New Colecovision schematics

[ChildOfCv]

Hi all,

 

First post. Hopefully it's a good one.

 

Searching high and low for CV schematics and finding only the same set everywhere, and noting that it's lacking in a few places (and in PDF format instead of a schematic file), I set out to engineer a new schematic myself. I still can't guarantee its total accuracy, of course. But I hope it is.

 

I used Kicad 5 for the project, which is also my first time using the package. So some things may not have been done optimally. Constructive criticism is welcome.

 

.zip'd project archive: https://drive.google.com/open?id=1lVH7_HhNAEBQ5yUY1j8FsJktdYjzmgUc

 

Last I checked, Google Drive thought it couldn't display the preview, but the download button should still work.

 

[Mr. Bildo]

This is great. This will be a great resource for people working on the system.

 

How annoying are the power flags in KiCad?

[ChildOfCv]

This is great. This will be a great resource for people working on the system.

 

How annoying are the power flags in KiCad?

Very. They don't even survive past passives like inductors. Yes, Kicad: inductors conduct.

[Keatah]

PDF is a good format because everyone has the capability to read it.

[mr_me]

PDF is a fine publish format but it's not good for someone who wants to edit the document.

[ChildOfCv]

PDF is a fine publish format but it's not good for someone who wants to edit the document.

I guess I can compromise and offer the PDFs within the .zip archive though. As you say, it's much easier to make the corrections as they come rolling in, if it's still in its generating format. But people who don't want to download KC may also want to see the end result.

 

I'll publish a new version of the project this evening. The new version also has footprints assigned (some of which are just placeholders though) and the beginnings of PCB placement. That should help people who want to know where that @#$#@ R26 is on the board, too.

Edited December 5, 2018 by ChildOfCv

[ChildOfCv]

New version uploaded. PDF file included. Even preview works now.

[ChildOfCv]

And yet another update. This time the PDF is in color. There is also a PDF picture of the PCB.

 

Unfortunately, it appears that Kicad converts the text into vector graphics, so I guess the files aren't searchable. Haven't found a solution to that yet, other than maybe to feed the PDF through OCR.

[ChildOfCv]

Fixed. The schematic and silkscreen PDFs are now searchable.

[Mr. Bildo]

Unfortunately, it appears that Kicad converts the text into vector graphics, so I guess the files aren't searchable. Haven't found a solution to that yet, other than maybe to feed the PDF through OCR.

 

If you use Acrobat (and maybe other programs, not sure) it does a pretty good job of automatic OCR. I've had good luck even with stuff like old service manuals. CTRL+F and it digs through it for you.

[matthew180]

Nice! KiCAD is my EDA software of choice, so it is nice to see this in that format. I have not had a chance to look at the files yet, but I was wondering if you reworked the schematics from scratch by reversing actual hardware, of if this was an effort to capture the existing schematics floating around out there? IIRC those older schematics have errors in them. At lease now corrections can be made and better schematics shared!

[ChildOfCv]

Nice! KiCAD is my EDA software of choice, so it is nice to see this in that format. I have not had a chance to look at the files yet, but I was wondering if you reworked the schematics from scratch by reversing actual hardware, of if this was an effort to capture the existing schematics floating around out there? IIRC those older schematics have errors in them. At lease now corrections can be made and better schematics shared!

 

I retraced my own board from scratch. Mine is rev J, so it's also possible that it has differences anyway. It would be nice to see a change log so we know what to expect between boards.

 

There were a number of omissions in the other schematics, also some mismatched labels and missing component values. Most critically, the Q' -> D connection on the initial clock divider. There is one other difference in the clock circuit, where the 3.58MHz bumps the 10.xxxMHz oscillator. I traced it out multiple times to verify, and keep getting the same result. Of course it could just be a confirmation bias brain fart too.

[Shawn]

What are the chances this turns into a run of new PCB's?

[ChildOfCv]

What are the chances this turns into a run of new PCB's?

 

It's tempting. The main thing left is to run the traces. The latest version has most of the parts placed, the board outline, and even the mounting holes in place. The joystick boards haven't been touched yet, but it at least seems possible.

 

Of course I also noted that others such as the CollectorVision people have already created a replacement PCB. They may even have the Gerber files in their archive.

