CV-EAV (Enhanced Audio and Video) ColecoVision Upgrade Board

[+mytek]

I first started this discussion in another topic about RAM, but I think it's time to give it its own topic for easier search and find.

 

So what is it?

 

It's a replacement for the stock RF Modulator, that is specifically designed to output high quality, Composite, S-Video, and Audio through onboard PCB mounted jacks. If this goes as planned, there will be NO WIRES required during installation. And aside from a heavy duty (high wattage) soldering iron to remove the original modulator, this thing will pop right onto the existing Main ColecoVision PCB. And with a commonly found stepped-drill some simple mods to the rear plastic case will give you full access to all the outputs (RCA's for Composite Video and Audio -- Mini-DIN for S-Video).

 

So although there were some initial schematics that got previously shared, suffice it to say that they are not based on the final design. There have been tweaks that were born out of having gained a better understanding of the LM1889 chip's color subcarrier output characteristics. To put it into a nutshell, the video quality has continued to improve, and it's now looking very nice. Even the composite output looks great.

 

So here we see a composite CRT on the Left, and an S-Video LCD on the right.

 

Keep in mind that the CRT is not some fancy Sony PVM, but just a cheap little DVD semi-portable player I purchased in 2004. The LCD is a bit newer by probably 5-6 years, being a 720P Vizio with a whole gamut of inputs ranging from Component, VGA, HDMI, Composite, and S-Video (that's the way things should have stayed -- oh well).

 

This newest video upgrade design puts out fully buffered versions of Composite and S-Video simultaneously without any cross talk between them. Jail bars are extremely minimal, and what's present is most likely due to the protoboard with the plugged in mess of wires that I'm using for my tests (you can see part of it off to right). I feel really good that once this gets made into a real PCB it'll be perfect.

 

The cost of the parts required is very low, and there are absolutely no hard to get components, although a few do need to be purchased from either eBay or AliExpress.

 

When I get the design locked in, I'll post the final schematics. And following that I'll post the PCB gerbers as well.

 

[+evg2000]

cool, any thoughts of adding it to the PCBway shared projects?
https://www.pcbway.com/project/

[+mytek]

5 hours ago, evg2000 said:

cool, any thoughts of adding it to the PCBway shared projects?
https://www.pcbway.com/project/

I wouldn't be opposed, although I've normally shared my stuff as gerbers, and if the board is fairly small, OSH Park. I think you'll find that in the long run it's cheaper to have boards made and shipped by JLCPCB then PCBway. At least this has been my experience. So if you have PCB gerber files in hand, you can take it anywhere and have boards made.

[+evg2000]

Oh, it the board is small then I would go OSH park also.

[+mytek]

On 9/30/2022 at 10:55 AM, evg2000 said:

Oh, it the board is small then I would go OSH park also.

I think this board will be on the threshold of what makes sense to have OSH Park do vs. someplace else. In other words it might be a bit pricey doing this through OSH Park when all things are considered. However their price does include both domestic and international shipping as part of it, so that'll have to get factored in when purchasing elsewhere. I'll see when I get done as to where things stand.

 

BTW, this project is integral to my CV-NUC+ Miniature ColecoVision console, and was one of the prerequisites as to whether I would pursue that.

 

EDIT: This project is 4th in line for the moment. Kinda focusing on the Mini CV-NUC+ project right now, and it consists of 3 boards.

[+mytek]

CV-EAV Board Development Back in the Queue

 

While I wait for the Mini CV-NUC+ boards to arrive from the manufacturer, I finished up the design for this project and have boards currently being made as I write this. This is the same video circuit used in the CV-NUC+, but made as an upgrade for the stock NTSC Colecovision instead. Also because much of the support circuitry is already on the motherboard, this part of the video circuit is much less involved than what went into the CV-NUC+.

