>
> It's an old embedded system which has basically no debugging hooks and has
> recently acquired a high failure rate where the system seems to load the
> first bit of the instruction stream correctly, but runs off into the weeds
> sometime before empowering any of the functionality that would allow us to
> debug the problem.
>
> --
> Aaron
In an old system that used to work fine and if initial boot code hasn't
changed, then it's probably hardware related. Most likely some device
out of spec or similar. What have you changed, either chip vendor or pcb
?. For example, is the device that the boot code loads from the same
spec in terms of access time ?.
A logic analyser is probably the best tool. An ice introduces it's own
timing skews...
Chris
--
Greenfield Designs Ltd
-----------------------------------------------------------
Embedded Systems & Electronics: Research Design Development
Oxford. England. (44) 1865 750 681
Reply by ●July 26, 20062006-07-26
Tom Lucas <news@remove_auto_this_flame_to_reply.clara.co.uk> wrote:
>
> <aborgman@redshark.goodshow.net> wrote in message
> news:1T3bajm5InjpN3l@redshark.goodshow.net...
>> Tom Lucas <news@remove_auto_this_flame_to_reply.clara.co.uk> wrote:
>
>>> What are you up to with your processor? Sounds like mischief to me
>>> ;-)
>>
>> It's an old embedded system which has basically no debugging hooks and
>> has
>> recently acquired a high failure rate where the system seems to load
>> the
>> first bit of the instruction stream correctly, but runs off into the
>> weeds
>> sometime before empowering any of the functionality that would allow
>> us to
>> debug the problem.
>
> I bet it doesn't do it with the emulator :-p Can you make "breakpoints"
> in the code that would wait for a pin to be toggled by a switch before
> progressing? That way you can break down into blocks to narrow down the
> problem.
It comes with the boot sector stuff preloaded in the eprom, so changing that
is basically impossible - and it's functionality for loading code depends on
it successfully getting through the initialization sequence.
--
Aaron
Reply by Tom Lucas●July 26, 20062006-07-26
<aborgman@redshark.goodshow.net> wrote in message
news:1T3bajm5InjpN3l@redshark.goodshow.net...
> Tom Lucas <news@remove_auto_this_flame_to_reply.clara.co.uk> wrote:
>> What are you up to with your processor? Sounds like mischief to me
>> ;-)
>
> It's an old embedded system which has basically no debugging hooks and
> has
> recently acquired a high failure rate where the system seems to load
> the
> first bit of the instruction stream correctly, but runs off into the
> weeds
> sometime before empowering any of the functionality that would allow
> us to
> debug the problem.
I bet it doesn't do it with the emulator :-p Can you make "breakpoints"
in the code that would wait for a pin to be toggled by a switch before
progressing? That way you can break down into blocks to narrow down the
problem.
Or a cheap quick test might be buying a different manufacturer's 186
clone and sticking it in to see if the problem is still there.
Reply by ●July 26, 20062006-07-26
Tom Lucas <news@remove_auto_this_flame_to_reply.clara.co.uk> wrote:
> <aborgman@redshark.goodshow.net> wrote in message
> news:2T3b8dhrI5q1N3l@redshark.goodshow.net...
>>I have a board using a 25MHz 80C186EB and I'm looking for an in circuit
>> emulator that will allow the following:
>>
>> Stepping through codestream instruction by instruction
>>
>> That is it. I want to yank the CPU off the nonfunctional board, drop a
>> socket, hook up the emulator and step through exactly what the on
>> board CPU
>> was seeing instruction by instruction.
>>
>> ...and the cheaper the better.
>>
>> Any recommendations?
>
> My project has been using two Noral 186 emulators for years but one now
> only works with hot air blown on it and the other is getting a bit
> twitchy as time goes on.
> We've been looking at replacements and the availalbe brands seem to be:
> Hitex,
> Ceibo,
> iSystems,
> Signum
> Lauterbach.
> You'll be looking at GBP5K for the cheaper ones and 20K for the
> Lauterbach but it is very good apparently.
>
> We have an 186EB attachment for our broken Noral emulator as well so we
> could have sold it to you if it worked :-( If you can find a secondhand
> one somewhere we might be able to help out.
