RGB/YPbPr Intellivision Thoughts

[yell0w_lantern]

Okay, The sound jacks are done and the mod board is wired to the MB. I am waiting on the Retrotink 2x Pro and should finish wiring the component jacks tomorrow. I have the jumper set for component. I don't need the sync or 5v from the output - just R, G, B, and Gnd, right? Any resistor changes recommended? I'm just using a standard 3 plug component wire.

[the_crayon_king]

On 3/27/2021 at 2:07 AM, yell0w_lantern said:

Okay, The sound jacks are done and the mod board is wired to the MB. I am waiting on the Retrotink 2x Pro and should finish wiring the component jacks tomorrow. I have the jumper set for component. I don't need the sync or 5v from the output - just R, G, B, and Gnd, right? Any resistor changes recommended? I'm just using a standard 3 plug component wire.

Sorry for the delay in response. I don't seem to be getting push notifications for this page.

 

If you solder down the "COL" jumper then R=Ry, G=Y, and B=By.

Sync is inside of Y so you don't need it. The 5v is a voltage ref only needed by some RGB scalers (so you don't need it either). Just take ground from that same area and you should be good to go.

 

I didn't test with the Retrotink 2x Pro (I use the retrotink 2x SCART).

It is entirely possible (but unlikely) that this might not work on the 2X Pro.

 

If you could please confirm it works when you get everything together so I can add it to the list of things this works on.

[+-^CrossBow^-]

38 minutes ago, the_crayon_king said:

(I use the retrotink 2x SCART).

What did you use for testing?

 

Because as I've stated I was getting constant video dropouts with the Tink2x SCART through this board. The second level of Thunder Castle with the wizards as an example is totally unplayable because the video drops out through the Tink2x SCART every few seconds and takes a few seconds to recover each time. Tried it on 3 different flat panels from different manufactures and it did it on all of them.

 

The LTO menu was usually stable, but did have some instances of blinking out here and there.

 

Any of the mazes on Diner that were using the bright deep blue color would also cause frequent video drop outs through the Tink2x SCART.

 

 

[the_crayon_king]

2 hours ago, -^CrossBow^- said:

What did you use for testing?

 

Because as I've stated I was getting constant video dropouts with the Tink2x SCART through this board. The second level of Thunder Castle with the wizards as an example is totally unplayable because the video drops out through the Tink2x SCART every few seconds and takes a few seconds to recover each time. Tried it on 3 different flat panels from different manufactures and it did it on all of them.

 

The LTO menu was usually stable, but did have some instances of blinking out here and there.

 

Any of the mazes on Diner that were using the bright deep blue color would also cause frequent video drop outs through the Tink2x SCART.

 

 

 

Sorry, I didn't notice your post because of the aforementioned lack of push notifications.

 

Anyway, is your issue with real games or only the LTO. I don't have a LTO or a way to get one but I have a suspicion it is doing weird things to sync. @mattyv316 made it sound like he was getting good results. Maybe the difference comes down to a different LTO revision?

 

If the issue happens with real games I might have some theories about that too but let me know.

[+-^CrossBow^-]

Yes I was only using my LTO. I have one of the earlier first releases of the cart and haven't had any issues with it in the past. Again, it was only certain screens that would trigger this effect.

 

WAIT I just remember that my actual cart of MARIA was dropping off video anytime the screen used Grey colored borders. Basically similar to how screens with lots of colors would trigger it. As an example, EVERY time a dragon would show up on the first level of Thunder Castle, the screen flashes and that flashing would cause the video to drop out.

 

 

[mattyv316]

I just wanted to clarify my experiences with this.

I was having drop outs on paging up/down browsing the LTO menu of roms, but this was only happening when I had the Fred K board installed. It was consistent on multiple TVs and had that board in an INTV II and INTV 3. Same behavior. I also have (2) 2x Scart boxes. Still consistent behavior.

I used the Crayola board (early revision) on a SSVA and have not seen drop outs. I would have to test @-^CrossBow^- 's use cases specifically, but I definitely have not seen them in the places I saw them when I use the Fred K board.

I liked the result in my SSVA, so I bought another Crayola board to swap out with the Fred K. board in my INTV 3. This one was the current revision and behaves the same as my SSVA.

