• Please review our updated Terms and Rules here

Commodore PET 2001 RAM adapter?

D

Dubis7

Experienced Member
Joined
Mar 20, 2018
Messages
105
I've got an early Commodore PET 2001 that needs a few parts. The Ram/Rom had gone bad, and I replaced those with one of the Rom/Ram replacement boards from Tynemouth Software (https://www.tindie.com/products/tynemouthsw/romram-replacement-board-for-commodore-pet/) which got it booting.

However, it still has an issue. The letters appear and react to key presses, but they flicker and change what characters they are constantly. It looks like my video RAM is bad and needs to be replaced.

I know a RAM adapter was made previously by D'Asaro Designs (http://www.dasarodesigns.com/product/mps-6550-commodore-pet-2001-ram-adapter/), but they appear to no longer be manufacturing them.

Does an adapter still exist? Or perhaps specs to manufacture one? I have to believe there's some solution to deal with bad video RAM that I just haven't run across.

If there is still an adapter for individual RAM and ROM chips out there, I'd definitely be interested. I'm fine with using the replacement motherboard, but if a product exists, I'd be even happier to replace all the chips individually.
 
It may be the video RAM, it may not be.

It may be the character generator ROM. It may be other things.

Any chance of posting a video of your problems?

You can also download a copy of my PETTESTER and see if you can get it burnt into a 2716 EPROM. This replaces the EDIT ROM in your machine. It will test the video RAM, ROM and DRAM for you.

Dave
 
Dubis7 said:
I've got an early Commodore PET 2001 that needs a few parts. The Ram/Rom had gone bad, and I replaced those with one of the Rom/Ram replacement boards from Tynemouth Software (https://www.tindie.com/products/tynemouthsw/romram-replacement-board-for-commodore-pet/) which got it booting.

However, it still has an issue. The letters appear and react to key presses, but they flicker and change what characters they are constantly. It looks like my video RAM is bad and needs to be replaced.

I know a RAM adapter was made previously by D'Asaro Designs (http://www.dasarodesigns.com/product/mps-6550-commodore-pet-2001-ram-adapter/), but they appear to no longer be manufacturing them.

Does an adapter still exist? Or perhaps specs to manufacture one? I have to believe there's some solution to deal with bad video RAM that I just haven't run across.

If there is still an adapter for individual RAM and ROM chips out there, I'd definitely be interested. I'm fine with using the replacement motherboard, but if a product exists, I'd be even happier to replace all the chips individually.
Click to expand...

I've repaired quite a lot of PETs (many 2001 with 6550/6540), it may or may not be the video RAMs, have you tried to swap them? Have you tried to use some other chips? I my experience it isn't usual that ALL the 6550 go bad. In my own 2001 for example, 50% of the
chips were bad, but the other 50% is working fine.
There's an adapter for sale here:
https://www.arcadeshopper.com/wp/?page_id=11#!/PET/c/28756047/offset=0&sort=nameAsc

That's designed by me. If you want to make your own (4 x SMD small chips under the socket), I can provide the gerber files to order your PCBs.
It uses 2114 RAM chip to replace a 6550 chip.

Frank IZ8DWF
 
I grabbed a video to show what I'm talking about. https://youtu.be/IQ6ptCQrzbM

I think my character ROM is fine because you can see the characters being generated, but I could be mistaken.

I realize I was mistaken. I thought all the RAM chips in the early PETs were 6550, but if I'm understanding correctly the video RAM is a standard 2114 and doesn't need an adapter?

Regardless, I'm definitely interested in picking up some standalone adapters for my main RAM and ROM chips, and it looks like you have both. Do you also sell the EProms to go in the ROM adapters, or will I need to find someone to burn a new set?
 
Dubis7 said:
I grabbed a video to show what I'm talking about. https://youtu.be/IQ6ptCQrzbM

I think my character ROM is fine because you can see the characters being generated, but I could be mistaken.

