mini-PCI card installed, but boots just fine
> without it installed. I am not sure if the FPGA is coming up on PCI in
> any of these situations - it's very difficult to check without having
> the mini-PCI card functional as well.
>
> Does anybody have any idea what could be going on here? I'm wondering
> if my kludge wires could be causing some sort of signal integrity
> issues for the PCI lines? (or maybe the floating AD25/AD26 pads that
> used to be connected through resistors to the FPGA/mini-PCI pins are
> causing issues?)
>
> Thanks!
>
> -Michael

This may be rather obvious but, the first thing I would make sure of
is that your new sbc is not expecting a 3.3 volt pci card which maybe
incompatible with your older pci card if it is expecting 5 volts.

On Sep 20, 12:17=A0pm, JE <jakeeng...@yahoo.com> wrote:
> On Sep 19, 12:00=A0pm, Michael <nleah...@gmail.com> wrote:
>
>
>
>
>
> > Hi there - we are working on a single board computer and having some
> > strange issues. The single board computer is a very strange form
> > factor (not any of the PC-104 variants or anything like that), and we
> > have it installed on a board that we built many years ago (and the
> > person who built it is not around). This board has a mini-PCI slot as
> > well as an FPGA that talks to the CPU over PCI. We are trying to
> > upgrade from an older version of this SBC to a newer version. The CPUs
> > are supposedly pin for pin compatible - but the new one refuses to
> > boot from our board. We found, however, that if you remove the mini-
> > PCI card it will boot without any troubles.
>
> > Strange, right?
>
> > Now I must admit I'm not very experienced with PCI. Everything I know
> > about it I've learned over the last couple days trying to figure out
> > what is going on with this problem. Anyways - the IDSEL pins on the
> > FPGA and the mini-PCI card are connected to address/data bus pins
> > (AD25 and AD25, respectively) through 33.2 ohm resistors. It seems
> > that the SBC ties certain address/data pins to specific IRQs. So you
> > have to use very specific AD pins for the IDSELs. It also seems that
> > the pins you can use changed from the old SBC to the new "pin for pin
> > compatible" SBC. On the new SBC you can use AD19 and AD20. I decided
> > it was time to do some kludging. So now I have the 33.2 ohm resistors
> > still connected to one of the pads they were originally connected to
> > (the mini-PCI and FPGA IDSEL pins) while their other pads (going to
> > AD25/AD26) are left open. The other sides of the resistors are wired
> > with little teeny kludge wires to the AD19 and AD20 pins on the mini-
> > PCI connector.
>
> > Unfortunately, this setup has not changed things. The board still
> > refuses to boot with the mini-PCI card installed, but boots just fine
> > without it installed. I am not sure if the FPGA is coming up on PCI in
> > any of these situations - it's very difficult to check without having
> > the mini-PCI card functional as well.
>
> > Does anybody have any idea what could be going on here? I'm wondering
> > if my kludge wires could be causing some sort of signal integrity
> > issues for the PCI lines? (or maybe the floating AD25/AD26 pads that
> > used to be connected through resistors to the FPGA/mini-PCI pins are
> > causing issues?)
>
> > Thanks!
>
> > -Michael
>
> On Sep 19, 12:00 pm, Michael <nleah...@gmail.com> wrote:
>
>
>
>
>
> > Hi there - we are working on a single board computer and having some
> > strange issues. The single board computer is a very strange form
> > factor (not any of the PC-104 variants or anything like that), and we
> > have it installed on a board that we built many years ago (and the
> > person who built it is not around). This board has a mini-PCI slot as
> > well as an FPGA that talks to the CPU over PCI. We are trying to
> > upgrade from an older version of this SBC to a newer version. The CPUs
> > are supposedly pin for pin compatible - but the new one refuses to
> > boot from our board. We found, however, that if you remove the mini-
> > PCI card it will boot without any troubles.
>
> > Strange, right?
>
> > Now I must admit I'm not very experienced with PCI. Everything I know
> > about it I've learned over the last couple days trying to figure out
> > what is going on with this problem. Anyways - the IDSEL pins on the
> > FPGA and the mini-PCI card are connected to address/data bus pins
> > (AD25 and AD25, respectively) through 33.2 ohm resistors. It seems
> > that the SBC ties certain address/data pins to specific IRQs. So you
> > have to use very specific AD pins for the IDSELs. It also seems that
> > the pins you can use changed from the old SBC to the new "pin for pin
> > compatible" SBC. On the new SBC you can use AD19 and AD20. I decided
> > it was time to do some kludging. So now I have the 33.2 ohm resistors
> > still connected to one of the pads they were originally connected to
> > (the mini-PCI and FPGA IDSEL pins) while their other pads (going to
> > AD25/AD26) are left open. The other sides of the resistors are wired
> > with little teeny kludge wires to the AD19 and AD20 pins on the mini-
> > PCI connector.
>
> > Unfortunately, this setup has not changed things. The board still
> > refuses to boot with the mini-PCI card installed, but boots just fine
> > without it installed. I am not sure if the FPGA is coming up on PCI in
> > any of these situations - it's very difficult to check without having
> > the mini-PCI card functional as well.
>
> > Does anybody have any idea what could be going on here? I'm wondering
> > if my kludge wires could be causing some sort of signal integrity
> > issues for the PCI lines? (or maybe the floating AD25/AD26 pads that
> > used to be connected through resistors to the FPGA/mini-PCI pins are
> > causing issues?)
>
> > Thanks!
>
> > -Michael
>
> This may be rather obvious but, the first thing I would make sure of
> is that your new sbc is not expecting a 3.3 volt pci card which maybe
> incompatible with your older pci card if it is expecting 5 volts.- Hide q=
uoted text -
>
Hi - so it has a mini-PCI card, not normal PCI card, attached to it -
and mini-PCI is only 3.3V, so I don't think that is it.

