29th Jan 2021, 7:36 pm | #81 | |
Octode
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
|
Re: Commodore PET 2001
Quote:
|
|
29th Jan 2021, 7:37 pm | #82 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,556
|
Re: Commodore PET 2001
Good job you didn't follow my advice, you'd be knee deep in IC sockets by now and the machine still wouldn't have worked.
|
29th Jan 2021, 7:56 pm | #83 | ||
Octode
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
|
Re: Commodore PET 2001
Quote:
|
||
29th Jan 2021, 11:28 pm | #84 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,587
|
Re: Commodore PET 2001
Have spent the evening plugging and unplugging RAM chips and this PET certainly isn't going down without a fight. With just the two pairs of RAMs the machine reported 1,023 bytes free which is correct. When I added the third pair the report became 2,050 whereas the figure is supposed to be 2,046. Close enough I thought. However when the next pair was added the PET displayed 1,250 bytes instead of the expected 3,070. Went back to just three pairs and got 1,250 again rather than 2,050. Since then I've tried various combinations and although I've found one more stuck chip the machine reports 1,250 bytes regardless of the number of chips fitted in addition to the basic four.
I'm confused! Alan |
30th Jan 2021, 8:53 am | #85 |
Dekatron
Join Date: May 2008
Location: Derby, UK.
Posts: 7,735
|
Re: Commodore PET 2001
The PET will be testing the amount of available memory by writing &00 to successive locations in turn, reading back the value written and keeping going until it reads back a non-zero value. This will obviously happen with an unpopulated area of the memory map, but it can also be caused by a faulty IC. In the latter case, the amount will not be a whole number of chip sizes.
An Open Source EPROM programmer ought to be able to test memory ICs. Even if there were no test for your devices with extra select lines, you could modify one for another device of the same capacity; the first part of the test program file will be describing what pin does what function, and hopefully commented well enough to be easily altered.
__________________
If I have seen further than others, it is because I was standing on a pile of failed experiments. |
30th Jan 2021, 10:51 am | #86 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,587
|
Re: Commodore PET 2001
Yes I can see how the system is reading the memory locations and is currently getting stuck at 1250 bytes. I have a feeling that I've now hit the 'poor quality socket' problem aggravated by the need to repeatedly remove and reseat the RAMs. Could keep me amused for hours.
Alan |
30th Jan 2021, 5:42 pm | #87 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,587
|
Re: Commodore PET 2001
Quick update. The PET is now operating as a 2001-4 rather than a 2001-8 with half the system RAM in position and is reporting 3071 bytes free which is correct. Of the remaining 8 RAM chips two are definitively faulty and the other 6 have question marks hanging over them. Part of the problem has undoubtedly been poor pin contacts but multiple chip failure can't yet be ruled out. I'll press on with pin/socket cleaning and testing.
Alan |
30th Jan 2021, 5:49 pm | #88 |
Dekatron
Join Date: May 2008
Location: Derby, UK.
Posts: 7,735
|
Re: Commodore PET 2001
Depending how you felt about originality, you could always bodge in a 62256 with some additional logic to generate the chip select, and get yourself an unofficial 2001-32!
__________________
If I have seen further than others, it is because I was standing on a pile of failed experiments. |
30th Jan 2021, 5:58 pm | #89 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,587
|
Re: Commodore PET 2001
The thought had crossed my mind Julie. At the moment the challenge is to establish which of the 6 RAM chips really are faulty without the benefit of specialist diagnostic tools. All good fun!
Alan |
30th Jan 2021, 6:09 pm | #90 |
Octode
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
|
Re: Commodore PET 2001
My PET 2001 bought by my father back in 1978 had 3 RAM chip failures in its first 2 years. One got replaced under warranty but when 2 more went we used it as a reason to upgrade to a 32k dynamic memory board that plugged into the CPU socket (The Star Trek program my father and I worked on was already busting the seams of 8k and went on to nearly fill the 32k!). 6550 RAM failures where commonly reported in the magazines for number of years. I suspect this is why Commodore moved to 2114s for the 2001 and then dynamic memories for their larger machines.
|
30th Jan 2021, 6:19 pm | #91 |
Dekatron
Join Date: May 2008
Location: Derby, UK.
Posts: 7,735
|
Re: Commodore PET 2001
Static RAM is easy enough to test if you have a modern computer and an "experimenter board" with enough I/O lines ..... it's not a specialist diagnostic tool if you have already got one ..... Otherwise, if you have enough good RAM, you can use this for a program to make the PET test its own memory; this can be more sophisticated and in-depth than the power-on test, including writing multiple values and reading several locations to ensure addresses are being decoded correctly.
