Jump to content
SiLic0ne t0aD

New Coleco RGB board?

Recommended Posts

3 hours ago, Mister Fab said:

I took a picture of each transistor, I think I didn't make a mistake about they're all npn. Aren't collector and base inverted on pnp?

In NPN, the PN junction of base to emitter points to emitter, while in PNP, the base to emitter PN junction points toward base.  So they need opposite polarities to allow collector->emitter current flow.

 

MPSA13 is a Darlington pair NPN transistor.  2N3904 is a standard NPN transistor, and 2N3906 is a standard PNP transistor.

Share this post


Link to post
Share on other sites
4 hours ago, Mister Fab said:

Let me know if you see some incoherences so I can correct them.

Could you recheck U29 pin 4?  It now shows a connection to ground at the capacitor.  I think that instead, the capacitor connects to the via that horeshoes between the ground points underneath.  This leaves room for U29 pin 4 to reach Q26/Q23.  And I suspect that U29 pin 6 goes to AV pin 7 also.

Share this post


Link to post
Share on other sites
2 hours ago, ChildOfCv said:

Could you recheck U29 pin 4?  It now shows a connection to ground at the capacitor.  I think that instead, the capacitor connects to the via that horeshoes between the ground points underneath.  This leaves room for U29 pin 4 to reach Q26/Q23.  And I suspect that U29 pin 6 goes to AV pin 7 also.

I just rechecked pin 4, I don't get any continuity signal with any pin of AV connector.

 

I get a continuity signal between U29 pin 4 and C156 negative, and on the other side to C147

 

Sorry, I don't understand the green part I underlined in your comment.

Edited by Mister Fab

Share this post


Link to post
Share on other sites
13 minutes ago, Mister Fab said:

I just rechecked pin 4, I don't get any continuity signal with any pin of AV connector.

 

I get a continuity signal between U29 pin 4 and C156 negative, and on the other site to C147

 

Sorry, I don't understand the green part I underlined in your comment.

Well, you've partially answered my question anyway :)  C156 and C156A have negative pole connected to pin 4.  Okay, looking at the schematic in progress again, I can see that.

 

There's a large ground rail that goes in a horseshoe pattern under the PCB which connects several pins.  I thought it might do that and come back up on top at that leftmost pad.

 

And you confirmed that U29 pin 6 connect to the collector of (omitted transistor) Q26?

 

So the Y signal controls Q23, which controls the U29 pin 6 net and goes to the sync detector/amplifier.  I think the purpose here is to mute the color signals during sync periods, but would have to see it in action to confirm that.

Looks like I now have a use for C140 too.

Share this post


Link to post
Share on other sites
1 hour ago, ChildOfCv said:

Well, you've partially answered my question anyway :)  C156 and C156A have negative pole connected to pin 4.  Okay, looking at the schematic in progress again, I can see that.

 

There's a large ground rail that goes in a horseshoe pattern under the PCB which connects several pins.  I thought it might do that and come back up on top at that leftmost pad.

 

And you confirmed that U29 pin 6 connect to the collector of (omitted transistor) Q26?

 

So the Y signal controls Q23, which controls the U29 pin 6 net and goes to the sync detector/amplifier.  I think the purpose here is to mute the color signals during sync periods, but would have to see it in action to confirm that.

Looks like I now have a use for C140 too.

1 hour ago, ChildOfCv said:

 

 

And you confirmed that U29 pin 6 connect to the collector of (omitted transistor) Q26?

 

Yes, I confirm. It should have been Q26

 

Share this post


Link to post
Share on other sites

Yes, LM1881 based sync separator/stripper.

But it was a false track.

 

@ChildOfCv RGB signal blanking is not overkill in a pure RGB input ? Ok it work on an analogue TV, but could perhaps disturb the analogue front end of the OSSC (which is not a TV front end).

Did you not miss a junction between the base of Q22 and the output of U29B ?

Share this post


Link to post
Share on other sites

I have been looking for the schematic of the board in the first post. I tried contacting the designer but no luck so far. It says it's open source... The schematic isn't on OSHpark either.

