Jump to content
jeff12345

ColecoVision Eating VRAM Chips?!

Recommended Posts

Posted (edited)

Hey everyone!

I'm currently trying to repair a ColecoVision and my Google searches make me believe that this is the forum where the experts reside :)

 

Originally the CV just produced a black screen and made a weird droning noise. I found that one of the SRAM chips and one of the DRAM chips had heavily corroded legs. I also checked all the busses with my logic analyzer. CPU-side looked normal to me, but the data bus of the VDP just remained low aside from a couple of quick bursts at the beginning. So I decided to change all the RAM chips and do the 5V mod along the way. The solder at the legs of the corroded DRAM chip unfortunately was like stone and I did not manage to get it out without destroying a couple of traces and had to add a couple of botch wires. (Overall I have to say the CV is the worst soldering experience I had so far, I constantly have to work with 400dC+ and still have a hard time getting the solder to meld). Added the new chips (SRAM: Intel SAB2114PTL, DRAM: TI TMS4164-20NL), powered on the console and... garbled graphics... But well, clear improvement over a black screen.

 

161260313_2020-07-3122_34_18.thumb.jpg.c7c0037a8ae132833df1f8a92c4b1085.jpg

 

First thing I thought is that I didn't solder properly somewhere. In the meantime, however, I checked for continuity and shorts 3 times and am fairly certain that everything is connected properly. So I assumed that maybe one of the DRAM chips was bad and replaced them one by one. Replaced the 4th, and.... SUDDENLY THE PICTURE LOOKED PERFECT! YAY!!!

 

IMG_0939.thumb.jpg.18f908c13a5f5e2888b366cf9f6f3e00.jpg

 

Until I tried to switch it on again the next day. Back to garbled graphics...

 

Coincidentally I also had a C64 with bad memory laying around, so I replaced the chips in there with sockets, added a new batch of my TMS4164 and ran a RAM tester for a couple of passes - it did not find any problems.

 

Then I moved the whole batch of tested chips into the CV. Turned it on and... everything was fine again! Until I tuned it off and on a couple of times. Back to garbled graphics... So I moved the whole batch from the CV into the C64 and reran the RAM test. And this time, one of the chips reported as bad. This makes me think that the CV kills the chips somehow.

 

Voltages at the chips are too low rather than too high for my taste: 4.5-4.7V when turned on. PSU when CV is turned off gives about 5.1V. I found someone else in this forum reporting the same voltage drop problems after doing the 5V RAM mod, but there was no resolution (I unfortunately can't find the post anymore).

 

I checked for shorts and found that I have a resistance of about 340Ohms between GND and +5V, which seemed weird to me. But according to this post, it seems to be normal.

 

I did the 5V mod following this document. I'm not really comfortable with the A7 pin of the 4164 being connected directly to VCC. I'm not super experienced with electronics, but shouldn't there be a resistor inbetween to limit the current flow when pulling that pin high?

 

Any ideas about what might be going on are highly appreciated!

Edited by jeff12345

Share this post


Link to post
Share on other sites
Posted (edited)

I'm sure others will be able to help better than I can, but I thought I'd at least let you know that the VCC at the VDP on two of my working (not VRAM modded) CVs was around 4.77V with known good power supplies, so your voltage reading definitely looks to be on the low side.

Edited by Ikrananka
  • Like 1

Share this post


Link to post
Share on other sites
Posted (edited)

I've done hundreds (literally) of these mods and have never experienced what you're describing.  After the mod, Pin 1 on the vram should be 0 volts, pins 8 and 9 should be +5 volts or around there.  Sounds like you have a broken bus line somewhere.  Here are some simple tests from the VDP to the vram chips.

https://console5.com/wiki/Colecovision_5V_RAM_Modification

 

Edited by Ruggers Customs
  • Like 2

Share this post


Link to post
Share on other sites

Thx for the quick replies! I know the Console5 site, this is the stuff I checked 3 times now. But I guess I will do pass #4 over the weekend.

 

Pin 1 is N/C, 8&9 are around 4.6V. What I don't understand though, is how a broken bus line could destroy the chips. Even if there would be a short somewhere, I don't see how this would happen.

 

What chips are you typically using, @Ruggers Customs?

 

Other things I tried but forgot to add above:

- Test voltages with everything socketed removed (VRAM, SRAM, ROM). Voltage slightly higher (~4.7V), but still rather on the low end.

- Test with lab PSU to make sure it's not the CV PSU. But basically the same.

Share this post


Link to post
Share on other sites
7 minutes ago, jeff12345 said:

Other things I tried but forgot to add above:

- Test voltages with everything socketed removed (VRAM, SRAM, ROM). Voltage slightly higher (~4.7V), but still rather on the low end.

- Test with lab PSU to make sure it's not the CV PSU. But basically the same.

That is very peculiar to lose so much voltage.  Could it be a faulty VDP that is somehow internally shorting after warming up?

Share this post


Link to post
Share on other sites

What also makes me question a broken bus line is that I twice had the effect that everything worked fine after putting in new chips. Then went bad again after turning on/off a couple of times without moving. Could of course be that the physical force when inserting chips closes a loose connection for a while, but seems like a rather big coincidence to me.