>
> What are you up to with your processor? Sounds like mischief to me ;-)
It's an old embedded system which has basically no debugging hooks and has
recently acquired a high failure rate where the system seems to load the
first bit of the instruction stream correctly, but runs off into the weeds
sometime before empowering any of the functionality that would allow us to
debug the problem.
--
Aaron
Reply by Tom Lucas●July 26, 20062006-07-26
<aborgman@redshark.goodshow.net> wrote in message
news:2T3b8dhrI5q1N3l@redshark.goodshow.net...
>I have a board using a 25MHz 80C186EB and I'm looking for an in circuit
> emulator that will allow the following:
>
> Stepping through codestream instruction by instruction
>
> That is it. I want to yank the CPU off the nonfunctional board, drop a
> socket, hook up the emulator and step through exactly what the on
> board CPU
> was seeing instruction by instruction.
>
> ...and the cheaper the better.
>
> Any recommendations?
My project has been using two Noral 186 emulators for years but one now
only works with hot air blown on it and the other is getting a bit
twitchy as time goes on.
We've been looking at replacements and the availalbe brands seem to be:
Hitex,
Ceibo,
iSystems,
Signum
Lauterbach.
You'll be looking at GBP5K for the cheaper ones and 20K for the
Lauterbach but it is very good apparently.
We have an 186EB attachment for our broken Noral emulator as well so we
could have sold it to you if it worked :-( If you can find a secondhand
one somewhere we might be able to help out.
What are you up to with your processor? Sounds like mischief to me ;-)
Reply by ●July 25, 20062006-07-25
Jim Stewart <jstewart@jkmicro.com> wrote:
> aborgman@redshark.goodshow.net wrote:
>> I have a board using a 25MHz 80C186EB and I'm looking for an in circuit
>> emulator that will allow the following:
>>
>> Stepping through codestream instruction by instruction
>>
>> That is it. I want to yank the CPU off the nonfunctional board, drop a
>> socket, hook up the emulator and step through exactly what the on board CPU
>> was seeing instruction by instruction.
>>
>> ...and the cheaper the better.
>
> If you're using it for production debug, you
> can get nearly the same results by just wiring
> a logic analyzer to a chipclip and hooking that
> onto the board's flash/EPROM program chip.
>
> Connect the address and data lines to the
> inputs, qualify OE/ with CE/ and use that
> for your clock. Set the logic analyzer
> to state and format it to show the address
> and data lines as hex or binary.
>
> Arm the analyzer and hit reset on the target
> board. After some garbage, you should see
> the address lines go to 0xFFFF0 and data
> to the first byte of code.
>
> With a hex listing of the code, you should be
> able to find bad address and data lines, etc.
>
> I've done it and it works.
I've done it, and it works for a few instructions. After that it gets dodgy.
--
Aaron
Reply by Jim Stewart●July 25, 20062006-07-25
aborgman@redshark.goodshow.net wrote:
> I have a board using a 25MHz 80C186EB and I'm looking for an in circuit
> emulator that will allow the following:
>
> Stepping through codestream instruction by instruction
>
> That is it. I want to yank the CPU off the nonfunctional board, drop a
> socket, hook up the emulator and step through exactly what the on board CPU
> was seeing instruction by instruction.
>
> ...and the cheaper the better.
If you're using it for production debug, you
can get nearly the same results by just wiring
a logic analyzer to a chipclip and hooking that
onto the board's flash/EPROM program chip.
Connect the address and data lines to the
inputs, qualify OE/ with CE/ and use that
for your clock. Set the logic analyzer
to state and format it to show the address
and data lines as hex or binary.
Arm the analyzer and hit reset on the target
board. After some garbage, you should see
the address lines go to 0xFFFF0 and data
to the first byte of code.
With a hex listing of the code, you should be
able to find bad address and data lines, etc.
I've done it and it works.
Reply by ●July 25, 20062006-07-25
I have a board using a 25MHz 80C186EB and I'm looking for an in circuit
emulator that will allow the following:
Stepping through codestream instruction by instruction
That is it. I want to yank the CPU off the nonfunctional board, drop a
socket, hook up the emulator and step through exactly what the on board CPU
was seeing instruction by instruction.
...and the cheaper the better.
Any recommendations?
Thanks,
--
Aaron