Jump to content

Photo

Intellivision ntsc rgb mod design help

intellivision

59 replies to this topic

#51 emerson OFFLINE  

emerson

    Chopper Commander

  • Topic Starter
  • 193 posts
  • Location:Northeast Ohio

Posted Tue May 7, 2019 5:00 PM

 

Good to hear. I will be working on a similar project in the near future using different components namely a FPGA, ADV7125. 
Anyway I don't know if I skimmed over it but have you tried taking sync from the bits 1 X X 0 1 instead of wherever you are taking it ?

I just wonder if one is better than the other; it could possibly save me a lot of grief in the future.

 

I think in general using C-sync instead of sync on xxx is better. Also if you use those bits you could use the THS7374 on the outputs (SRGB) to easily toggle C-Sync 75 or C-Sync TTL with just resistors and jumpers.

The lm1881 can then be done away with entirely. I can go way more into detail if needed.

 

When you say "bits 1xx01" I assume you mean V5-V1 from the AY-3-8915 correct? Looks like it based on the datasheet. I decided to pull sync directly from the composite video signal in an attempt to ensure compatibility with the OSSC. Fred's mod seems to derive it's sync signals based on V5-V1 but somewhere there is a timing issue. My circuit monitors V5 and blanks the screen whenever it goes high. It does not specifically look for the "1xx00" blank condition from the AY-3-8915 truth table.

 

 

Also I dunno if it helps but the external video is half the length of RGB on the same rising edge (I believe)., which is specific to the intellivision 1. I have no idea what the others do.

At any rate that signal should be a square wave input that is 1111-1111, 255, aka white. I don't know anything about the device you are using or programable circuits in general (yet).

 

Not to be rude but I have no idea what you're talking about here. Can you please elaborate?



#52 the_crayon_king OFFLINE  

the_crayon_king

    Chopper Commander

  • 189 posts

Posted Tue May 7, 2019 7:50 PM

 

When you say "bits 1xx01" I assume you mean V5-V1 from the AY-3-8915 correct? Looks like it based on the datasheet. I decided to pull sync directly from the composite video signal in an attempt to ensure compatibility with the OSSC. Fred's mod seems to derive it's sync signals based on V5-V1 but somewhere there is a timing issue. My circuit monitors V5 and blanks the screen whenever it goes high. It does not specifically look for the "1xx00" blank condition from the AY-3-8915 truth table.

 

 

 

Not to be rude but I have no idea what you're talking about here. Can you please elaborate?

 

Yes blanking on 1 is a must. I don't get what you mean by composite ? Would that be a mod kit or do you mean just the ORing of H and V sync aka C-sync?

Looking at the datasheet H and V sync have different values for the resistors. I would expect them to be the same value but idk.

You could try ORing those two for C sync. those outputs ORed should be the same functionally as the output of the lm1881 but much cheaper.

Anyway I don't understand why a version using the bits would be worse. Ill find out soon enough. Thanks.

 

There is another video input signal on the intellivision 1 for the keyboard component.

From the AY-3-8915 Datasheet:

 

The required color to be displayed for each 280ns PIXEL is
decoded on a four line binary coded input. This selects one of
sixteen possible colors. An external resistor network completes
the D to A function as shown in the schematic of Fig. 1. The
waveform pi us table illustrates the use of the five inputs to produce
composite sync, color burst, line blanking, frame blanking and
video.The external video input pin provides the ability to superimpose
white high resolution (140ns wide) video information over the
picture (color image).

 

By 1111-1111 I mean 714mv or 1.428v (unterminated) meaning that external video input would be that same as white but at half the length. Aka 0111. V5=0 V4=1 V3=1 V2=1 V1=1

I don't know the limitations of what you are using but a FPGA would be able to superimpose by ORing say 1111-1111 and 0000-0000 (or any other value) across all rgb lines for 140ns.

Part of the confusion is probably that im dealing with the decimal or binary you are probably looking more at the voltages.



#53 emerson OFFLINE  

emerson

    Chopper Commander

  • Topic Starter
  • 193 posts
  • Location:Northeast Ohio

Posted Wed May 8, 2019 6:21 PM

When I say composite I mean the composite video mod in my circuit. I too would expect the hsync and vsync resistors to be the same. Not too sure what's going on there either...

 

I didn't mean to imply ORing hsync and vsync is worse than pulling sync from the video signal. All I'm saying is the firmware in Fred's rgb mod doesn't work with the OSSC so I decided to take a different approach.

 

My circuit works by taking the 5bit code on V5-V1 and expands it to 9 bits, or 3 bits for each red green and blue and runs that through a 2:1 resistor ladder. The bits and voltages aren't the issue. The problem is I forgot about the external video input on the cartridge port. I honestly couldn't say how my circuit behaves with it since I don't have an ECS to test with. Based on the datasheet description it seems like the AY-3-8915 superimposes the external video for you. How else would the external video get placed on V4-V1? My guess is the AY-3-8915 generates a 140ns "0111" state on V4-V1 when the external video pin gets toggled but this is just speculation. I wouldn't think the external video input would be analog, especially just for one color. It would have to start as analog, get converted to the 4bit color code then back to analog. That seems a little backwards but I could be wrong.

 

I'm using a GAL22V10 which is an older programmable logic chip. In the programming software I was able to set up a truth table of color code inputs and decoded outputs so hopefully you can do something similar with your code. It makes adjusting the color palette a breeze. Since you're going with a 24bit rgb color code you'll be able to directly port color palettes from jzintv which will be cool. Do you plan to have a program header so people can customize their palettes?