 

Board Top

 

Board Bottom

 

Schematic: CV-EAV_V1.1_schema.pdf

 

Similar to how the original RF Modulator sits in a stock NTSC Colecovision, this board which takes the place of the modulator gets flipped upside down before mounting to the CV motherboard. All the required connections to the motherboard are made via header J4, with audio and video outputs via J1-J3 (RCA and Mini-DIN4). Two metal standoffs secure the EAV board to the CV motherboard, with one of the holes already being present on the motherboard, and the other can be marked and drilled. Minimal modification is required to the case, requiring the addition of one extra hole to accommodate the audio RCA jack.

 

The S-Video output is of a very high quality, and suitable for conversion to HDMI via numerous such adapters on the market.

 

This board version will not work for a PAL Colecovision, only rendering a B&W signal due to the missing color burst oscillator. To add this extra circuit would require a considerable amount of additional components, and was not the focus of this first design. Perhaps in the future there will be a PAL derivative developed that will accommodate this.

[+-^CrossBow^-]

4 hours ago, mytek said:

Board Development Back in the Queue

 

While I wait for the Mini CV-NUC+ boards to arrive from the manufacturer, I finished up the design for this project and have boards currently being made as I write this. This is the same video circuit used in the CV-NUC+, but made as an upgrade for the stock NTSC Colecovision instead. Also because much of the support circuitry is already on the motherboard, this part of the video circuit is much less involved than what went into the CV-NUC+.

 

Board Top

 

Board Bottom

 

Schematic: CV-EAV_V1.1_schema.pdf

 

Similar to how the original RF Modulator sits in a stock NTSC Colecovision, this board which takes the place of the modulator gets flipped upside down before mounting to the CV motherboard. All the required connections to the motherboard are made via header J4, with audio and video outputs via J1-J3 (RCA and Mini-DIN4). Two metal standoffs secure the EAV board to the CV motherboard. Minimal modification is required to the case, requiring the addition of one extra hole to accommodate the audio RCA jack.

 

The S-Video output is of a very high quality, and suitable for conversion to HDMI via numerous such adapters on the market.

 

This board version will not work for a PAL Colecovision, only rendering a B&W signal due to the missing color burst oscillator. To add this extra circuit would require a considerable amount of additional components, and was not the focus of this first design. Perhaps in the future there will be a PAL derivative developed that will accommodate this.

This looks great! But I'm curious. You state only 1 extra hole is needed for the audio? Then how is the s-video and cvbs jacks being accessed? I'm assuming the pcb mount s-video comes through the original opening of the RF modulator but the composite RCA?

 

[+mytek]

10 hours ago, -^CrossBow^- said:

This looks great! But I'm curious. You state only 1 extra hole is needed for the audio? Then how is the s-video and cvbs jacks being accessed? I'm assuming the pcb mount s-video comes through the original opening of the RF modulator but the composite RCA?

The Mini-DIN4 S-Video jack lines up perfectly with the original RF output hole which is more than big enough. And the RCA composite video output lines up with the rectangular channel switch hole, but that hole will need to have one side lengthened with a round file (1 CM diameter file is pretty close to perfect). For the audio RCA jack, a step drill is recommended.

Amazon Link to Step Drill Set

 

 

EDIT: I'll be providing a PDF for printing a paper template for locating the audio jack hole. The modification to the Channel Switch hole is fairly obvious, and kinda locates itself.

[+mytek]

Final Version of CV-EAV Board Done, Installed, and Working

 

Needs the entire RF modulator removed (lots of heat required -- pain in the ass ), and a new header installed, plus one hole drilled for for a standoff.

 

Board gets installed component side down.

 

Produces nice looking S-Video and Composite output.

 

Looks like something that could've been BITD.

 

To make life easier with the holes, I just made them U-Shaped after first using a Uni-Bit to drill the holes.

 

I kinda doubt that I'll release this, since it really requires a lot of heat to remove the RF modulator. So much so that I slightly warped the Colecovision PCB. Anyway at least after all the work it served as a model for the video circuits I used in my CV-NUC+ design, and it does produce a great looking image from my original Colecovision console 

[Pixelboy]

Nice! So... could this be used on a breadboard ColecoVision? 

