PET 3032 issue

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

PET 3032 issue

Francesco Messineo
Hi all,
I made no real progress on the 3032 so far. Using the pettester.bin
image (is on zimmers.net) it seems that the RAM is good (I get a
screen full of g's, alternating with the all characters screen). So
the kernel ROM is addressed correctly.
I could exclude the CPU (I have a spare one) and the ROMs. I
substituted kernel and both basic ROM's sockets too, however I have
continuity on all pins, so it wasn't strictly needed.
16 to 1 decoder looks as it's working, when the CPU enters a loop, I
see it's cycling on most of the ROMs. Address buffers (LS244) looks
good too, but I lack a logic analyzer to really have a good picture of
what's going on :(
Data path must be ok, otherwise there're few chances that the
pettester.bin could work.
The stock kernel gives a completely black screen most of the times (it
clears the video ram, then gets stuck somewhere or enters a loop
displaying a few characters).
Next try will be swapping the ram bank's CAS lines as this problem is
starting to smell like RAM addressing error or mirror (not detected by
the simple pettester.bin).
I'm tempted in writing a custom ROM to make tests easier, but maybe
someone else already made debug ROMs for the PETs?
Any hint is welcome anyway :)

Frank  IZ8DWF

       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: PET 3032 issue

Francesco Messineo
Ok some progress.
It turns out that the petester.bin image only tests one 16K bank, not
both (on my board, only the innermost bank).
Swapping the CAS lines (R41, R42) did show some errors and I
substituted one 4116, I thought I was lucky since the petester didn't
show any error anymore, so I put back the original kernel ROM and it
did indeed boot fine.
I swapped back in their original places R41's and R42's ends and...
black screen again.
I socketed and substituted another 3x4116 but that didn't revive the
PET. So I swapped again the CAS lines and I'm going to write some test
programs to understand what are the bits that don't work as expected.
Frank IZ8DWF


On Fri, Feb 3, 2017 at 1:35 PM, Francesco Messineo
<[hidden email]> wrote:

> Hi all,
> I made no real progress on the 3032 so far. Using the pettester.bin
> image (is on zimmers.net) it seems that the RAM is good (I get a
> screen full of g's, alternating with the all characters screen). So
> the kernel ROM is addressed correctly.
> I could exclude the CPU (I have a spare one) and the ROMs. I
> substituted kernel and both basic ROM's sockets too, however I have
> continuity on all pins, so it wasn't strictly needed.
> 16 to 1 decoder looks as it's working, when the CPU enters a loop, I
> see it's cycling on most of the ROMs. Address buffers (LS244) looks
> good too, but I lack a logic analyzer to really have a good picture of
> what's going on :(
> Data path must be ok, otherwise there're few chances that the
> pettester.bin could work.
> The stock kernel gives a completely black screen most of the times (it
> clears the video ram, then gets stuck somewhere or enters a loop
> displaying a few characters).
> Next try will be swapping the ram bank's CAS lines as this problem is
> starting to smell like RAM addressing error or mirror (not detected by
> the simple pettester.bin).
> I'm tempted in writing a custom ROM to make tests easier, but maybe
> someone else already made debug ROMs for the PETs?
> Any hint is welcome anyway :)
>
> Frank  IZ8DWF

       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: PET 3032 issue

Gerrit Heitsch
On 02/04/2017 07:00 PM, Francesco Messineo wrote:

> Ok some progress.
> It turns out that the petester.bin image only tests one 16K bank, not
> both (on my board, only the innermost bank).
> Swapping the CAS lines (R41, R42) did show some errors and I
> substituted one 4116, I thought I was lucky since the petester didn't
> show any error anymore, so I put back the original kernel ROM and it
> did indeed boot fine.
> I swapped back in their original places R41's and R42's ends and...
> black screen again.
> I socketed and substituted another 3x4116 but that didn't revive the
> PET. So I swapped again the CAS lines and I'm going to write some test
> programs to understand what are the bits that don't work as expected.

Doesn't the 3032 have TIM? You should be able to find the broken RAMs
from within the Monitor.

  Gerrit



       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: PET 3032 issue

Francesco Messineo
On Sat, Feb 4, 2017 at 8:21 PM, Gerrit Heitsch
<[hidden email]> wrote:

> On 02/04/2017 07:00 PM, Francesco Messineo wrote:
>>
>> Ok some progress.
>> It turns out that the petester.bin image only tests one 16K bank, not
>> both (on my board, only the innermost bank).
>> Swapping the CAS lines (R41, R42) did show some errors and I
>> substituted one 4116, I thought I was lucky since the petester didn't
>> show any error anymore, so I put back the original kernel ROM and it
>> did indeed boot fine.
>> I swapped back in their original places R41's and R42's ends and...
>> black screen again.
>> I socketed and substituted another 3x4116 but that didn't revive the
>> PET. So I swapped again the CAS lines and I'm going to write some test
>> programs to understand what are the bits that don't work as expected.
>
>
> Doesn't the 3032 have TIM? You should be able to find the broken RAMs from
> within the Monitor.

sure it has it. But the only way I've found to have it "boot" was
swapping the CAS lines.
I'm going to test the ram with these lines swapped right now.

Frank

       Message was sent through the cbm-hackers mailing list
Reply | Threaded
Open this post in threaded view
|

Re: PET 3032 issue

Francesco Messineo
Interestingly, the failed RAM chips were UI2 and UJ2, both of them on D7 line.
The only explanation for this coincidence is that one of them failed
and "hit" the other sitting on the same bus line.
Another interesting thing is that to identify UI2 I had to fill that
bank with some locations having D7 high and some having D7 low, and
then it started to make D7 output all 1 in all locations.
That's why petester didn't identify this one (it identified problems
with UJ2 though, after swapping the CAS lines).
I think that the 3032 isn't yet fully functional. Basic runs fine for
as much as I could test, but
if I try to activate an option ROM installed in UD3 (marked SuperTool
3032, see my thread of some months ago) with SYS37000, it goes into
the monitor.
People trying this option in VICE told me it works.
So still searching for problems.
And, almost forgot, looking for hot chips, I spotted a way too hot
MC3446, so now I'm looking for one :/

Frank IZ8DWF

On Sat, Feb 4, 2017 at 8:26 PM, Francesco Messineo
<[hidden email]> wrote:

> On Sat, Feb 4, 2017 at 8:21 PM, Gerrit Heitsch
> <[hidden email]> wrote:
>> On 02/04/2017 07:00 PM, Francesco Messineo wrote:
>>>
>>> Ok some progress.
>>> It turns out that the petester.bin image only tests one 16K bank, not
>>> both (on my board, only the innermost bank).
>>> Swapping the CAS lines (R41, R42) did show some errors and I
>>> substituted one 4116, I thought I was lucky since the petester didn't
>>> show any error anymore, so I put back the original kernel ROM and it
>>> did indeed boot fine.
>>> I swapped back in their original places R41's and R42's ends and...
>>> black screen again.
>>> I socketed and substituted another 3x4116 but that didn't revive the
>>> PET. So I swapped again the CAS lines and I'm going to write some test
>>> programs to understand what are the bits that don't work as expected.
>>
>>
>> Doesn't the 3032 have TIM? You should be able to find the broken RAMs from
>> within the Monitor.
>
> sure it has it. But the only way I've found to have it "boot" was
> swapping the CAS lines.
> I'm going to test the ram with these lines swapped right now.
>
> Frank

       Message was sent through the cbm-hackers mailing list