5th Apr 2021, 12:29 am | #1321 |
Octode
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
|
Re: Non-working Commodore PET 3016
While this is true, I just wanted to raise the possibility that on PETs with CRTC chips the RAM refresh may have been accomplished using the CRTC address counters to refresh the RAM, and since I make no attempt to initialise the CRTC if fitted the RAM refresh may not happen correctly. Its possible the PET Kernal initialises the CRTC before making any use of memory thus starting up the RAM refresh, Since my "expertise" is in the early 2001 pets that had fully static RAM I just wanted to make the point that my diagnostic may misbehave on the so-called "Dynamic" PETs. Early "Dynamic" pets had no CRTC and therefore had TTL based DRAM refresh.
Last edited by Slothie; 5th Apr 2021 at 12:30 am. Reason: Last sentance |
5th Apr 2021, 1:00 am | #1322 |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,394
|
Re: Non-working Commodore PET 3016
Yes, having a look at 8032 etc. schematics, it does seem that before the 8296 / SuperPET, the 'Dynamic PET's' still used 2114 SRAM's for the video RAM - But the main RAM was multiplexed between the 6502 and the Refresh circuitry. With a bit of a description about this here:
http://www.zimmers.net/anonftp/pub/c...ynamic_pet.pdf (So no extra software support for DRAM's on these versions, without a CRTC at least should be required). I have also noticed that the PET (2001(N?) at least) seems to have built-in diagnostics in the ROM, where you have to pull the diagnostic sense line on one of the external ports low, and it then jumps to a diagnostic routine that puts ascending characters on the screen / slowly clears it with blank ones. And there's also some diagnostic loopback plugs to go with this: https://www.c64-wiki.com/wiki/PET_2001 - So providing the original ROM's are working, then this could be useful, but there doesn't seem to be too much info on how the diagnostics function / how much hardware / memory needs to be running for it to be useful. |
5th Apr 2021, 2:31 am | #1323 | |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,265
|
Re: Non-working Commodore PET 3016
Quote:
Is there enough space in the PET to allow a ZIF socket to be fitted for the ROMs? |
|
5th Apr 2021, 4:21 am | #1324 | |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
|
Re: Non-working Commodore PET 3016
Wow, I look away and and I hardly remember where we were up to.
Quote:
The solution to all this is to 'adapt' the Slothie EPROM itself. Bend out pin 18 horizontally so that when the IC is inserted into the new turned pin socket, pin 18 does not go into the socket. Then, take a micro-fine piece of wire and solder it between the upper shoulder of pin 20 (close to the body of the IC) and pin 18, so that they are linked together. This (disconnect pin 18, connect it to pin 20) is exactly what the adaptor does anyway, but by eliminating the adaptor you remove any danger of the 'fat' pins of the adaptor damaging the new socket on the mainboard. The reason I did it with an adaptor originally was to allow for the possibility of EPROMs to go back and forth with different bits of test code in them, according to need. The need for it to work reliably and not damage the motherboard socket overrides that convenience. |
|
5th Apr 2021, 4:37 am | #1325 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
|
Re: Non-working Commodore PET 3016
The idea from Julie and others that the strange 'message' Colin is seeing - presumably at the end of the 5-second screen test? Is an interesting one. If it's doing the screen test (albeit showing the wrong characters) for the usual 5 seconds before showing a corrupt message that suggests the Slothie EPROM and its connections to UD9 socket are OK, possibly, but due to Colin's observation that everything is now a very loose fit in the UD9 socket, I think that may as well be replaced anyway.
It could be worth just trying the Daver2 EPROM as UD8 (with original UD9 fitted in UD9 position) to see if the behaviour with that setup is the same as it was before, also removing and refitting the UF10 character PROM IC. |
5th Apr 2021, 10:49 am | #1326 |
Dekatron
Join Date: Apr 2012
Location: Worksop, Nottinghamshire, UK.
Posts: 5,549
|
Re: Non-working Commodore PET 3016
Surely the sacrificial socket aught to be a cheap one in a new turned pin one that has been fitted into the board.
In the past I have made adapters with hard wiring done on chips I found that pins can be removed from turned pin sockets by heating the solder end of the pin and then pushing it through from the solder side. |
5th Apr 2021, 10:59 am | #1327 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
|
5th Apr 2021, 11:08 am | #1328 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
Here's this morning's update.
First boot with the Slothie EPROM in UD9 gave me zero page error - stopped. Subsequent boots gave me :%2/ 0!`% %22/2 - 34/00%$ Removing Slothie EPROM and refitted UD9, followed by removing UD8 and fitting daver3 gave me the same outcome as in post 1301 Removing Daver3 EPROM and refitting UD8, followed by refitting Slothie EPROM gives me zero page error - stopped. I will attempt to get it failing again and then check continuity on UD9 to see if there's any flaky pins. I'll read the messages above again - if I'm honest, I'm slightly confused by which route to take. I have replacement turned pin and 'normal' 24 pin sockets which I could use to put into the existing socket, or replace the existing socket. I'm unclear whether there's a consensus of opinion on whether if I replace the existing socket with a turned pin one, I should put a sacrificial 'normal' socket into the new turned pin or another turned pin? Or no sacrificial socket at all? Colin. |
5th Apr 2021, 11:27 am | #1329 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
OK - after removal of the motherboard and continuity checking from the EPROM pins to the pins on the bottom of the motherboard, I get readings on all pins, but some were difficult to get - required some subtle pushing of the EPROM to make a connection.
