ColecoVision diagnostics shield

[ChildOfCv]

Yeah, for the PAL SCART board I only had a chance to do the video part in the schematic.  I don't have an actual board to measure placements, so I couldn't update the PCB.  I was using another user's pictures to trace the connections, and not all components had their references visible.  I don't know what happened to him though, I haven't noticed any more posts in a while.  Hopefully nothing bad.

It's likely that what's left outside of the video part is similarly numbered though, but of course, the component layout is shifted.  So while I'd hate to throw away the PCB part entirely, I admit that it's not a good reference.

 

[omf]

12 hours ago, ChildOfCv said:

Yeah, for the PAL SCART board I only had a chance to do the video part in the schematic.  I don't have an actual board to measure placements, so I couldn't update the PCB.  I was using another user's pictures to trace the connections, and not all components had their references visible.  I don't know what happened to him though, I haven't noticed any more posts in a while.  Hopefully nothing bad.

It's likely that what's left outside of the video part is similarly numbered though, but of course, the component layout is shifted.  So while I'd hate to throw away the PCB part entirely, I admit that it's not a good reference.

 

i may be able to help with that if interested, i do have a pal / pal scart / ntsc board so i can provide reference, i also dont mind removing the odd component if the wording is under a chip or whatnot to find out the correct references

 

[Zak]

On 8/26/2020 at 5:49 AM, ChildOfCv said:

It might be.  I have checksums for about 5 variations of ROM, including aftermarket patches, but... like Pokemon, gotta collect em all!

 

ok, there are the Colecovision "Kenia Fueguina" (Argentina)  BIOS.   Tnks to Alejandro Smukler @Alesmukler who took the trouble to unsolder the BIOS and assemble the necessary wiring to be able to read it. 

 

COLECO_KENIA_ROM.ZIP

[Ikrananka]

2 hours ago, Zak said:

 

ok, there are the Colecovision "Kenia Fueguina" (Argentina)  BIOS.   Tnks to Alejandro Smukler @Alesmukler who took the trouble to unsolder the BIOS and assemble the necessary wiring to be able to read it. 

I can confirm that the Colecovision "Kenia Fueguina" (Argentina) BIOS is 100% identical to the UK/European CBS ColecoVision PAL/SECAM BIOS.

 

[ChildOfCv]

On 8/29/2020 at 5:13 AM, omf said:

i may be able to help with that if interested, i do have a pal / pal scart / ntsc board so i can provide reference, i also dont mind removing the odd component if the wording is under a chip or whatnot to find out the correct references

That would be useful.  At the moment I'm working on the PAL board though, so I'm not ready to take on another project *yet*.  I'll try to finish that up quickly.

[ChildOfCv]

On 8/30/2020 at 11:55 AM, ChildOfCv said:

That would be useful.  At the moment I'm working on the PAL board though, so I'm not ready to take on another project *yet*.  I'll try to finish that up quickly.

Here's what @Mister Fab was working on.  Most of the component references are readily visible, but there are a few that will need to be peeked at.  For the image he was working on, he was using Paint.NET, which seems to be a fairly competent graphics platform, though with limited file format support.

 

Anyway, if you are interested in finishing the work, that would be helpful.  Also, getting an exact measurement of the PCB dimensions would help.

[omf]

16 hours ago, ChildOfCv said:

Here's what @Mister Fab was working on.  Most of the component references are readily visible, but there are a few that will need to be peeked at.  For the image he was working on, he was using Paint.NET, which seems to be a fairly competent graphics platform, though with limited file format support.

 

Anyway, if you are interested in finishing the work, that would be helpful.  Also, getting an exact measurement of the PCB dimensions would help.

im not too sure what you need here, are we consentrting on the yellow outlines or the non yellowd components?

i can see an issue with the picture already however,  some of the transistors near the rgb out socket have verticle resistors on some of the transistor base pins,  this is not highlighted  as having anything so that will need documenting.

 

i wont have access to my scart colecovision for 1.5 weeks but after that i should be able to do something if i know what is needed

 

[ChildOfCv]

2 hours ago, omf said:

im not too sure what you need here, are we consentrting on the yellow outlines or the non yellowd components?

Every component needs a designator, such as R158, so that it can be matched with the component on the board and on the schematic.  It may be hard to notice at first, but there are several components that do not have their designations "in the open" where they are easy to read.  For instance, you see one diagonal resistor, R158.  Just below it, there is a capacitor straddling another resistor.  Neither designator is visible there.  What are those two?  At the top, there is a row of resistors and one diode.  R174, R170, R175, R172, and then a diode whose designator seems to be printed underneath.

 

Those are the first problems, because it's much easier to position the components when you can find them by the designator.

[omf]

ok ill look for those when im able.

 

in the meantime i sourced a cartridge connector 3d model, unfrtunately its not the right colour's, but i guess you cant have everything

ENG_CVM_CVM_5530841-2_B.3d_stp.zip

 

datasheet:  https://www.te.com/commerce/DocumentDelivery/DDEController?Action=srchrtrv&DocNm=5530841&DocType=Customer+Drawing&DocLang=English

 

mfg's part number: 5530841-2

[ChildOfCv]

Update:  Currently building 5 diag board sets.