I sill have my INTV II on the Fred mod, so I will try to do another comparison in the next day or so. I hope this info helps

[the_crayon_king]

49 minutes ago, -^CrossBow^- said:

Yes I was only using my LTO. I have one of the earlier first releases of the cart and haven't had any issues with it in the past. Again, it was only certain screens that would trigger this effect.

 

WAIT I just remember that my actual cart of MARIA was dropping off video anytime the screen used Grey colored borders. Basically similar to how screens with lots of colors would trigger it. As an example, EVERY time a dragon would show up on the first level of Thunder Castle, the screen flashes and that flashing would cause the video to drop out.

 

 

MARIA is another non-OEM game correct?

 

I checked my code sync is active on

10001 --I have never seen this trigger.

10101

11001

11101

 

Compare that to the AY-3-8915 datasheet:

 

Sync is written as 1XX01 which are all included in the code I wrote.

 

The kit puts out what it is given. So the only possibility AFAIK is that these carts are putting out a sync code when they are not supposed to or they are not putting out a sync code when they are supposed to.

Unless maybe you are having some colors that are wrong on top of the sync issue. That would mean one of the pins (probably V2) is in the wrong place.

 

If I had one of these carts I could probably find the issue in a few mins. If you have an oscope you could look for the issue similarly.

 

The last thing I can think of is maybe it has something to do with PAL. I know that crappy scaler will output PAL onto my TV when nothing else I have will. I don't know if there is a difference between the code that would be ran on a PAL/NTSC cart. I know your internal timing would be NTSC (assuming you have a NTSC console).

Edited March 30, 2021 by the_crayon_king

[+-^CrossBow^-]

My SSVA is actually a standard 2609 board and EXEC since the original board set didn't work on it when I got it about 20 years ago. Having said that I have these drop outs on the LTO when navigating yes, but also in the instances I mentioned. Maria is a homebrew from Cote Gamers in France yes but the games aren't coded for NTSC or PAL, they run as the console dictates. Mine is most definitely an NTSC console. 

 

About the only thing I can do is get another Intelly model 1 to install the board in and see if what I'm getting with your board is something with this particular intellivision or not.

 

So to sum up:

 

Your first board didn't work at all of for me regardless of what I tried.

Second board does work great all the time but only through the SCART to HDMI converter.

Second board has frequent video dropouts through my Tink2x SCART on multiple games that I played.

 

 

[the_crayon_king]

9 minutes ago, -^CrossBow^- said:

My SSVA is actually a standard 2609 board and EXEC since the original board set didn't work on it when I got it about 20 years ago. Having said that I have these drop outs on the LTO when navigating yes, but also in the instances I mentioned. Maria is a homebrew from Cote Gamers in France yes but the games aren't coded for NTSC or PAL, they run as the console dictates. Mine is most definitely an NTSC console. 

 

About the only thing I can do is get another Intelly model 1 to install the board in and see if what I'm getting with your board is something with this particular intellivision or not.

 

So to sum up:

 

Your first board didn't work at all of for me regardless of what I tried.

Second board does work great all the time but only through the SCART to HDMI converter.

Second board has frequent video dropouts through my Tink2x SCART on multiple games that I played.

 

 

I thought we figured that you cooked that first one by shorting the outputs to ground?

 

When you say sync drop through multiple games are any of those OEM carts?

Anyway, I think what I am trying to relay is that sync output is whatever is input. If sync is jacked up (more than normal) then it is not being put in correctly somewhere. I might be able to work around it if I had one of these non-working carts but it doesn't seem like something the kit is doing wrong. I mean I won't go so far as to say it is impossible; I just don't see how it could be an error in my sync code when it is following the datasheet 1:1.
 

Probably not relevant but anything input over 5.5V for a sustained period is bad. I have seen something like 7V on one of my model 1's shortly before it cooked itself. It didn't live long enough to get a kit installed. The 12 ohm resistor in it and another model 1 shorted which seemed to fry a few transistors. I don't think any of that is related to your issues but it wouldn't hurt to measure the AY-3-8900-1 supply voltage and make sure it is not excessive. I am not sure how any of that could be relevant to your issue, just throwing it out there.