[Shawn]

 

It's tempting. The main thing left is to run the traces. The latest version has most of the parts placed, the board outline, and even the mounting holes in place. The joystick boards haven't been touched yet, but it at least seems possible.

 

Of course I also noted that others such as the CollectorVision people have already created a replacement PCB. They may even have the Gerber files in their archive.

 

 

None have yet to be offered for sale though. I think a drop in brand new PCB would be warmly welcomed as long as it was offered at a reasonable price.

[Shawn]

Even a bare PCB that the end user populates themselves.

[matthew180]

 

I retraced my own board from scratch. Mine is rev J, so it's also possible that it has differences anyway. It would be nice to see a change log so we know what to expect between boards.

 

There were a number of omissions in the other schematics, also some mismatched labels and missing component values. Most critically, the Q' -> D connection on the initial clock divider. There is one other difference in the clock circuit, where the 3.58MHz bumps the 10.xxxMHz oscillator. I traced it out multiple times to verify, and keep getting the same result. Of course it could just be a confirmation bias brain fart too.

 

That is fantastic that you re-traced the PCB, and very interesting error you mention about Q' and D on the clock divider. Something about that circuit always bugged me, and that was probably it. When I get a few minutes I'll definitely go over it, maybe a new set of eyes will help with any potential errors.

 

Oh, and KiCAD 5.0.2 was released today. :-)

[matthew180]

Even a bare PCB that the end user populates themselves.

 

Once the work of routing the board in KiCAD is done, anyone can have a PCB made quickly and easily from OHSpark (or a similar service). I'm not sure what good/advantage you get from a full size reproduction using the same original parts? An accurate schematic is nice for troubleshooting and making emulators, but due to PCB revision and less than desirable circuit design, if a new board is being designed it seems that you would want to correct some of the deficiencies of the original instead of reproduce them.

 

Also, due to the PCB size, having a new PCB made would probably cost more than buying a CV on ebay or where ever.

[Shawn]

 

Once the work of routing the board in KiCAD is done, anyone can have a PCB made quickly and easily from OHSpark (or a similar service). I'm not sure what good/advantage you get from a full size reproduction using the same original parts? An accurate schematic is nice for troubleshooting and making emulators, but due to PCB revision and less than desirable circuit design, if a new board is being designed it seems that you would want to correct some of the deficiencies of the original instead of reproduce them.

 

Also, due to the PCB size, having a new PCB made would probably cost more than buying a CV on ebay or where ever.

 

 

That makes a good point and I agree that making corrections and eliminating anything that can be made redundant from the original retail version is a great idea. I myself find the pcb's in colecovisions to be very fragile and the traces on the original boards are not exactly super high quality not have they stood the test of time. I like the idea of a new pcb that could be populated using the original "brains" of the coleco with a matching footprint of the original shell. Call it a face lift project if you will. Same old machine, brand new skeleton to wrap around the organs in the same old familiar body.

[ChildOfCv]

Newest version uploaded. There were a couple of minor fixes after going over the board again with a multimeter.

 

All components on the main PCB are placed, which should make it easier to locate the component in question.

 

The same link should work: https://drive.google.com/open?id=1lVH7_HhNAEBQ5yUY1j8FsJktdYjzmgUc

[ChildOfCv]

And one more update. PCB diagrams for the RF board, standard controller, and Super Action Controller are included. Standard controller corrected to account for the difference in pinout between the SAC keypad and the standard keypad.

[ChildOfCv]

And yet another update. Now includes roller controller schematic and PCB diagram.

[ChildOfCv]

Latest update: Now includes Expansion Module 1 (Atari-compatible console). I am now 90% sure the schematics are accurate.

[Ikrananka]

Latest update: Now includes Expansion Module 1 (Atari-compatible console). I am now 90% sure the schematics are accurate.

 

This is fantastic - thank you for all of your hard work.

[ChildOfCv]

You're welcome. Now I'm 99% sure the schematics are correct, after highlighting each net in the PCB program and testing them with the multimeter. The remaining 1% is the pins that appear to be unconnected, but could indeed be hiding traces underneath the chip. But I'm pretty sure I got all the traces.