Jump to content
IGNORED

AY-3-8915 to RGB adapter.


MrPix

Recommended Posts

18 hours ago, MrPix said:

If you want, we can combine projects. If not, that’s fine too. 

I appreciate, but as all my hobby projects always suffer indefinite delays at my grand despair, I will be more a millstone than anything else.

But I will follow you with great interest and will help as I can.

Link to comment
Share on other sites

Small update.

 

I have enough pins free to offer quite deep color, which means I’ll be able to hit both the NTSC and PAL palettes very accurately, within the limitation that RGB has a slightly smaller gamut than YPbPr.

  • Like 7
Link to comment
Share on other sites

4 minutes ago, grips03 said:

Will the video mod board have RGB, S-video and composite or just RGB?  Would be great if it had all 3.

 

So happy to see RGB board for Intv that would work with OSSC.

Me, too!  I'd build a prototype if there were schematics.

  • Like 1
Link to comment
Share on other sites

54 minutes ago, grips03 said:

Will the video mod board have RGB, S-video and composite or just RGB?  Would be great if it had all 3.

The board would have RGB output. A second board I designed takes RGB input and converts it very cleanly to S-Video and composite, with really good dot crawl performance. So all three can be had together.

54 minutes ago, grips03 said:

So happy to see RGB board for Intv that would work with OSSC.

Just a matter of reconstructing a few missing sync pulses.

48 minutes ago, KylJoy said:

Me, too!  I'd build a prototype if there were schematics.

All my projects are open hardware, and I typically release the schematics right away, and the gerbers after 3 months, so I can sell through a run to recover the development cost.

  • Like 1
Link to comment
Share on other sites

Sooo, with a better understanding of the problems ahead... :)

I think for the initial revision I will have to walk back talk of fixing the sync problem that trips up OSSC. Not that it's a huge problem, but getting the circuit basically working and used on a few select systems to see how people do with the palette - collecting feedback, that's a priority. It does seem people are more sensitive about colors than about OSSC compatibility right on the first try.

I'm also thinking of ways I might read the data into a frame buffer for upscaling and HDMI output a different way. One that's a lot cheaper than OSSC and other traditional HDMI solutions. Maybe using a Pi Zero and one of the open source bare metal framebuffer projects.

  • Like 1
Link to comment
Share on other sites

I looked at it. To reproduce his palette requires 18 bit color data, which is really inconvenient. I mean, it's doable, but with just a little bit of massaging that gets it into 16 bits allows a lot of scope for configurability of palettes later.

Link to comment
Share on other sites

  • 2 weeks later...

I did a little work on this, over the last week. As I don't own or have access to an Inty in person, I have been working backwards, while keeping an eye on the auctions for a basically functional unit to snag. 

So, the video output section has been done, and I plan to build it with the R2R ladders for the analog RGB output, based on fake input to the logic block. After that, hooking up the input logic to the output logic is simply a case of making a truth table of which output color relates to which input color. I have a few spare pins, so I will be able to have a jumper for selecting PAL/NTSC palettes. 

When I get that all together, I'll post some captures of the colors and get some feedback.

  • Like 1
Link to comment
Share on other sites

At the moment, I'd take a replacement for the RA-3-9600.  Well, I'm still building a test harness for this unholy mess, but it is such a spider web of functionality that even when I get the test hardware built and programmed, I can't even be sure if the chip is being tested properly.

Link to comment
Share on other sites

I looked at it but it has its tentacles into everything and is an order of magnitude more complex than I want to tackle right now. I just imagine people would be grateful for a simple and easy to install RGB solution. Tackling the STIC would push me out of small CPLD territory into FPGA territory and working out some quite complex logic interactions based on hidden internal registers.

Which, eww, by the way.

Link to comment
Share on other sites

The RA-3-9600 is actually the jack-of-all-trades chip that serves the STIC.  But yeah, I know.  It's about 350 words of RAM, a bus forwarder between the CPU and the graphics busses, an address counter, and a graphics memory tiler and server for the STIC.  The one-trick nature of the TMS9928 is appreciated far more after seeing this.