 

Maybe a model 2 would do something different? I certainly don't think a different model 1 would do anything significantly different.

 

[mattyv316]

Ok....so, I was able to do a little more testing.

SSVA w/Crayola mod -->2x Scart:

LTO - I was able to experience a few drop outs after playing with browsing/paging up and down on the menu after a while. Definitely WAY less than what I had with the Fred K mod since I really had to play with the browsing quite a bit to get it to trigger (w/Crayola).

However, I played Thunder Castle with this setup for about 20-30 minutes with no dropouts or issues.

SSVA w/Crayola mod -->Rad2x Genesis 2 cable:

LTO - No issues at all. I tried browsing the menu for a while and never experienced and drop outs. Also, playing Thunder Castle and other games did not trigger any drop outs.

 

[+-^CrossBow^-]

My SSVA has both regulators replaced along with full recap about 3 years ago and the 12ohm resistor has been replaced on the power board. The rest of it is still stock minus the composite board I also have installed. But I know that my +5 reads at about 5.04v at the ferrite bead just after the ribbon cable. In most other places the +5 drops down to about 4.85 to most of the ICs. Also both of the 3906s were replaced out this summer just for good measure as those do burn out commonly on the model 1.

 

Thunder castle would only do it when the screen flashed, and every few seconds on the second screen with the wizards. I couldn't get any further because of the video drop outs I couldn't see what was going on.

 

The first board might have fried because of what you described, but that is because every other RGB board I've installed you remove the 75Ω and bridge them because they are usually wired in series and not tied to ground so I wasn't ware that you them arranged that way. BTW on this new board I did confirm that removing the 75Ω resistor and leaving the pads empty, causes a VERY dark image to appear on my setup so the 75Ω must be there in order to get a picture to output.

 

I do have a model 2 but I've not done anything with it as I don't use it very often. The power/reset is flaky and I'm not a fan of the controllers. But I can give it a shot and see if it behaves any differently.

 

[the_crayon_king]

25 minutes ago, -^CrossBow^- said:

My SSVA has both regulators replaced along with full recap about 3 years ago and the 12ohm resistor has been replaced on the power board. The rest of it is still stock minus the composite board I also have installed. But I know that my +5 reads at about 5.04v at the ferrite bead just after the ribbon cable. In most other places the +5 drops down to about 4.85 to most of the ICs. Also both of the 3906s were replaced out this summer just for good measure as those do burn out commonly on the model 1.

 

Thunder castle would only do it when the screen flashed, and every few seconds on the second screen with the wizards. I couldn't get any further because of the video drop outs I couldn't see what was going on.

 

The first board might have fried because of what you described, but that is because every other RGB board I've installed you remove the 75Ω and bridge them because they are usually wired in series and not tied to ground so I wasn't ware that you them arranged that way. BTW on this new board I did confirm that removing the 75Ω resistor and leaving the pads empty, causes a VERY dark image to appear on my setup so the 75Ω must be there in order to get a picture to output.

 

I do have a model 2 but I've not done anything with it as I don't use it very often. The power/reset is flaky and I'm not a fan of the controllers. But I can give it a shot and see if it behaves any differently.

 

No I understand completely how/why it happened. What you did was very much something I would have thought to do with literally any other kit.

 

The model 2 is so much less painful of an install imo. You might have difficulties with placing the connector you are using but that is about it. You may know this but if you lift up the little metallic cap you should just be able to clean it and pop it back in.

 

-------------------------------------------

I mean we know there is an issue., just would help to figure out where it is. My money is still on the carts doing something weird. You have the cart, the console, the kit, the scaler, the TV. Any of those could be at fault.

 

I can do something in the future that should be able to fix about any sync issue. It is just going to take a bit of a redesign but the code is already written. It was the waveform I was showing you guys earlier in this thread. I believe that should work on the OSSC as well. I need to get a logic level translator and then I should be able to at least test that. My ultimate goal is to just shoot out HDMI but I am lacking the free time to learn how to do that atm.

 

[yell0w_lantern]

Yes, I will do that. I just found out that I have some repair work to do on my Intv though, so it may be a little bit.