I realize I was mistaken. I thought all the RAM chips in the early PETs were 6550, but if I'm understanding correctly the video RAM is a standard 2114 and doesn't need an adapter?

Regardless, I'm definitely interested in picking up some standalone adapters for my main RAM and ROM chips, and it looks like you have both. Do you also sell the EProms to go in the ROM adapters, or will I need to find someone to burn a new set?
Click to expand...

First of all: from the video, I think that defect might be really either the ROM, the VRAMs or any associated logic and even a combination
of more than one fault. Sure it looks like the Char ROM addresses aren't stable, but that might be caused by bad RAMs (you should
really try to swap the two VRAM's positions and see if the flickering chars change shape) or might be caused by unstable VRAM addresses
(the '157 multiplexers could be bad, or dirty/bad sockets or bad solder joints) or even bad/floating address decoding inside the char ROM.
There's always need for good troubleshooting to avoid just blindly swapping parts.
I've showed quite a few PET repairs on my youtube channel:
https://www.youtube.com/iz8dwf
If you have a 2001 revision with 6550 RAMs, then also the VRAMs are 6550, if you have a 2114 based 2001 PCB, then also the VRAMs are 2114. I've never seen any hybrid board (Though I can't exclude that they actually exist).
The adapters for sale on that site are my own design, but I'm not the owner of that site. So I'd need to contact him to see if he could program the EPROMs. I live in Italy and of course I could do adapters and EPROMs but shipping overseas would cost much more than
typical shipping inside the same continent.

HTH
Frank IZ8DWF
 
First thing. Is your character generator ROM in a socket? If so, can you pull it and tell us what you see on the screen (or post another video) please.

Second thing. Are your video RAMs in sockets?

Dave
 
daver2 said:
First thing. Is your character generator ROM in a socket? If so, can you pull it and tell us what you see on the screen (or post another video) please.

Second thing. Are your video RAMs in sockets?

Dave
Click to expand...

All 2001 boards I've seen have all RAMs and ROMs socketed. If he removes the char ROM we'll see the solid screen I'm sure, since characters on some screen position (or some characters out of the ROM) are good, so all logic *after* the ROM output is already working good.
If he removes the VRAMs we might see flickering stuff (that points to some bad ROM or bad socket/solders) or still some solid screen and that means bad VRAMs and/or bad addressing logic to the VRAMs.
I've already asked a few times if he could simply swap positions to the two VRAMs.

Frank IZ8DWF
 
I've created another video to try and answer all your questions. I show my display with and without the video ROM chip. https://youtu.be/U5ikSsLh_bI

This isn't a hybrid board. It has 6550 RAMS throughout. I got turned around before because everyone kept mentioning 2114 chips, but if my understanding is correct I still need that adapter board to use the 2114 chips in my unit. I've ordered two, one for each VRAM. At the very least I'll more or less future proof it that way.

I did try swapping some of my standard RAM chips in for the VRAM chips. That didn't appear to have any impact, so perhaps the VROM is my issue? If so, where would could I get a new EPROM burned? I know I'll need to grab an adapter board for that as well.
 
I have adapter boards that convert 6550 to 2114 i think they are dated 1980. Theae are not for individual ics but for yhe primary ram banks . Pm me if interested. These are the adapter boards http://mikenaberezny.com/hardware/pet-cbm/ods-ph-001-2114-ram-adapter/

iZ8dwf. I have a 1977 6550 2001 board and a 1978 2001 2114 board. There are NO boards that mix the two types of ram.

Dubis my 2001 which had 6550 ram had bad video ram. Tynemouth sold me a ram board for my video ram.
 
Last edited:
Dubis7 said:
I've created another video to try and answer all your questions. I show my display with and without the video ROM chip. https://youtu.be/U5ikSsLh_bI

This isn't a hybrid board. It has 6550 RAMS throughout. I got turned around before because everyone kept mentioning 2114 chips, but if my understanding is correct I still need that adapter board to use the 2114 chips in my unit. I've ordered two, one for each VRAM. At the very least I'll more or less future proof it that way.