 

I started by soldering on the Atmel, oscillator, ICSP components, FTDI, and power filtering components.  Then made sure it would accept USB connection and program through the Atmel programmer.  Loaded up boot loader and a simple diag program that checks for shorts between adjacent pins.  So far so good.

 

Next is to add the remaining components, load in the Coleco diag program, and make sure it works on my CV.

[omf]

On 9/2/2020 at 12:27 AM, ChildOfCv said:

there is a row of resistors and one diode.  R174, R170, R175, R172, and then a diode whose designator seems to be printed underneath.

VR3

 

 

On 9/2/2020 at 12:27 AM, ChildOfCv said:

you see one diagonal resistor, R158.  Just below it, there is a capacitor straddling another resistor.  Neither designator is visible there.  What are those two?

there appears to be no designator for these along with the resistors attached to the base pin of the transistors

 

 

ill look for some more later

 

[omf]

c-5530841-2-b-3d.stp

On 9/2/2020 at 11:13 AM, omf said:

 

in the meantime i sourced a cartridge connector 3d model, unfrtunately its not the right colour's, but i guess you cant have everything

 

c-5530841-2-b-3d.stp

[ChildOfCv]

On 9/26/2020 at 7:11 AM, omf said:

c-5530841-2-b-3d.stpUnavailable

 

c-5530841-2-b-3d.stp 427.58 kB · 3 downloads

Hmm... For some reason it called this reply up as if it were posted again yesterday.

Well, anyway, I now have 5 Diag main PCBs that seem to work.  Now to finish some of the cartridge port test components.

In other news, I created 3D drawings of the DE9 connectors and the epoxy-coated inductors.

[ChildOfCv]

The latest version of the schematic set is now posted.  This includes the ColecoVision PAL D board, and the latest version of the diagnostics boards.  Remember that the SCART version is only expected to be accurate for the video section.

Coleco.zip

[Falonn]

Nice work!  I'm surprised at how similar everything in the PAL version--except the A/V subsystem--is to the NTSC model.  (Although, in retrospect, I suppose I shouldn't be.  Reusing those other blocks is just good engineering on their part.)

 

There is a lot more to take in on the PAL A/V page though, sheesh!  It's sort of a combination of the NTSC's A/V stuff, the RF board, and another clock circuit all rolled into one.  I can't imagine how long it must have taken to trace all of that out.  Thanks again for your generosity to the community!  This is a great resource.

[Zak]

after a long time I was finally able to get down to business and test the amazing diagnostic PCB that ChildOfCv developed.

 

 

I was testing the same with two Colecovision, one that works OK to see the successful tests and then I started working with a CV that I have to repair.

In principle, the CV showed a black screen and a constant tone in the sound.

I ran test 1 of voltages: OK

then the test battery No. 2. there it threw the following error, it seems that there were problems with the U5 decoder:

 

 

I replaced it, but the black screen continues. Then I ran the tests again and the failures of test 2 no longer appear, however I have doubts about the meaning of  "PIN TEST".  What are low and high "stuck pin test" ?

 

In Faulty CV  only apears D1 and D7 lines,  but in Working ok CV  apears  D0 D1 D5 and D7  lines.  

what does it mean that only D1 and D7 appear?

 

then, when running the video test it throws the following error:

 

 

 

up to here I come. any idea where to go next? bad video IC ?

 

tnks!

 

cc @Alesmukler

 

[Ruggers Customs]

38 minutes ago, Zak said:

after a long time I was finally able to get down to business and test the amazing diagnostic PCB that ChildOfCv developed.

 

 

I was testing the same with two Colecovision, one that works OK to see the successful tests and then I started working with a CV that I have to repair.

In principle, the CV showed a black screen and a constant tone in the sound.

I ran test 1 of voltages: OK

then the test battery No. 2. there it threw the following error, it seems that there were problems with the U5 decoder:

 

 

I replaced it, but the black screen continues. Then I ran the tests again and the failures of test 2 no longer appear, however I have doubts about the meaning of  "PIN TEST".  What are low and high "stuck pin test" ?

 

 

In Faulty CV  only apears D1 and D7 lines,  but in Working ok CV  apears  D0 D1 D5 and D7  lines.  

what does it mean that only D1 and D7 appear?

 

then, when running the video test it throws the following error:

 

 

 

up to here I come. any idea where to go next? bad video IC ?

 

tnks!

 

cc @Alesmukler

 

Black screen with a tone is normally either a bad VDP or CPU.

[Zak]

1 hour ago, Ruggers Customs said:

Black screen with a tone is normally either a bad VDP or CPU.

apparently the CPU, ROM and RAM is ok

 

 

 

[+-^CrossBow^-]

I actually was running into similar results with my Diag Shield this past Saturday when using it on a working CV. First the +5 was reading out of spec but my meter was showing it to be about 4.9v so that was odd. And the CPU test would frequently fail with the diag shield telling me the CPU wasn't at the expected address? 