[+-^CrossBow^-]

11 hours ago, the_crayon_king said:

The model 2 is so much less painful of an install imo. You might have difficulties with placing the connector you are using but that is about it. You may know this but if you lift up the little metallic cap you should just be able to clean it and pop it back in.

Is there a way to get the Intelly 2 to work without the RF Modulator? I seem to recall seeing vids where others have tried to install the composite kits into Intelly 2s and when they removed the RF modulator, the unit would no longer come up or produce any picture at that point? If it could be removed and made to still work, then the connectors I use should still be able to be soldered down to the main PCB where the modulator was located and just use the hole from the RF output for the new connector. That would be the most ideal in this case.

 

And yeah I know to clean the switch on the Intelly 2, I'm just not really a fan of the thing and just prefer to use the original model systems.

 

But the Intelly 2 is easier to get into since it doesn't have all of that shielding on it. So definitely worth a try. Could be some time before I can get to this though as I have other systems on the way to me I have to take care of first.

 

 

[the_crayon_king]

1 hour ago, -^CrossBow^- said:

Is there a way to get the Intelly 2 to work without the RF Modulator? I seem to recall seeing vids where others have tried to install the composite kits into Intelly 2s and when they removed the RF modulator, the unit would no longer come up or produce any picture at that point? If it could be removed and made to still work, then the connectors I use should still be able to be soldered down to the main PCB where the modulator was located and just use the hole from the RF output for the new connector. That would be the most ideal in this case.

 

And yeah I know to clean the switch on the Intelly 2, I'm just not really a fan of the thing and just prefer to use the original model systems.

 

But the Intelly 2 is easier to get into since it doesn't have all of that shielding on it. So definitely worth a try. Could be some time before I can get to this though as I have other systems on the way to me I have to take care of first.

 

 

I disabled RF on mine a long time ago. The composite mod is mixing from the outputs of the AY-3-8915; this kit is mixing from the inputs of that same ic. So RF shouldn't matter.

[the_crayon_king]

I have a logic analyzer now so I have a possibly more accurate overall representation of the total waveforms.

 

Attached is the file of this capture which can be viewed with KingstVIS.

For example you can type "0b1_1001" (correlates to 11001) to see the number of vsync pulses.

Doing this has revealed that 100XX, and 1101X are never active. Not useful to know but interesting nonetheless.

 

I have finally ordered a OSSC so I will see what can be done to make it function.

As mentioned previously it is unlikely that the current kit can be made to work with the OSSC. I should be able to use the "sync fix" code to make it work (with a dev board I have) and then see how much I can simplify code from there.

 

Use the AY-3-8915 datasheet for reference. I hope now that we collectively have enough data.

 

timing.kvdat

[yell0w_lantern]

Okay, I fixed my power board and finally tested the Retrotink 2x Pro and I am pleased to report that it works.

[the_crayon_king]

38 minutes ago, yell0w_lantern said:

Okay, I fixed my power board and finally tested the Retrotink 2x Pro and I am pleased to report that it works.

Thank you for reporting back. It is good to know what exactly these work on.

 

I meant to mention that if you do not plan on using RF that you can repurpose the RF switch to change one of the toggle options YPBPR, SOG, Color Palette. There should be a free floating pin to attach to. That is what I did on my personal console. I actually am not sure if it effects RF at all, but to be safe I removed the trace that wasnt ground on mine.

Edited April 6, 2021 by the_crayon_king

[the_crayon_king]

So I was looking at generating composite and s-video just because. Anyway, the on board "composite" is noisy AF. So I ran the RGB through a BH encoder and generated composite that way. Neither signal is 100% correct since I did this with components I just happened to have around. I think the composite from RF needs amplified by 6X. The stupidest way to do this would be to run the same signal three times through a THS7314. I will do that later. If anyone has scope shots of an actual composite mod I'd like to see those as well.

 

Below is a comparison:

 

Edited April 7, 2021 by the_crayon_king

[+-^CrossBow^-]

Installed the board into my Intell2 yesterday and do get different results, although I wouldn't say they are that much better.

 

Basically the video dropouts were less than they were on my model 1, and on Thunder Castle (Using an actual cart btw), it wasn't dropping out when the lightning flashes would occur. However, within a few minutes I began to notice some odd flickering in the yellows off the left sides of characters and the score. On the second maze screen with the wizards, it also wasn't dropping video but was starting to get progressively worse with the flickering on the left hand side of yellow objects. Eventually they would start to turn orange and red on the edges? And then the video drop outs started to occur constantly again.

