In the previous post, I had repaired a CBM 4016 board.
I had reached the stage where it was working with a PET ROM/RAM
board to replace the ROMs and known working 65xx chips.
The video RAM was also replaced with a 2114 video RAM replacement board.
At that point, I would normally leave it. The board is working, the
replacement parts are all working, the power consumption is down, only the
6545 is running warm, but that is an original part.
This board is a little unusual in being a factory 12" 4016 board, I don't
think there would have been many of those, everything was 4032 after this
point when 32K became the norm.
Let's see if I can get this working with mostly original chips.
RAM
The 16K of 4116 DRAM is all testing correctly, so nothing to do there. I would certainly not consider upgrading it to 32K by fitting eight more 4116s.
Character ROM
There isn't much to say about the character ROM. The one that was fitted looks
original for the board (late 1981), that was working and all the characters looked correct, so I haven't touched it.
System ROMs
The board arrived with the original editor ROM and four of the five ROMs
replaced with 2532 EPROMs.
There were also a set of replacement ROMs, which were some of the originals, and some others dated early 1980, probably from a 9" 40xx machine or maybe an early 12" 8032.
It is tempting to just make up a full set and try them out, but that can be
counter productive and can lead you down all sorts of wrong paths.
The way something like the PET is wired up, all those chips are wired in
parallel. The same address and data lines and the only difference being the
chip select line.
We know the chip select was bad due to the 74154 failing, so all of those ROMs
would have been selected at the same time, and all trying to write to the
databus at the same time. You read $B000 and only the $Bxxx ROM should be
selected, but all are selected and so you get the value that should be at
$B000 written to the bus along with the value at $C000, $D000, $E000, $F000.
Video RAM from $8000 and the RAM from $0000, $1000 etc.
The chips with the biggest drive transistors win and anything trying to drive
the other way gets slowly burned out.
They run hot, and fitting heatsinks isn't the solution.I see these machine full of heatsinks and worry that it's only a temporary solution and just delays the inevitable.
Find the bad ones and get them out.
Or if you want an easy life, take them all out and fit a PET/ROM RAM.
Testing the ROMs
When you are in a situation like that, one bad ROM can make it look like all ROMs are bad, or intermittent, and lead you down all sorts of wrong paths trying to identify the fault. I find it safer to test the ROMs in isolation.
This is how I test it, I fit a ZIF socket to one of the ROM sockets to make it
easier and save wear and tear. It doesn't matter that it is the wrong socket, the PET diagnostics will still calculate the checksum.
I'll give you some examples.
This is what a good chip should show:
Only that ROM, and the other slots empty.
And another good one.
Some of the ROMs however misbehave.
This one was responding to an address in the RAM, but not to it's own address.
This one was responding to everything (sorry about the photo, I don't like to
leave these running too long when I know they are in conflict).
This is the dangerous sort which will eventually take out other ROMs. The 6540
ROMs in the original PET 2001 are particularly bad at this. I would not
consider trying to get a set of those running, other that for a museum piece
that will rarely be powered on. For anything you actually want to use, get
those ROMs out and use a PET ROM/RAM board.
You can use individual ROM replacements if you prefer, but it's easier to just
replace the lot, with the bonus of the options of BASIC 1, 2 or 4 and up to
32K RAM. I'm not trying to do a hard sell, I'm a rubbish salesman, but I'd
rather you didn't waste time and money doing something which is ultimately
doomed to fail, having seen several people fail to keep a 2001 running with
6540s.
During the testing, it reads each chip 16 times. Normally the result should be
the same each time, when there are bad ROMs, you don't always get the same
result, which is shown as "Inconsistent CRC" (yes, I know it's a checksum not
a CRC, I ran out of space). This chip is definitely bad, it is also appearing
in the other ROM slots.
Other ROMs don't appear to respond at all.
Sometimes that is down to bad contacts, or bad sockets. I use the ZIF socket
to rule the socket out, if other ROMs read OK, it's down to the particular
chip.
The editor ROM was giving no response and showing signs of corrosion. I tried
cleaning that, but no change.
I was surprised to find all four of the 2532 EPROMs supplied do not respond
either. Nothing on the diagnostics, and the Stag programmer was not able to
get any response either.
The end result was three of the supplied ROMs worked, but I was still missing
two and I didn't have any spares of those, so time for EPROM replacements.
I used a (46)2532 to replace the 901465-21 ROM,which is a 4K ROM with a
non-standard pinout which matches the 4K masked ROMs in the PET.
A 28C16 EEPROM replaces the 2K editor ROM, 901498-01, 50Hz 40 column editor
with the normal keyboard.