I'd like to incorporate with attribution the circuit into the output stage of an Adam CPU board replacement in the future, so it has native RGB.

Has anyone seen the schematic anywhere?

Share this post


Link to post
Share on other sites
7 hours ago, emmanuelf said:

Did you not miss a junction between the base of Q22 and the output of U29B ?

Well, one thing I know is that R167 is 82 ohms, not 68 :P

 

I checked Mister Fab's trace work and don't see any such connection.  Q22 is the Y preamp, just like Q25 and Q20 are the R-Y and B-Y preamps, and it follows the pattern of the other two, so that's probably correct.  The sync detector samples pin 4 instead of the Y output because as soon as sync is signaled, the component outputs are shorted together and forced to a bias voltage by the network of U29C, U30A, and U30B.  At least that's my theory.

Share this post


Link to post
Share on other sites
33 minutes ago, ChildOfCv said:

He never provided the schematic as far as I know, but the board can be manufactured without one.

 

https://oshpark.com/shared_projects/WhLjVooT

Right, but I am trying to incorporate the circuit into a new PCB that replaces the Adam CPU/video board.

30 minutes ago, mobiusstriptech said:

The schematic is posted in this thread. It's open source and has been around for 20+ years

I will reread the thread. I likely missed it as I read the whole thread at like 3am ;)

The circuit is quite simple. I will be happy to provide ongoing supplies of these to my own design, ready assembled, if there's any demand for them.

Thanks folks!

Share this post


Link to post
Share on other sites

It's near the beginning. Be prepared for a lot of support if you offer these for sale. The original circuit needs many modifications and I spent a considerable amount of time dealing with people who were upset about the quality. Along with the need for an oscilloscope to properly dial it in, which most users won't have.

Share this post


Link to post
Share on other sites
2 minutes ago, mobiusstriptech said:

It's near the beginning. Be prepared for a lot of support if you offer these for sale. The original circuit needs many modifications and I spent a considerable amount of time dealing with people who were upset about the quality. Along with the need for an oscilloscope to properly dial it in, which most users won't have.

I didn't see a schematic for the PCB though.  I see the TMS document saying how you'd build one with mostly discrete components, but you'd have to cull out the composite output and 3.58MHz clock, then replace those sync separator components with an LM1881.

 

However, I'm starting to believe that the PAL version is good enough for both.  Citrus cautioned that it's "only for PAL", but that is incorrect.  The 9928 also has a color burst voltage level output (the 9928 documentation is flat wrong about that), and if you hook the 9928 directly to a TV you get ultra-blue--or at least I do.  The "PAL" version uses a sample and hold circuit to suppress that color burst.

Share this post


Link to post
Share on other sites

There was never a schematic for the PCB. It was made from that schematic. Anyone who wants to take it on is of course welcome to do so. I've been working on a replacement for a long time now but I have lots of projects and not a lot of time. So it is not a quick fix.

Share this post


Link to post
Share on other sites
6 minutes ago, mobiusstriptech said:

It's near the beginning. Be prepared for a lot of support if you offer these for sale. The original circuit needs many modifications and I spent a considerable amount of time dealing with people who were upset about the quality. Along with the need for an oscilloscope to properly dial it in, which most users won't have.

That's why I was hoping to get the evolved circuit and not have to re-do all the trial and error work. 

My first goal is to incorporate it into a complete replacement PCB where I can do the full set-up and control all the variables. The RGB-out would be buffered, and also sent to a converter IC (AD725) for S-Video and a much cleaner color composite out.

If I were to redo the board for fitting to the back of existing soldered components, I would design the board slightly differently, to minimize install-time errors. 

I have an oscilloscope. I have a full design and SMD production capability here. :D

In the future we're moving into a low cost level shifter plus Pi Zero option to emulate the 9928, with scaled HDMI output straight from the Pi Zero frame buffer. This will likely be 1/2-1/3rd the cost of the F18A MKII solution. However, that's just in the planning stages and I want to lay out this video circuit next week. ;)

I'd hate to have to order a board, and reverse engineer the schematic, just to get an open source circuit. I'd rather ask for help and give credit!