 

Now, what I discovered is that it appears I have a bad STIC in my intelly 2 causing these video issues. I put another STIC in it from a model 1 and the video flickering anomolies went away. To be sure it wasn't some fluke I put the original back in and the flickering and color oddness game back after about 10min of being on.

 

Anyway, with the replacement STIC, the same video dropouts I was getting in the model 1 are still happening in this model 2 with my Tink2x SCART and both of my flat panels I tested it on. One is a lower end Toshiba 1080P, the other is my slightly higher end Sony Bravia 1080P. So I decided to try something else.

 

I cut the traces to the Intelly 2 channel select switch (Because I removed the RF modulator to install this), and attached it back to the YUV/RGB selector pads. I then wired up a temp set of RCAs for component. Connected this to my CRT and it works just fine without any video dropouts at all. But the overall color palette seems a tad darker on some shades (Especially greens) compared to the RGB but still looks good. I then connected it up to my Extron 7SC component inputs and I'm pleased to report it also works great there as well. However, not everyone will have this setup of course. I did NOT try and connect to the OSSC but I can try and do that a little later. It isn't that easy to get to my OSSC as I have it mounted and out of the way but will see and report back. So at least for me, using this board works good for my needs in YUV output mode vs RGB. If it works with OSSC that way, then I would just offer it as a component output solution. I will purchase a TRRS jack and TRRS component cable breakout to attach to it so that technically my Intelly 2 can be used wither way using the channel switch.

 

 

[the_crayon_king]

7 hours ago, -^CrossBow^- said:

Installed the board into my Intell2 yesterday and do get different results, although I wouldn't say they are that much better.

 

Basically the video dropouts were less than they were on my model 1, and on Thunder Castle (Using an actual cart btw), it wasn't dropping out when the lightning flashes would occur. However, within a few minutes I began to notice some odd flickering in the yellows off the left sides of characters and the score. On the second maze screen with the wizards, it also wasn't dropping video but was starting to get progressively worse with the flickering on the left hand side of yellow objects. Eventually they would start to turn orange and red on the edges? And then the video drop outs started to occur constantly again.

 

Now, what I discovered is that it appears I have a bad STIC in my intelly 2 causing these video issues. I put another STIC in it from a model 1 and the video flickering anomolies went away. To be sure it wasn't some fluke I put the original back in and the flickering and color oddness game back after about 10min of being on.

 

Anyway, with the replacement STIC, the same video dropouts I was getting in the model 1 are still happening in this model 2 with my Tink2x SCART and both of my flat panels I tested it on. One is a lower end Toshiba 1080P, the other is my slightly higher end Sony Bravia 1080P. So I decided to try something else.

 

I cut the traces to the Intelly 2 channel select switch (Because I removed the RF modulator to install this), and attached it back to the YUV/RGB selector pads. I then wired up a temp set of RCAs for component. Connected this to my CRT and it works just fine without any video dropouts at all. But the overall color palette seems a tad darker on some shades (Especially greens) compared to the RGB but still looks good. I then connected it up to my Extron 7SC component inputs and I'm pleased to report it also works great there as well. However, not everyone will have this setup of course. I did NOT try and connect to the OSSC but I can try and do that a little later. It isn't that easy to get to my OSSC as I have it mounted and out of the way but will see and report back. So at least for me, using this board works good for my needs in YUV output mode vs RGB. If it works with OSSC that way, then I would just offer it as a component output solution. I will purchase a TRRS jack and TRRS component cable breakout to attach to it so that technically my Intelly 2 can be used wither way using the channel switch.

 

 

The sync output and the sync going to the DAC are the same so I don't know why one would have dropout and not the other. The respective devices must handle it differently.

 

After looking at a video for Thunder Castle I see now where the issue is happening...It's in the over-scan which is right around where sync is all jacked up.

There is a large gap of no color data in the data I posted that would be the overscan.

The area is all of the 0b1_1000 (11000) around vsync.

I need one of these intermittently working games. I would prefer the LTO (still waiting).