I did try swapping some of my standard RAM chips in for the VRAM chips. That didn't appear to have any impact, so perhaps the VROM is my issue? If so, where would could I get a new EPROM burned? I know I'll need to grab an adapter board for that as well.
Click to expand...

I don't recall what size the 6550 RAMs are. When you swap them around, pay attention to how the flakiness looks exactly the same of differs. If moving them around doesn't change the address, you may be looking at a slight difference. Thinks like ti was change some letter from as A to B. but after moving is around it was changing F to Z ( no real values here ). It is still is a difference. The point is, it may have changed. What the letters change from to is a hint as to what may be the problem. Don't expect a bad RAM in location A moved to location B to fix the problem. What is important is exactly how the problem changed or remained exactly ( I really means exactly ) the same.
Dwight
 
Dubis7 said:
I've created another video to try and answer all your questions. I show my display with and without the video ROM chip. https://youtu.be/U5ikSsLh_bI

This isn't a hybrid board. It has 6550 RAMS throughout. I got turned around before because everyone kept mentioning 2114 chips, but if my understanding is correct I still need that adapter board to use the 2114 chips in my unit. I've ordered two, one for each VRAM. At the very least I'll more or less future proof it that way.

I did try swapping some of my standard RAM chips in for the VRAM chips. That didn't appear to have any impact, so perhaps the VROM is my issue? If so, where would could I get a new EPROM burned? I know I'll need to grab an adapter board for that as well.
Click to expand...

One question: when there's no char ROM in the socket, is the flashing cursor solid alternating white and black? From the video it seems having stripes instead? As I had anticipated, the no ROM situation is a solid filled screen.
It would be much more interesting to see what happens with no VRAMs. The fact that you don't seen any change when trying to swap the VRAMs with other 6550s is another indication that the fault isn't probably related to the 6550 themselves.
I could burn and send EPROMs, but that would mean shipping from Italy to US (I assume from your accent) so not cheap.

Frank IZ8DWF
 
I grabbed two more videos. One shows the display with the Character ROM in and no VRAM, and then again when I remove the Character ROM and VRAM. The second gets a closer look at the cursor. It's flashing on and off, but instead of being one white box it's black and white stripes.

https://youtu.be/__ARxKbGNW8

https://youtu.be/VacPKXvzNPc

I may be interested in a new Character ROM, if that's what this is pointing to, but I'll also see if there's anyone in the states who has the equipment to burn one. I've got some RAM adapters I'm planning to use for the VRAM, so that should rule those out as a source of the problem once I get those installed.
 
Dubis7 said:
I grabbed two more videos. One shows the display with the Character ROM in and no VRAM, and then again when I remove the Character ROM and VRAM. The second gets a closer look at the cursor. It's flashing on and off, but instead of being one white box it's black and white stripes.

https://youtu.be/__ARxKbGNW8

https://youtu.be/VacPKXvzNPc

I may be interested in a new Character ROM, if that's what this is pointing to, but I'll also see if there's anyone in the states who has the equipment to burn one. I've got some RAM adapters I'm planning to use for the VRAM, so that should rule those out as a source of the problem once I get those installed.
Click to expand...

Hm... really difficult to tell without looking at the logic signals at this point. I'll try on my 2001 to see if the no-char-rom cursor is solid (as I think it should be). I believe the VRAM isn't the problem, but as I said, it's really difficult to guess without probing signals around.
It seems a composite fault (more than one faulty chip). Even a simple logic probe could shed some light on the fault(s).