Share this post


Link to post
Share on other sites
1 hour ago, ChildOfCv said:

Well, one thing I know is that R167 is 82 ohms, not 68 :P

 

I checked Mister Fab's trace work and don't see any such connection.  Q22 is the Y preamp, just like Q25 and Q20 are the R-Y and B-Y preamps, and it follows the pattern of the other two, so that's probably correct.  The sync detector samples pin 4 instead of the Y output because as soon as sync is signaled, the component outputs are shorted together and forced to a bias voltage by the network of U29C, U30A, and U30B.  At least that's my theory.

Ok, I miss  the ac coupling via C147.

Yes that is my understanding too.

Share this post


Link to post
Share on other sites

At this point, my skills in electronics are just too low compared to yours, but I'll try to be good at drawing and to be the most accurate I can to finish to copy the circuit without any errors. It will take some time. I have nearly finished the back side. The component side is a bit tricky due to the tracks running under the large ICs.

Share this post


Link to post
Share on other sites
Quote

@ChildOfCv RGB signal blanking is not overkill in a pure RGB input ? Ok it work on an analogue TV, but could perhaps disturb the analogue front end of the OSSC (which is not a TV front end).

Forget what I said. During blanking, RGB signal should be very low, normally at black level but near 0v with some equipments to get them working.

 

All of these will for me to put mine under the scope ;-)

Need a little bit of motivation and some time and space. Will try to arrange that.

Share this post


Link to post
Share on other sites
1 hour ago, emmanuelf said:

Forget what I said. During blanking, RGB signal should be very low, normally at black level but near 0v with some equipments to get them working.

 

All of these will for me to put mine under the scope ;-)

Need a little bit of motivation and some time and space. Will try to arrange that.

Yeah it would be useful to find out when it activates those muting switches, and what voltage it creates.  Also, it would be useful to know what pins the DA15 to SCART adapter uses.  The pinout from CV.DK definitely has pins 4 and 5 backwards though.  OSSC expects R on SCART pin 15, G on 11, B on 7, and sync on 20.  So using the layout we've puzzled out:
 

DA15              SCART
5                 7
3                 11
1                 15
7                 20
6,8,14            5,9,13,14,17,18,21
15                2,6

Everything else is irrelevant to the OSSC, according to its schematic.

Share this post


Link to post
Share on other sites

I've been following this topic for more than a year and it's been exciting to see the flurry of new activity in the past week.  Something ChildOfCv said finally made me decide to come out of lurker mode to ask some questions:

 

On 2/4/2020 at 1:23 PM, ChildOfCv said:

I see the TMS document [...], but you'd have to cull out the composite output and 3.58MHz clock, then replace those sync separator components with an LM1881.

I spent a few minutes cleaning up the image from the TMS document: removing the 3.58MHz clock, the "Composite Video Encoder", and the "PAL Encoder" blocks.  This version is a lot easier to look at.  (See attached.)

 

Now for questions:

 

1. Was the LM1881 recommendation just a simple reduction in discrete component count?  Or does the LM1881 do something better/smarter than the "Sync Separator" block in the TMS document?  It certainly seems to do more, looking at the block diagram in the LM1881 datasheet.  With V-sync (and almost H-sync) output, you'd nearly get VGA for "free" in addition to RGB, right?

 

2. Judging by component count (and the lack of CD4066Bs, transistors, diodes, etc.) it looks like the citrus3000psi board has only replicated the "R-G-B Converter" block from the TMS document.  Is that correct?  If so, why aren't the "DC Restoration" and "Sync Separator" blocks from the TMS document also needed?  Their outputs seem to feed directly into the "R-G-B Converter" block's inputs.

 

3. Most/all of the R and C component values used on the citrus3000psi board are different than the TMS schematic.  Even things like the LM318's take -5V and +12V vs. the single +5V inputs to the LM359M on the citrus board.  Is this just because we have access to better rail-to-rail op-amps now?  The Coleco board gives us easy access to -5 and +12, so hopefully it wasn't left out just to simplify the circuit.  If we were to build exactly the circuit as presented in the TMS document (with the unnecessary composite blocks removed), would it perform better than the citrus board?