 

[+-^CrossBow^-]

5 hours ago, mytek said:

I kinda doubt that I'll release this, since it really requires a lot of heat to remove the RF modulator. So much so that I slightly warped the Colecovision PCB. Anyway at least after all the work it served as a model for the video circuits I used in my CV-NUC+ design, and it does produce a great looking image from my original Colecovision console 

I've never really had any issues removing the RF modulator. I do it pretty often when I'm requested to replace the capacitors since there are usually three of them hiding inside the RF. So I'm still curious and quite interested in getting a hold of one to perhaps put into my spare CV that is still stock. I would put it into my CV with the TMS-RGB setup, but I already have RCAs installed from the earlier composite kits I've had in it in the past. It might also interfere with my RGB output since I installed that just left of where you have your s-video using the original RF opening.

 

[+mytek]

10 hours ago, -^CrossBow^- said:

I've never really had any issues removing the RF modulator.

Well I removed everything, The RF Modulator PCB and the metal surround shield. At least this version of the CV-EAV can't fit inside the original metal shielded box. My reason behind this approach was due to needing 3 jacks instead of the original one RF jack. So if I had left the metal shielded box, it would have required extensive mods to the box itself, probably every bit as difficult as simply removing the entire thing as i did.

 

The other aspect that I'm also not 100% happy about is that I can see faint jail bars in the S-Video, and not so faint ones in the composite output. You normally can only see them in a large area being painted the same color. I know its not my video cable, especially on the S-Video side, since I am using a triple shielded very high quality cable with individually shielded coax for each signal (Luma & Chroma), with an overall separate shield wrapped around both coaxes. Normally jail bars are an S-Video thing where the luma and chroma are inadvertently bleeding over into each other. However since I see it a bit worse over the composite output which is intentionally the product of mixing those two signals, the source of the jail bars has got to be elsewhere (e.g., LM1889 chip).

 

I've checked out the reviews on the AV-Mod boards, and quite frankly I don't think people are as picky as me and rarely ever mention the jail bar issue, which i know must exist, and have seen whenever an actual screenshot has been shown. I guess compared to RF which can be far worse in all regards, they are happy to at least have something better than that, and don't really notice the jail bars.

[+-^CrossBow^-]

I'm guessing the RF must be done differently then because on all of mine I just desolder the interconnect block pins on the bottom of the PCB, and use my desoldering iron to remove the solder around the tabs. The entire thing outer shield housing and all just comes right up off the board. Done it many times and not had much trouble at all doing it.

 

So with that method, I don't have to remove the daughtboard inside because it comes up with the rest of the housing as a single unit.

 

 

[+mytek]

1 hour ago, -^CrossBow^- said:

I'm guessing the RF must be done differently then because on all of mine I just desolder the interconnect block pins on the bottom of the PCB, and use my desoldering iron to remove the solder around the tabs. The entire thing outer shield housing and all just comes right up off the board. Done it many times and not had much trouble at all doing it.

 

So with that method, I don't have to remove the daughtboard inside because it comes up with the rest of the housing as a single unit.

Well I think the problem is that on my board someone got over zealous with soldering the metal box tabs to the board (lots of solder was used). I ended up having to use a solder gun, big solder sucker, wick, and prying to get it separated from my board.

 

I also noticed when I was removing parts during my experiments, that Coleco obviously never heard of using thermal relief pads on their PCB when connecting to large copper ground areas, so that too took a lot of heat to break those components free from the PCB. They also liked to use both a top and bottom copper ground area connected with pads on both sides, each pad without the use of thermal relief. Definitely not a fun board to work on .

 

I applied solder to the top because I wasn't sure if there were actual plated-thru vias still intact after my previous de-soldering and subsequent replacing of those components. At one point all of the components had been removed so as to prevent their influence when I was developing the video circuits for use in the CV-NUC+. New components got used in their place when they got replaced, hence the reason it looks different than when it originally left the factory.

 

Below is an example of no thermal relief (Figure A) which was done in many places on the Colecovision, and proper thermal relief (Figure B).

 

[Ruggers Customs]

I'm with @-^CrossBow^- here on the ease of removing the rf module.  I've removed hundreds (literally) and it's never been an issue either with wick or a gun.  With a gun it taks me less than 8 seconds.  Here's one I removed completely and put back on less than 2 hours ago so I could replace the caps and fit a TMS-RGB mod.