Frank IZ8DWF
 
Lets go over a few things. First is that the video address counters are working, as the characters are going to the right places. They also don't seem to be overwritten, as they do hold the values written from the CPU but they are flickering to alternate values. That lets me think that the RAM is likely good.
It looks like it is flickering between characters and graphics.
That is a signal from the 6522. If it didn't have a good select, it might be outputting alternate graphics on/off. I wonder if the 6522 can be removed and it would still boot? The graphics signal comes from that chip.
The other thought is that it might be selecting data from the main bus. One could try momentarily shorting C2-13 to ground. That would block select 8 from the address decoder. If the decoder chip had a fault it might be selecting the video memory periodically. I think this is less likely as it might also overwrite some of the data. The fact that the letters are in the right place on the screen means the character counters are working through the 74LS157s but the address could be jumping around. if the selects from the processor was not held off.
One other possibility is that it is doing reverse video but I don't think so. If one looks at the Y in ready. It looks like it is missing parts of the Y and not reverse videoing it.
My thinking is it is randomly selecting reverse video. One could carefully bend pin 39 of the 6522 out and then ground pin 18 of the video ROM to ground to see if the flickering stops. ( I don't like grounding signals from these chips but most TTL is current limited and can handle sorts to ground ). This is the graphics select.
My thoughts.
Dwight
 
What I am seeing in your first video is correct. With the two (2) video rams removed (C3 and C4), the default data outputs (well, actually the inputs to the character generator A2) will all be HIGH - thus selecting the character corresponding to all '1's.

Removing the character generator could give a totally blank screen or a totally white screen depending upon the PET.

With the character generator IN - and the two VIDEO RAMS removed - it should be possible to connect the data outputs of the RAM sockets (pins 13, 14, 15 and 16) to GND/0V (pin 22). I would use some low-valued resistors if you have them (say 100R). If you do one pin at a time, you should see various characters displayed on the screen corresponding to the link you install. It doesn't really matter what character is actually displayed, only that it changes depending upon the video ram data pin you have connected to GND/0V, and it is stable.

If this is OK, the next thing would be to check the video RAM itself by substitution.

Also known to fail are the data bus buffers B3 and B4.

Hi Dwight!

I like Dwight's idea of connecting C2 pin 13 to GND/0V though to see if the video RAM is being selected from the main CPU when it shouldn't be.

Dave
 
I'm also strongly thinking it is flicking between text and graphics. The more I look at that first video the more it looks like that. For a while I thought maybe it was the video blanking but even place where there should be a space, I'm seeing pixels showing up. That eliminates the video blanking.

On grounding C2-13, do let it boot first to load the RAM. It looks to be loading the RAM correctly. That is why I don't think it is a RAM fault. RAMs are usually good or always bad. The letters are still where they should be, just not being held on.
After watching the video and counting pixels, I'm thinking it to be a bad 74LS157. Watching the video again, at part way along the line that he is typing it starts jumping around on the screen. Counting pixels, that is around address D0 hex. That would be changes in the mux chip D3. If I was just going to change a chip, without diagnosing what to change, it would be D3. Also, from the last entry on the line above to the location junk started showing up, 48 pixels, on a later line. That is 110000b, address bits 4 and 5. This again implicates D3 as not working right.
Dwight
 
Dwight Elvey said:
I'm also strongly thinking it is flicking between text and graphics. The more I look at that first video the more it looks like that. For a while I thought maybe it was the video blanking but even place where there should be a space, I'm seeing pixels showing up. That eliminates the video blanking.

On grounding C2-13, do let it boot first to load the RAM. It looks to be loading the RAM correctly. That is why I don't think it is a RAM fault. RAMs are usually good or always bad. The letters are still where they should be, just not being held on.
After watching the video and counting pixels, I'm thinking it to be a bad 74LS157. Watching the video again, at part way along the line that he is typing it starts jumping around on the screen. Counting pixels, that is around address D0 hex. That would be changes in the mux chip D3. If I was just going to change a chip, without diagnosing what to change, it would be D3. Also, from the last entry on the line above to the location junk started showing up, 48 pixels, on a later line. That is 110000b, address bits 4 and 5. This again implicates D3 as not working right.
Dwight
Click to expand...