 

4. Kind of a bonus: wouldn't it be easier to get component video out?  Couldn't we drop most of the "R-G-B Converter" block, leaving just the 1.14x and 2.03x correction factors but not adding Y to B-Y and R-Y with the op amps?  Isn't that already a "native" component video signal?

 

Thanks!

TMS9928-RGB-only.png

Share this post


Link to post
Share on other sites
5 hours ago, Falonn said:

1. Was the LM1881 recommendation just a simple reduction in discrete component count?  Or does the LM1881 do something better/smarter than the "Sync Separator" block in the TMS document?  It certainly seems to do more, looking at the block diagram in the LM1881 datasheet.  With V-sync (and almost H-sync) output, you'd nearly get VGA for "free" in addition to RGB, right?

Yeah, the main reason for the 1881 is to reduce component count.  It may also do a better job than the discrete circuit, but keeping the size of the board down has a higher priority.

 

5 hours ago, Falonn said:

2. Judging by component count (and the lack of CD4066Bs, transistors, diodes, etc.) it looks like the citrus3000psi board has only replicated the "R-G-B Converter" block from the TMS document.  Is that correct?  If so, why aren't the "DC Restoration" and "Sync Separator" blocks from the TMS document also needed?  Their outputs seem to feed directly into the "R-G-B Converter" block's inputs.

Without completely reverse engineering the original board, I can guess.  But it may be due to the summing circuit that TI chose.  Perhaps with a different construction, the signal can be summed properly without resorting to this method.  You'll notice that during sync pulse, it's grounding the color components.

5 hours ago, Falonn said:

3. Most/all of the R and C component values used on the citrus3000psi board are different than the TMS schematic.  Even things like the LM318's take -5V and +12V vs. the single +5V inputs to the LM359M on the citrus board.  Is this just because we have access to better rail-to-rail op-amps now?  The Coleco board gives us easy access to -5 and +12, so hopefully it wasn't left out just to simplify the circuit.  If we were to build exactly the circuit as presented in the TMS document (with the unnecessary composite blocks removed), would it perform better than the citrus board?

Definitely because of access to better op-amps.  LM359 may still be available, but there are better, higher bandwidth alternatives that have rail range that makes +5V usable.  It's true that where the PSI3000 mounts, access to +12 and -5 aren't that close.  But if you have op-amps that can do their thing with +5V and ground, why bother grabbing other rails?

5 hours ago, Falonn said:

4. Kind of a bonus: wouldn't it be easier to get component video out?  Couldn't we drop most of the "R-G-B Converter" block, leaving just the 1.14x and 2.03x correction factors but not adding Y to B-Y and R-Y with the op amps?  Isn't that already a "native" component video signal?

Few TVs these days have component video in.  Even among those that do, you have to suppress the TMS9928's color burst voltages.  It seems that many TVs pick the period just after sync pulse to sample the color components for DC restoration.  But that's a bad time for the TMS because it's outputting a voltage that's intended to create the color burst for your TV system.  So when it returns to what it thinks of as neutral DC voltage, the TV ends up with weird tint.  On NTSC systems, that's a blueish background.  When it wants to make a blue screen, it assaults your eyes with the brightness of the blue.  PAL systems get both blue and green from that.

  • Thanks 1

Share this post


Link to post
Share on other sites
3 hours ago, ChildOfCv said:

Few TVs these days have component video in.  Even among those that do, you have to suppress the TMS9928's color burst voltages.  It seems that many TVs pick the period just after sync pulse to sample the color components for DC restoration.  But that's a bad time for the TMS because it's outputting a voltage that's intended to create the color burst for your TV system.  So when it returns to what it thinks of as neutral DC voltage, the TV ends up with weird tint.  On NTSC systems, that's a blueish background.  When it wants to make a blue screen, it assaults your eyes with the brightness of the blue.  PAL systems get both blue and green from that.

Well thank you - that explains the blue tint I have with my component modded NTSC CV.

Share this post


Link to post
Share on other sites
6 hours ago, Ikrananka said:

Well thank you - that explains the blue tint I have with my component modded NTSC CV.

Was this one of the common mods, or simply a direct connection to component outputs?

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...