Link to comment
Share on other sites

Luckily, where I have decided to intervene in the video generation in the easiest spot, in terms of desoldering the smallest IC possible and simplest access to the needed signals. I'm not touching any super hard to find ICs, and not asking people to desolder 40 pin ICs where there's a better than even chance they'll trash their console.

Why does the Inty get no love? I looked around and there's hardly anything for it.

Link to comment
Share on other sites

I keep meaning to post some files here (sorry, too many other things going on), but I do have some logic analyzer captures of the AY-3-8915.  These are from the Sigrok open-source tools.  With a loaner Intellivision on the way, the files might not be as interesting now.  However, if you're still interested, I can post the files here to help out.

 

 

Link to comment
Share on other sites

Not sure if this is useful, but here's the logic analyzer capture file (at bottom) using Sigrok software of the AY-3-8915 (software download at https://sigrok.org/wiki/Downloads).  The screenshot is from the FWDiagnostics ROM since it shows all 16 colors in a simple pattern.

 

FWDiag.gif.935e62e7eb494a3d2d571e5d9c084e91.gif

 

I recommend ignoring the lines in the capture file labelled as RF1 thru RF8.  These are the outputs to the resistors and I was hoping that they would be "logic-like" if the threshold was set right (here the threshold was set at 0.2V).  However, these pins are moderately analog in nature.

 

Color Processor 8915 - LA - RFx at 0.2V with 5V 18KOhm pull-up.sr

Link to comment
Share on other sites

I’ll look at this tomorrow and see if it will speed things up. Any additional info helps as the datasheet is a bit lean. The logic is straightforward, but the way the outputs are used to generate NTSC is, to be kind, obtuse. I understand it after other projects, but still, recreating it in a color system like RGB that has a smaller color gamut, never an ideal situation. 
 

I’ll do the best I can and most people will hopefully be happy. 
 

So many shades of green!

Link to comment
Share on other sites

  • 3 weeks later...

Cool! Glad to see there's more progress being made on different RGB options for the Intellivision! I too have one of the previously mentioned RGB boards made by Fred in my console, which I'm happy with for the most part, except for the OSSC incompatibility. The OSSC is my main upscaler (besides the cheap Chinese one that's only used for the Inty), so I (and many others here) will be thrilled if that issue gets resolved at some point. I'm not extremely picky about colors, personally.. compatibility is more important to me, but that's just me. Nevertheless, I'll be sure to stay tuned to this thread and watch any progress. I'm also up for testing, if need be as well. Thanks for giving the Inty a shot! :)

Edited by SiLic0ne t0aD
.
Link to comment
Share on other sites

I just wanted to drop by and say: I just found out elsewhere that I'm following in the messy tracks of someone called juicebox. I don't know who they are or what they did (though from reading some threads they were a bit... yeah...) but I don't work like they do, I don't have any desire to follow in their footsteps or be tarred by the same brush.

  • Like 3
Link to comment
Share on other sites

I did some work on this yesterday, and made some progress. Enough to define what the sync problem is, and two possible solutions to bypass/fix it.

It seems the most likely assured success will be by using a small microcontroller to count and average the sync pulses, and if one doesn't turn up within 2-3 clocks pf the expected sync, to generate it. This would be 2 or 3 cycles late at 16MHz, so still within the same clock cycle at 3.5MHz. This should be enough to get OSSC to maintain synch in the dead period. This has the advantage that it would work automatically with NTSC, PAL and SECAM.

The other method is simpler and would use a counter and preset value all within the CPLD. This, however, would require specific testing and calibration for each system. What it saves in design complexity and build cost it takes back in lack of resiliency.

The third option is to alter the ROM slightly so the correct sync signals are generated in the first place - but this would require having varying ROMs in different INTYs, and some people might not be happy with that. It's also not in my core skills area.

What do people think?

Link to comment
Share on other sites

I don't think modifying the intellivision rom can help.  The intellivision video signal is created in the stic.  Maybe the ossc firmware can be improved to better handle the intellivision video signal.

Edited by mr_me
Link to comment
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.
Note: Your post will require moderator approval before it will be visible.

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...