I've seen more than once some 6550 RAMs fail in such a way that they still output the "old" value when the address changes from (for example) xx0111 to xx1000 (multiple 1 to 0 transitions simultaneously).
I have also showed that on a few videos.
Also, yes, the PET will boot fine without the 6522.
The fault isn't easy to diagnose because some characters on some screen position appear to be the right ones and not flickering at all.
The idea to drive the char ROM with resistors is good, but I've also seen a couple of 6540 ROMs get "slow" in selecting the right byte on multiple '1' to '0' address transistions, so a static address might cause the ROM to work as espected and fail when address "wiggle" as they should.
I've made a microcontroller based test device that can identify all these (and other) strange 6550 and 6540 failure modes.

Frank
 
My RAM adapter boards came in today. I've replaced both VRAM chips with the adapters and a set of 2114 RAM chips. No change, so now I've confirmed that the RAM chips themselves aren't the issue.

I removed the 6522 and tried booting. My video signal went away entirely. It sounds like it booted successfully, as the tape drive whirred and stopped like it usually does, but I'm not seeing anything on the display.

That being said, I'm going to look into Dwight's suggestions about the D3 74LS157. I've also ordered a ROM adapter board and am planning to replace the character ROM with a new EPROM as a matter of future proofing, even if that's not my actual issue.
 
Oh, and is there a resource out there for identifying which chips are which so I can compare "D3" to the physical board I'm looking at? I'm going to look for one myself but if anyone has a reference document that I'm not finding I'd appreciate being able to look it over.
 
iz8dwf said:
I've seen more than once some 6550 RAMs fail in such a way that they still output the "old" value when the address changes from (for example) xx0111 to xx1000 (multiple 1 to 0 transitions simultaneously).
I have also showed that on a few videos.
Also, yes, the PET will boot fine without the 6522.
The fault isn't easy to diagnose because some characters on some screen position appear to be the right ones and not flickering at all.
The idea to drive the char ROM with resistors is good, but I've also seen a couple of 6540 ROMs get "slow" in selecting the right byte on multiple '1' to '0' address transistions, so a static address might cause the ROM to work as espected and fail when address "wiggle" as they should.
I've made a microcontroller based test device that can identify all these (and other) strange 6550 and 6540 failure modes.

Frank
Click to expand...

For the 6550s, swapping the RAMs should change the behavior in how the characters change if one of the RAM output data was changing. It will still be flaky but in different ways. I do wish he'd posted a video with the RAMs swapped. What is the same for one person might be different when looking for differences between the two displayed outputs.
Assuming there was no difference and looking at the symptoms, it is clearly an address problem and not the ROM as the RAM isolates the video ROM from RAM addresses. Again, this can be seen in the video were it jumps by 48 locations. The Video ROM has no ability to jump a screen address. The symptoms point to an addressing problem, either in a RAM or the 74LS157s.
The video ROM was eliminated with the 48 character shift. ( assuming only one problem sourse )
At this point, another couple of videos, with just one single RAM removed, would isolate a RAM issue. That would clear up the possibility that a RAM input address pin is causing the issue. This would assume that both RAMs didn't have exactly the same intermittent input address pin. Using C4 socket would be best to look at as it would be mostly printed characters. This would eliminate any single RAM input address issue. ( Assuming both RAMs didn't have the same input address intermittent. ) If one was not stable and the other was, it would be a RAM input issue.
If you get this far and the problem still remains it leaves the address lines from the 74LS157s.
Getting this far, we have eliminated the ROM and the RAM. That only leaves one of the 74LS157s.
D3 is the only one that effects addresses on 48 character boundaries.
I hope that is methodical enough for you, Frank?
Dwight
 
You must log in or register to reply here.
Back
Top