Share this post


Link to post
Share on other sites
Posted (edited)

Well if all your continuity checks are coming back okay then it's something else obviously.  Is your power switch clean?  That can cause garbled graphics, same with a loose power connection on the rear of your console.

 

If it was an issue with your ram or a bad or broken bus line then the garbled image would be the same each time you power the console on and off.  Since it's different each time, that means that it's possibly power related.

Edited by Ruggers Customs

Share this post


Link to post
Share on other sites
1 minute ago, Ikrananka said:

That is very peculiar to lose so much voltage.  Could it be a faulty VDP that is somehow internally shorting after warming up?

Maybe. I have to see if I find a cooling spray or something to test. But the first time I was bad immediately after turning it on after a night, so the VDP shouldn't have been particularly hot.

Share this post


Link to post
Share on other sites
1 minute ago, Ruggers Customs said:

Well if all your continuity checks are coming back okay then it's something else obviously.  Is your power switch clean?  That can cause garbled graphics, same with a loose power connection on the rear of your console.

I cleaned it with contact cleaner, but didn't solder it out to take it apart. I soldered a bridge across the 5V line though: gave me maybe 0.1V more on the board, but errors were still the same. Plug is good point, I didn't think of that yet. I hooked up my lab PSU to the pints in the plug, maybe I should try connecting it directly to the board.

Share this post


Link to post
Share on other sites

Regarding A7:  4164 is a 64Kb chip, but the VDP only uses 16K.  The 4116's that came with it use the A7 pin as VCC.  It's way easier to just let it be +5V than to change that.  A7 needs to be tied to some constant value--0 or +5V.  Either one works, but as I say, +5V is easier.  I guess it's okay for pin 1 to stay at -5 since it's supposedly not connected inside of the chip.

 

As for A7 (pin 9), it's fine to just tie it to a voltage.  It already has meg-ohms of resistance internally anyway.

 

BTW, did you add sockets to the CV board?

Share this post


Link to post
Share on other sites
Posted (edited)

Anyway, looking at the failure, it seems that U11 is the biggest culprit since everything that's wrong is only off by 1.  But it's also not consistent which off-by-1 it is.  For instance, "BEFORE" replaces only the first E with a D.  That seems to indicate that the Q output is floating rather than shorted.  The connection is not making it back to the VDP.

 

Oh wait.... TI, for some reason, reversed the order of the lines so that the least significant bit is D7.  It's hard to remember which ones are compensated for, so it could actually be U16 that's the problem.  And U16 also has a dedicated Q output to the 9928 which is not multiplexed with an address line.

 

Well, D0 connects to CD7.  AD7 connects to VA0.  So yeah, it's probably U11 that is suspect.  Make sure pin 2 of U11 has continuity with pin 5, and nothing else.  On all the other DRAMs, it should only connect to pin 2.  On the 9928, it should only connect to pin 3.

Edited by ChildOfCv

Share this post


Link to post
Share on other sites
17 hours ago, Ruggers Customs said:

Well if all your continuity checks are coming back okay then it's something else obviously.  Is your power switch clean?  That can cause garbled graphics, same with a loose power connection on the rear of your console.

 

If it was an issue with your ram or a bad or broken bus line then the garbled image would be the same each time you power the console on and off.  Since it's different each time, that means that it's possibly power related.

I think the part about the plug was the right hint. I bridged the switch again and connected the lab PSU directly with the board. CV start up nicely most of the time. I cleaned the plug of the original PSU thoroughly with contact cleaner and I feel like it made it better. But I can reproduce the error by fidgeting around with the ground or 12V pin of the plug.

 

IMG_0944.thumb.JPG.d84430640d27fbc3384cb198637185fe.JPGIMG_0943.thumb.JPG.025aac26e6435d4ccb097d2ff044beef.JPG

  • Like 1

Share this post


Link to post
Share on other sites

What still scares me a bit that one of the RAM chips reported as bad after I moved it back from the CV into the C64. But maybe it was just by circumstance. Really don't know what would fry these chips with <5V if you guys say having A7 directly on VCC is fine.

Share this post


Link to post
Share on other sites
Posted (edited)
42 minutes ago, jeff12345 said:

What still scares me a bit that one of the RAM chips reported as bad after I moved it back from the CV into the C64. But maybe it was just by circumstance. Really don't know what would fry these chips with <5V if you guys say having A7 directly on VCC is fine.

The only time you would ever have to worry if you fried a chip is if you accidentally forget to disable the 12 volt line feeding into Pin 8 on the ram chip. When you remove inductors L2 and L3 then you disable the -5 and +12.  Those chips have a threshold of 5.5 volts so anything above that will fry them.  Pin 9 will always be 5 volts but on the 4516s that's a non connected pin so it's perfectly fine.  So if pin 1 is 0 volts and pin 8 and 9 are +5 then you're fine.  Now there is the remote possibility that ESD from you personally can blow a chip but as long as you're being safe then everything should be fine.  :)

Edited by Ruggers Customs
  • Thanks 1

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...