 

Additionally, it seems the shield may not be able to figure out all CV variants too well. This CV seems to be pretty new as it only had a very tiny small blob of solder near the expansion board for the top RF shield instead of the braided ground strap I would normally see. Near as I could tell, I was the first person inside this console in about 20 years. Anyway, the rom test would always come up with unknown on that test.

 

After cleaning the cartridge slot and expansion connector along with the PCB and connector on the Diag shield parts, some of the tests ran a little better such as the CPU. RAM test would also nearly always fail with U4 and U5 not responding? But again, this CV works.

 

In the end I think I might have found the issue. It seems the connectors that the ribbon cable between the two shield boards is where the fault lies. I sprtized some alcohol into the connectors on the PCBs and reattached the ribbon cable and then started to get consistent results with the shield. The voltages were now reading what I was reading from the inductor with my meter, the CPU test would still occasional fail to start unless I ran the VDP tests first and then ran the CPU test? Ram test started working properly as well.

 

So I still think it is an excellent diagnostic kit to have on hand but I might remove the connectors and just solder wiring straight to the PCBs to see if it helps in that regard. I'm assuming that if left unused over time, the pins start to oxidize a bit and that would skew the readings of course.

 

[ChildOfCv]

Some of that could be the connection to the system.  I noticed similar problems while developing due to residue of flux getting on everything.  Before connecting to the expansion port, clean it as well as you can.  You may even want to go over the card edge contacts on the diag board to make sure they remain uncontaminated.  Contact cleaner may help, but make sure it has no alcohol or other conductive stuff in it or else you have to clean it up after use too.  DeoxIT cleaners claim to have no alcohol and claim to be nonconductive.  But anyway, if it can't make a good connection, it's not driving the system.  It's only making suggestions.

 

To answer the question about what the stuck pins test does, it uses a 10K resistor to pull the pins in question high, then tests if there is anything fighting against that.  Then it repeats the test while pulling everything low.  It could be that 10K isn't aggressive enough, but it seems to me that if no chip is attempting to drive the bus, it ought to be enough.  However, it is possible that some chips are "leaky" but not enough to keep the system from failing outright.  Generally if one chip is only sorta shut down, another chip that wants to drive the bus can do so, but at the cost of increased power.  I've wavered between whether or not to use a smaller resistance to do the tests with, or to say that maybe there is a chip that's on the verge of failing, or is at least not within spec, and it really should be tracked down and replaced.  On the other hand, it's a game console.  What do you expect?  

 

Still, since Zak's system seems to have a bad VDP, it's possible that it is the culprit in the stuck pins too.

[ChildOfCv]

Also, note that the text stated that if one of the select lines (x000) are implicated, then you should look at U5.  I guess it mentioned them while activating them specifically, but the only thing implicated in those tests was the shield line.  I guess that's to be expected though when you've removed the RF shielding, and therefore it has no conductivity to ground.  Maybe I should make a note about that too.

 

Hmmm, I just realized that it's the lower half of the shield that's supposed to connect the cartridge shield ground.  So removing the upper part wouldn't be a problem.  If someone removed both halves though, you have a problem and need to solder a wire from the shield connection to ground.

 

I'll also make an attempt to adjust the code so that it doesn't print the troublesome message unless one of the select lines failed to activate.

Edited July 11, 2021 by ChildOfCv

[Ruggers Customs]

Yep can confirm.  If the bottom half of the shield is gone, it will tell you to check U5.  I just solder a wire from the center screw hole to ground so shield 13 is always grounded.  

 

If memory serves me correctly, I think that was also the fix for the Risky Rick game?

[NIAD]

39 minutes ago, Ruggers Customs said:

If memory serves me correctly, I think that was also the fix for the Risky Rick game?

That was the fix for Penguin Adventure by Opcode as well.

[ChildOfCv]

I did some analysis of leakage current for chips that are on the bus but not sending.  Resistance of a transistor is, of course, not constant.  But I used datasheet values to estimate a resistance range to see what they all end up doing together.

 

Input ICs:  135K each

SRAM: 400K (there are two, of course, but they are divided between the data lines so a single data line only considers one of them)

CPU: 500K

ROM: 240K

VDP: 52.5K (yikes)

 

Parallel resistance:  Around 23.5K.  So that does put up a decent fight with the 10K resistors.  Perhaps they should be reduced to 2K.  If you have the solder skills and confidence, try replacing R12-R15 and R22-R25 with 2K resistors.

[ChildOfCv]

13 hours ago, -^CrossBow^- said:

So I still think it is an excellent diagnostic kit to have on hand but I might remove the connectors and just solder wiring straight to the PCBs to see if it helps in that regard. I'm assuming that if left unused over time, the pins start to oxidize a bit and that would skew the readings of course.

That's bad news.  I wonder if it's the ribbon cable connectors or that the pins might be too skinny for spec.

 

Well, one alternative might be to use dupont connectors.  They'll definitely fit those pins and sockets.  Just make sure to match up pin 1 etc