And that's a full set of ROMs working (for the moment)
Would a full set of EPROMs be better? Probably
Would a PET ROM/RAM board be even better? Absolutely.
All tests passed.
I left that running for a while, and no problems were found.
65xx chips
I had been using two W65C21N and one W65C22N chips, alongside a vintage 6545.
Time to add a vintage 6502 to the mix and let it run by itself for the first
time. Two were supplied. The one which was probably the original was dead, and
the other was dated 1979, presumably from yet another PET. That did work, but
had a patch of superglue presumably from a heatsink stubbornly stuck to it, so
one from my stash of around the right vintage was selected.
Great, it's running under it's own steam.
The supplied single 6520 was bad, so I found two of those and I think the
supplied 6522 was OK.
If you don't look too closely, you probably wouldn't spot the damage that was
fixed last week.
Again, it's a choice of running old chips or new ones. There's no choice with
the 6502 as the W65C02S is not compatible. There are later 6845 chips that can
be used to replace the 6545, but they are still 30+ years old.
Video RAM
The final step to putting this all back to good was the video RAM. I had
fitted two sockets for the good chip and the bad chip, so I could use the 2114
RAM replacement.
I found a similarly enigmatic 2114 in my spares, from another PET where the
markings had wiped off or maybe melted off.
All done
That all seems to be working.
Ah, one last thing to do, the piezo transducer needs to be replaced
(properly).
The original was missing, it seems to be in such a position that you naturally
pick the board up by that point and could push it out. I've almost done that a
couple of times.
I had temporarily replaced that by one of the piezos I use on the Mini PETs,
but it's not ideal, so I looked for something like the original.
I found these, the right size, just with wires, but it was easy enough to take
the black wire off and shorten the red one and solder it back in the original
place.
That all looks good, it fits nicely, and sounds just like it should, and
finishes off a sort of restoration of this unusual CBM 4016 board.
Testing
I have left this running my old BASIC RAM test program. This is quite slow and
inefficient, as it is written in BASIC, but that means it exercises all the
ROMs as well as the RAM, so is quite a good soak test.
I can't load 3D Monster Maze or Tut-Tut in 16K, but Invaders is playing
nicely.
I sometimes wonder about the order of events. You end up with a whole stack of
bad chips, and you wonder which failed first, or was it something else like a
voltage spike or static shock?
Was it trying to repair one fault that caused the others? How much run time
did it have with the dead or dying 74154? That probably took out a lot of the
ROMs and even the EPROMs, but was that first, or did something drag that down?
Post script
I was looking through some previous repairs and came across this 8032 board I repaired in 2017.
It is the same revision board as the 4016, with similar date codes, but with 32K and 80 column video.
That has all four MOS 40 pin chips in sockets, but the 6545 is soldered directly
and is a Hitachi part. I think the 1G3 on the top row is a date, the 1 meaning
1981. The replacement that was fitted to the 4016 when it arrived was marked
4E4 (I presume 1984? G3 and E4 seem like the codes TI use for package styles, but they could also be part of the date code).
That appeared to be factory soldered. Not sure why that didn't warrant a
socket? Maybe they thought it was unlikely to fail (which is very much not the
case).
What is it with the MOS 40 pin chips of this era, they all have this same
damage near pin 1 where it looks like the clear coat has flaked off or something?
Also note the option ROM chip with the markings painted over.
That was the protection chip for VisiCalc (or possibly some similar
packages). If I remember correctly, it is a Commodore printer ROM
chip, and all they do is check for certain bytes being present in the right places
before launching the software. The black paint was I presume to stop people realising that.
I did get one PET which had the original printer ROM without the black
paint, and another where it looks like someone sanded off part of the
black paint and then covered it back over with sharpie.
Adverts
All of the PET repair parts I used are available from my Tindie store:
PET ROM/RAM:
PET Diagnostics:
PET Dual Userport Joystick:
I have also listed the LED tester I made for the userport:
Mini PET 40/80 Internal
Or if you decide you want to take the easier route, I have recently built some
more Mini PET 40/80 Internal boards, that are a drop in replacement for boards
like this one.
Shipping
I can still ship worldwide from my Tindie store and US orders are now
tariffs paid.
Currently it looks like Royal Mail to the US is working. I pay the 10% tariffs
as part of the postage to (hopefully) avoid delays and additional costs at
customs. The increased US postage costs cover this.
Patreon
You can support me via Patreon, and get access to advance previews of
development logs on new projects like the Mini PET II and Mini VIC and other
behind the scenes updates.This also includes access to my Patreon only Discord
server for even more regular updates.