__________________
If I have seen further than others, it is because I was standing on a pile of failed experiments. |
30th Jan 2021, 6:33 pm | #92 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,587
|
Re: Commodore PET 2001
Perhaps specialist is the wrong word Julie. Truth is it would take me longer to learn how to do the things you've suggested than to use the old fashioned method I've adopted. Whichever approach is taken I still also have to address the poor contacts issue.
Alan |
30th Jan 2021, 6:50 pm | #93 |
Octode
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
|
Re: Commodore PET 2001
In "the pet revealed" there is a simple memory test program:
Code:
5 INPUTA,B 10 PRINT"[CLEAR]":TI$="OOOOOO" 20 FORI=ATOB 21 FORY=1TO4 22 READN 23 POKEI,N:X=PEEK(I) 24 IFX=NTHEN26 25 GOSUB200 26 NEXT 27 RESTORE 30 DATA0,85,170,255 110 PRINT"[HOME][DOWN*11]";I-1024,I,TI$ 120 NEXT 130 PRINT"END OF TEST" 140 END 200 IFX=1ORX=2ORX=4ORX=8ORX=81ORX=84ORX=87ORX=93THEN300 210 IFX=162ORX=164ORX=171ORX=174ORX=247ORX=251ORX=253ORX=254THEN300 220 IFX=16ORX=32ORX=64ORX=128ORX=21ORX=69ORX=117ORX=213THEN350 230 IFX=42ORX=138ORX=186ORX=234ORX=127ORX=191ORX=223ORX=239THEN350 300 A$="I":GOT0400 350 A$="J" 400 IFI<=2047THEN500 410 IFI<=3071THEN510 420 IFI<=4095THEN520 430 IFI<=5119THEN530 440 IFI<=6143THEN540 450 IFI<=7167THEN550 460 GOTO560 500 B$="2":GOTO600 510 B$="3":GOTO600 520 B$="4":GOTO600 530 B$="5":GOTO600 540 B$="6":GOTO600 550 B$="7":GOTO600 560 B$="8":GOTO600 600 PRINT"YOU HAVE A FAULT AT ADDRESS " ;I;"IN ROW ";A$;B$;".",N,X 605 RETURN |
30th Jan 2021, 7:11 pm | #94 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,587
|
Re: Commodore PET 2001
Thanks Slothie. I'm fairly confident that the first 4k is working fine so not testing the first 2k isn't a problem. However, the datasette currently resides in the loft of my previous abode although it did work with VIC20s and C64s albeit a few years ago. I will get hold of it in due course.
Alan |
30th Jan 2021, 11:53 pm | #95 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,587
|
Re: Commodore PET 2001
RAM saga over. There turned out to be six faulty system RAM chips. Four caused a loop resulting in the frozen random character screen. For obvious reasons it wouldn't have been possible to identify these by running the memory test programme on the machine. The other two faulty 6550s allowed the system to boot ok but would not return the correct number of free bytes. This is an alarming failure rate which tends support Slothie's suggestion that Commodore moved over to 2114s because of problems with the 6550s.
The good news is that the PET is running happily as a 2001-5 with 4095 free bytes. It's satisfying to have got to this point but I now need to think about what happens to the PET next as I don't have any interest in keeping it in the longer term. More to follow. Alan |
31st Jan 2021, 9:54 am | #96 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,587
|
Re: Commodore PET 2001
Does anybody have any theories as to what might have caused such a catastrophic failure of RAM? It seems strange that so many chips failed, presumably at around the same time.
Alan |
31st Jan 2021, 10:30 am | #97 |
Octode
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
|
Re: Commodore PET 2001
6550 RAMs don't have a good reputation. Its possible that they can fail over time even when switched off. MOStek IC's have a bit of a reputation for being unreliable in the long run so I think you were just unlucky. My 2001 was fully working when packed away 30 years ago but I fully expect to have to replace some of the chips when I finally get it out of storage. I also have the exciting prospect of replacing one of the keycaps on the chicklet keyboard...
|
31st Jan 2021, 1:11 pm | #98 |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,587
|
Re: Commodore PET 2001
Please see attached ......
|
31st Jan 2021, 1:15 pm | #99 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,556
|
Re: Commodore PET 2001
How sweet!
I think you would probably get quite a decent sum for it now even with the fault honestly described, since all the prospective owner has to do is source replacement RAM or a more modern replacement in the form of a replacement RAM board, or make one themselves. |
31st Jan 2021, 1:18 pm | #100 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,556
|
Re: Commodore PET 2001
I did find this, probably only one of several solutions available, but I think finding and fitting something like that would be a matter for the new owner.
https://www.nightfallcrew.com/19/08/...-xad-manosoft/ Since the images are so good you could even lash one up on stripboard just to try it out. |