I've put the motherboard back in and got the %2/ 0!`% %22/2 - 34/00%$ message. Aftere several prods of the EPROM, I now get the proper zero page error - stopped message. So my current conclusion is that the 'old' motherboard socket is faulty and needs replacing. I'll hold off for others to comment. Colin. |
5th Apr 2021, 12:31 pm | #1330 |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,394
|
Re: Non-working Commodore PET 3016
Is it possible the slight flexing of the PET's PCB, when removing / inserting the 2 test EPROM's / adaptors, is actually disturbing an intermittent connection elsewhere / on another original socket? - hence the rather intermittent results
As only certain characters on the message change (although strangely not the '-'), it does seem to point towards the character ROM (or maybe Screen RAM's if these are socketed). So it may be worth removing the IC's and re-inserting these a few times to clean the socket contacts / IC pins a bit (or do this manually). Working out the difference in the PETSCII codes between correct & corrupt, may also reveal something like a constant bit-stuck error. Another thought is maybe something is not quite initialising / a slight timing issue in accessing the screen RAM. Have you tried resetting / power-cycling it say 10 times, without disturbing anything, to see if message is always consistent or it may alternate between OK and the corrupt one (which I presume is always corrupt in the same way?) |
5th Apr 2021, 12:34 pm | #1331 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
|
Re: Non-working Commodore PET 3016
Agree. So my suggestion, fit whichever socket you prefer, and do the small mod to the Slothie eprom so that you can fit it directly into the socket without the adaptor.
|
5th Apr 2021, 1:32 pm | #1332 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
Just checked - every single character code with the exception of the spaces (32) and the dash (45) is 32 lower in the 'garbled' error message than the 'real' error message.
Don't know if that means anything, but it's interesting. Colin. |
5th Apr 2021, 1:34 pm | #1333 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
Thanks. I'll get on with removing the old and I'll fit a new turned pin socket. I'll then give your post earlier a good read and modify/fit the Slothie EPROM and let you know later on today.
I'm assuming this will take us back to where we were with a consistent zero page error - stopped message? Colin. |
5th Apr 2021, 2:03 pm | #1334 | |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,394
|
Re: Non-working Commodore PET 3016
Quote:
Is the corrupt version always repeatable, upon several resets / power-cycles, without disturbing anything on the board? / does it change if pressing down on any of the ROM's etc. whilst resetting? I can't really see that fitting the Slothie EPROM without the extra IC socket adaptor under it would change this, but no harm in trying if there are some suspicions over socket contact reliability with it. |
|
5th Apr 2021, 2:47 pm | #1335 | |
Dekatron
Join Date: May 2008
Location: Derby, UK.
Posts: 7,735
|
Re: Non-working Commodore PET 3016
Quote:
If you've previously achieved a full character set on screen, it's just about vaguely possible there's another 74LS244 somewhere that is so marginal, it cares which ROM chip is fitted.
__________________
If I have seen further than others, it is because I was standing on a pile of failed experiments. |
|
5th Apr 2021, 4:05 pm | #1336 |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,394
|
Re: Non-working Commodore PET 3016
Yes, it does appear that the corrupt version is with bit 5 always high - so actually 32 higher, not lower, than the correct message (which always has bit5 low, except for Space and '-' characters)
And so this would explain why these weren't affected, as I'd previous suggested, if problem was with bit5 being stuck high (which it could float to, if not connected properly) That's assuming you have the later (for BASIC 2 / 4) 901447-10 character ROM (which used +32 character codes for lower case letters - so was closer to proper ASCII), rather than the original (BASIC 1) 901447-08 version (which used different +64 character codes for lower case). Which I discovered from here, that does seems to explain it all fairly-well: https://www.atarimagazines.com/compu...ACTER_SETS.php And looking at the schematics: http://www.zimmers.net/anonftp/pub/c...1/320137-3.gif I also discovered that bits A0..A2 of the Char.Gen ROM are used to select which row of 8 pixels is to be outputted. So D0..D6 from the Video RAM's actually connect to A3..A9 on the (IC ref. A2) Char. Gen ROM - which has it's A10 input controlled by a 'GRAPHIC' latch that selects the lower or upper half of the Character 2KB ROM to switch between the Standard and Alternate Character sets, when you POKE'd 59468 with a value, to set or clear the latch). And D7 from the video RAM's actually controls the reverse video circuit (rather than connecting to the Character ROM) So I 'd check that IC A2 (2316 character ROM) Pin23 (A8) = 'SD5' of is making reliable contact to IC C4 (2114 - High Nibble Video RAM) Pin 13 (D1) = 'SD5' etc. Last edited by ortek_service; 5th Apr 2021 at 4:12 pm. |
5th Apr 2021, 5:03 pm | #1337 | ||
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
This is (well was until I took the old socket out - I'll let you know) consistently repeatable and did not alter between reboots, nor with any gentle pushing/prodding of the EPROM with the exception of going back to plain text some of the time.
Colin. Quote:
|
||
5th Apr 2021, 5:04 pm | #1338 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
|
Re: Non-working Commodore PET 3016
I think there clearly is a physical-intermittent problem with UD9 socket which was precipitated by plugging that adaptor into it (my bad) and that will have to be sorted first otherwise we won't know which weirdness is coming from where.
After that there may indeed be a separate fault with the char gen circuit, with any luck just a bad UF10 (character PROM) socket. |
5th Apr 2021, 5:04 pm | #1339 | ||
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
I've just realised I'm looking at the Commodore 64 PETSCII codes so I need to go dig out the PET PETSCII codes...
Colin. Quote:
|
||
5th Apr 2021, 5:07 pm | #1340 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
|
Re: Non-working Commodore PET 3016
Oh, may even be worth removing UF10 and inspecting the pins in case one of them got folded over the last time the IC was reinserted.
|