[+mytek]

On 2/25/2023 at 4:43 PM, Pixelboy said:

Nice! So... could this be used on a breadboard ColecoVision? 

 

Oops I missed this when you first posted it.

 

The circuit I used in the CV-EAV could certainly be used, but this particular board is reliant on circuits that already exist in a stock Colecovision. In a breadboarded system it would be better to use the entire video circuit from the CV-NUC+ project instead. Final schematics for that system should get released by April.

[+mytek]

12 hours ago, Ruggers Customs said:

I'm with @-^CrossBow^- here on the ease of removing the rf module.  I've removed hundreds (literally) and it's never been an issue either with wick or a gun.  With a gun it taks me less than 8 seconds.  Here's one I removed completely and put back on less than 2 hours ago so I could replace the caps and fit a TMS-RGB mod.

Apparently I have an unusual case, because looking at your photo it only appears that 4 points got soldered, whereas mine got soldered down in 5 places. And a lot more solder got used as well. Perhaps someone was monkeying around with my CV before I got my hands on it. Anyway with only 4 spots soldered it would certainly be easier to remove it.

 

So getting back to the faint jail bars I was seeing. It was actually worse when viewed on my CRT with Composite, than it was over S-Video to my VIZIO LCD. This meant that the jail bars were most likely being produced from either the TMS or LM1889 color encoder. I also recall reading somewhere that if the chrominance is too hot that it might cause this effect to happen in our systems. So I ran a bit of an experiment today and tried using a smaller capacitor between the LM1889 subcarrier output (pin-13) and my chroma buffer transistor circuit.

 

Right Click on the image below and select open link in a new tab, and then if your cursor is a magnify symbol (+) click it again. Ignore the slight contrast shift between the two images, in reality they are the same, but my camera did something odd.

Since this is being displayed on a CRT the shadow mask pixels are easily seen, but what I'm hoping will show up in this photo comparison is that there are evenly spaced vertical darker areas on the left image which I am referring to as jail bars.

 

Although the subcarrier and colorburst are being reduced by roughly a factor of 2 in the 2nd image to the right, the overall color saturation is not being affected, but the jail bars vanish.

[Ruggers Customs]

4 hours ago, mytek said:

Apparently I have an unusual case, because looking at your photo it only appears that 4 points got soldered, whereas mine got soldered down in 5 places. And a lot more solder got used as well. Perhaps someone was monkeying around with my CV before I got my hands on it. Anyway with only 4 spots soldered it would certainly be easier to remove it.

 

 

The pic is after I put it back on and all 5 points are soldered.   It's just not necessary to flood the entire area around the tabs with solder.  As far as removal, all 5 points were flooded before.

Excellent work on the removing the jailbars! 

[+mytek]

1 hour ago, Ruggers Customs said:

The pic is after I put it back on and all 5 points are soldered.   It's just not necessary to flood the entire area around the tabs with solder.  As far as removal, all 5 points were flooded before.

Usually I'm pretty good with the soldering iron, but maybe it was just a bad day for me. Anyway you guys have convinced me that I should go ahead and release the gerbers for this project, and I will do that once I put together a BOM.

 

1 hour ago, Ruggers Customs said:

Excellent work on the removing the jailbars! 

Thanks   I'm quite happy to have figured it out, because I was wracking my brain trying to come up with a solution. Did lots of searches for anything remotely connected to the problem, but really wasn't finding much. Then when I saw the blurb about having too much chroma perhaps being the problem a light bulb went on, and I recalled an earlier prototype of the design that had a circuit producing less chroma looking fine on composite. Probably if I hadn't diligently tried multiple ideas to perfect this board, I could have missed that. And also being rather resistant about putting something out there with even a tiny flaw, forced me to make it right.

[Derf_Jagged]

Did you ever release the final schematics and gerbers anywhere?

[mattsoft]

Would love to see the schematics available for this. Interested in modifying the design and replacing the s-video with a single TRSS jack to eliminate the need to drill holes in the CV case.