I should mention we have this newer SBC being used in a different
system, with the same mini-PCI card, and it works just fine. So
they're definitely compatible.

Thanks,

-Michael

JE wrote:
> 
> On Sep 19, 12:00 pm, Michael <nleah...@gmail.com> wrote:
> > Hi there - we are working on a single board computer and having some
> > strange issues. The single board computer is a very strange form
> > factor (not any of the PC-104 variants or anything like that), and we
> > have it installed on a board that we built many years ago (and the
> > person who built it is not around). This board has a mini-PCI slot as
> > well as an FPGA that talks to the CPU over PCI. We are trying to
> > upgrade from an older version of this SBC to a newer version. The CPUs
> > are supposedly pin for pin compatible - but the new one refuses to
> > boot from our board. We found, however, that if you remove the mini-
> > PCI card it will boot without any troubles.
> >
> > Strange, right?
> >
> > Now I must admit I'm not very experienced with PCI. Everything I know
> > about it I've learned over the last couple days trying to figure out
> > what is going on with this problem. Anyways - the IDSEL pins on the
> > FPGA and the mini-PCI card are connected to address/data bus pins
> > (AD25 and AD25, respectively) through 33.2 ohm resistors. It seems
> > that the SBC ties certain address/data pins to specific IRQs. So you
> > have to use very specific AD pins for the IDSELs. It also seems that
> > the pins you can use changed from the old SBC to the new "pin for pin
> > compatible" SBC. On the new SBC you can use AD19 and AD20. I decided
> > it was time to do some kludging. So now I have the 33.2 ohm resistors
> > still connected to one of the pads they were originally connected to
> > (the mini-PCI and FPGA IDSEL pins) while their other pads (going to
> > AD25/AD26) are left open. The other sides of the resistors are wired
> > with little teeny kludge wires to the AD19 and AD20 pins on the mini-
> > PCI connector.
> >
> > Unfortunately, this setup has not changed things. The board still
> > refuses to boot with the mini-PCI card installed, but boots just fine
> > without it installed. I am not sure if the FPGA is coming up on PCI in
> > any of these situations - it's very difficult to check without having
> > the mini-PCI card functional as well.
> >
> > Does anybody have any idea what could be going on here? I'm wondering
> > if my kludge wires could be causing some sort of signal integrity
> > issues for the PCI lines? (or maybe the floating AD25/AD26 pads that
> > used to be connected through resistors to the FPGA/mini-PCI pins are
> > causing issues?)
> >
> > Thanks!
> >
> > -Michael
> 
> On Sep 19, 12:00 pm, Michael <nleah...@gmail.com> wrote:
> > Hi there - we are working on a single board computer and having some
> > strange issues. The single board computer is a very strange form
> > factor (not any of the PC-104 variants or anything like that), and we
> > have it installed on a board that we built many years ago (and the
> > person who built it is not around). This board has a mini-PCI slot as
> > well as an FPGA that talks to the CPU over PCI. We are trying to
> > upgrade from an older version of this SBC to a newer version. The CPUs
> > are supposedly pin for pin compatible - but the new one refuses to
> > boot from our board. We found, however, that if you remove the mini-
> > PCI card it will boot without any troubles.
> >
> > Strange, right?
> >
> > Now I must admit I'm not very experienced with PCI. Everything I know
> > about it I've learned over the last couple days trying to figure out
> > what is going on with this problem. Anyways - the IDSEL pins on the
> > FPGA and the mini-PCI card are connected to address/data bus pins
> > (AD25 and AD25, respectively) through 33.2 ohm resistors. It seems
> > that the SBC ties certain address/data pins to specific IRQs. So you
> > have to use very specific AD pins for the IDSELs. It also seems that
> > the pins you can use changed from the old SBC to the new "pin for pin
> > compatible" SBC. On the new SBC you can use AD19 and AD20. I decided
> > it was time to do some kludging. So now I have the 33.2 ohm resistors
> > still connected to one of the pads they were originally connected to
> > (the mini-PCI and FPGA IDSEL pins) while their other pads (going to
> > AD25/AD26) are left open. The other sides of the resistors are wired
> > with little teeny kludge wires to the AD19 and AD20 pins on the mini-
> > PCI connector.
> >
> > Unfortunately, this setup has not changed things. The board still
> > refuses to boot with the mini-PCI card installed, but boots just fine
> > without it installed. I am not sure if the FPGA is coming up on PCI in
> > any of these situations - it's very difficult to check without having
> > the mini-PCI card functional as well.
> >
> > Does anybody have any idea what could be going on here? I'm wondering
> > if my kludge wires could be causing some sort of signal integrity
> > issues for the PCI lines? (or maybe the floating AD25/AD26 pads that
> > used to be connected through resistors to the FPGA/mini-PCI pins are
> > causing issues?)
> >
> > Thanks!
> >
> > -Michael
> 
> This may be rather obvious but, the first thing I would make sure of
> is that your new sbc is not expecting a 3.3 volt pci card which maybe
> incompatible with your older pci card if it is expecting 5 volts.

Why did you quote the complete original message, without even
snipping, twice?  At least remove the second quote.

-- 
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: <http://cbfalconer.home.att.net>;
            Try the download section.