If this problem is solvable on a CPLD then I will solve it on the next revision "Baked Potato V4".

It might work on the new code I have already written for that upcoming board.

 

Anyway, V3 will still not work on the OSSC. It is because of the sync (for sure). I know this because I made code that works on the OSSC. I don't want to get into how the sync correction works but it was a pain.

That is well and true corrected Intellivision sync that works on the OSSC.

After using the OSSC on 5X mode I can see the hype. I ran the OSSC through the datapath which adds lines on the video but it still looks amazing. I am probably not making anymore boards of the old version. The next version will be 30-35 just like this one but will work on the OSSC and hopefully correct the overscan thing as well. 

Edited April 11, 2021 by the_crayon_king

[+-^CrossBow^-]

1 hour ago, the_crayon_king said:

The sync output and the sync going to the DAC are the same so I don't know why one would have dropout and not the other. The respective devices must handle it differently.

 

Anyway, V3 will still not work on the OSSC. It is because of the sync (for sure). I know this because I made code that works on the OSSC. I don't want to get into how the sync correction works but it was a pain.

That is well and true corrected Intellivision sync that works on the OSSC.

After using the OSSC on 5X mode I can see the hype. I ran the OSSC through the datapath which adds lines on the video but it still looks amazing. I am probably not making anymore boards of the old version. The next version will be 30-35 just like this one but will work on the OSSC and hopefully correct the overscan thing as well. 

Yes, the Extron 7SC is basically a very versatile commerical grade video converter. It can accept composite, s-video, components, RGBs, and RGBHV and outputs everything as VGA signal locked at whatever resolution I want it to be. It is quite likely that your RGB would work on it, but it is a pain for me to get behind it and connect it and besides not everyone has an Extron like device that can convert the video in that manner. But the component was something I wanted to test more as I thought it might be easier for using with other devices. It does work on CRTs or at least the one I have in my spare room for testing stuff on, and it works through my Extron without any issues at all. I've not tried to connect it to the OSSC but as you have found now that you have an OSSC in your hands, it likely wouldn't work under component either.

 

Keep me up to date on that new revision board that works with the OSSC. Because that has been the golden goose we needed for our eggs.

 

[the_crayon_king]

Main parts are ordered. PCB designed mostly. I don't want to go too much further without prototyping things out by hand.

 

If someone has a lead on a LTO let me know. I see them for like 400$ and I'm not paying that much.

I really can't correct issues without first recreating them.

 

Alternatively, if someone knows of OEM games where sync is lost from overscan stuff let me know so I can try to find one of those. So far we have Thunder Castle. This problem would probably manifest on games that flash something quickly in the overscan area (outside the normally viewable area).

[+-^CrossBow^-]

10 hours ago, the_crayon_king said:

Alternatively, if someone knows of OEM games where sync is lost from overscan stuff let me know so I can try to find one of those. So far we have Thunder Castle. This problem would probably manifest on games that flash something quickly in the overscan area (outside the normally viewable area).

Not that is helps as it is pricey as well, but I saw similar drop outs from some color combos used in levels of the game Diner as well. I'm trying to thing of other games I might have tested that did similar drop outs. As I stated, the game Maria has some color combos on the screens that would make it drop out.

 

Keep in mind, that all of these video drop outs were ONLY happening through RGB output through my Tink2x SCART. They didn't happen through the cheap SCART to HDMI converter and I'm not seeing them in component output mode through my Extron either.

 

[the_crayon_king]

26 minutes ago, -^CrossBow^- said:

Not that is helps as it is pricey as well, but I saw similar drop outs from some color combos used in levels of the game Diner as well. I'm trying to thing of other games I might have tested that did similar drop outs. As I stated, the game Maria has some color combos on the screens that would make it drop out.

 

Keep in mind, that all of these video drop outs were ONLY happening through RGB output through my Tink2x SCART. They didn't happen through the cheap SCART to HDMI converter and I'm not seeing them in component output mode through my Extron either.

 

Believe it or not I actually have Diner. Under what conditions would this happen?

From what I have seen of the game I haven't seen any overscan weirdness but maybe I haven't made it far enough.

And are you certain this issue wouldn't have been the STIC thing?

I will play this for a bit and see what happens.