#54 the_crayon_king OFFLINE  

the_crayon_king

    Chopper Commander

  • 189 posts

Posted Thu May 9, 2019 12:36 PM

When I say composite I mean the composite video mod in my circuit. I too would expect the hsync and vsync resistors to be the same. Not too sure what's going on there either...

 

I didn't mean to imply ORing hsync and vsync is worse than pulling sync from the video signal. All I'm saying is the firmware in Fred's rgb mod doesn't work with the OSSC so I decided to take a different approach.

 

My circuit works by taking the 5bit code on V5-V1 and expands it to 9 bits, or 3 bits for each red green and blue and runs that through a 2:1 resistor ladder. The bits and voltages aren't the issue. The problem is I forgot about the external video input on the cartridge port. I honestly couldn't say how my circuit behaves with it since I don't have an ECS to test with. Based on the datasheet description it seems like the AY-3-8915 superimposes the external video for you. How else would the external video get placed on V4-V1? My guess is the AY-3-8915 generates a 140ns "0111" state on V4-V1 when the external video pin gets toggled but this is just speculation. I wouldn't think the external video input would be analog, especially just for one color. It would have to start as analog, get converted to the 4bit color code then back to analog. That seems a little backwards but I could be wrong.

 

I'm using a GAL22V10 which is an older programmable logic chip. In the programming software I was able to set up a truth table of color code inputs and decoded outputs so hopefully you can do something similar with your code. It makes adjusting the color palette a breeze. Since you're going with a 24bit rgb color code you'll be able to directly port color palettes from jzintv which will be cool. Do you plan to have a program header so people can customize their palettes?

 

I had planed to bypass the AY-3-8915 entirely if possible is the only reason I asked.

 

I understand what you are doing now.

 

I haven't looked at the signal personally I just assume it has to be digital since it should only have two states on/off.

 

I haven't thought too far ahead on any board specifics. First I would be building the code; this would be my first foray into programming at all.

I do deff plan on having jumpers to toggle between RGB and component video. If able I'd have multiple pallets set with jumpers.

 

What I was assuming about the ext video input is that you can clock it at half the clock of RGB then OR those two together.

Say RGB is black 0000-0000 and ext video is white (on) then you end up with two signals 0000-0000 280ns and 1111-1111 140ns

when ORed I thought it should be just like the datasheet specifies.

 

You should be able to get away with 1 clock if you can specify length in the code somehow. Like if you trigger off the 280ns clock but only have that portion on for 140ns. Im really not sure if that's possible. Ill make my own thread sooner or later.

 

Here is a sample of my truth table. I don't want to paste the whole thing as it would take up alot of the screen.

I didnt put alot of time or thought into decimal to analog voltage value since its not specifically relevant to me besides testing.

 

V5,V4,V3,V2,V1
================================
red states
000 0000-0000
128 1000-0000
203 1100-1011
255 1111-1111

red for example can only be:
255 - 714mv
203 - 571mv
128 - 357mv
000 - ground
================================
green states
000 0000-0000
064 0100-0000
128 1000-0000
159 1001-1111
255 1111-1111

green for example can only be:
255 - 714mv
159 - 446mv
128 - 357mv
064 - 178.5mv
000 - ground
================================
blue states
000 0000-0000
101 0110-0101
128 1000-0000
255 1111-1111

blue for example can only be:
255 - 714mv
128 - 357mv
101 - 287mv
000 - ground
================================


Edited by the_crayon_king, Thu May 9, 2019 12:44 PM.


#55 hitokage OFFLINE  

hitokage

    Combat Commando

  • 4 posts
  • Location:Frederick, MD

Posted Sun May 12, 2019 1:50 AM

I wanted to try to give some kind of help with this and hopefully it makes some sense. First, I don't have an ECS myself, but based on what I can find online it doesn't seem to do any video generation itself. So, it should work with this mod or any others without any trouble.

 

The external video input pin on the cartridge port seems to have been used for two other things though. The first being the "Keyboard Component" and the second being the "System Changer".

 

From what I can find the external video input as used for the keyboard component, used genlocked video - meaning it was a composite video signal that was synchronized to the Color Processor Chip. If this is the case, there isn't a good solution to make it work.

 

The system changer also used composite video, but wasn't genlocked (so they also changed which pin on the cartridge port was used for external video). That's why the system changer only works with the Intellivision II, unless the Intellivision (all others) itself is modified. There is not a way to make this work without modifying the system changer. Since the system changer is just a 2600 that plugs into the Intellivision, with the Intellivision not really doing anything, it probably isn't worth trying to make it work.


Edited by hitokage, Sun May 12, 2019 1:51 AM.


#56 emerson OFFLINE  

emerson

    Chopper Commander

  • Topic Starter
  • 193 posts
  • Location:Northeast Ohio

Posted Sun May 12, 2019 8:31 AM

Thanks for the info hitokage. So the AY-3-8915 does convert an analog composite video signal to the 4bit digital color code? I wouldn't have guessed that.



#57 hitokage OFFLINE  

hitokage

    Combat Commando

  • 4 posts
  • Location:Frederick, MD

Posted Mon May 13, 2019 1:04 AM

I don't think so, it just switches to the external video source like a AV switchbox. Genlocked video could get switched seamlessly between internal and external composite video - like a video switcher used in a broadcast studio (the screen refresh is synchronized). Non-genlocked video, like used with the system changer, would cause some distortion when switching between the two signals - like a basic mechanical switchbox.

 

Here is a link to the Intellivision Wiki that describes how to modify an Intellivision (any non-Intellivision II) to use the system changer - http://wiki.intelliv...rt_Modification

 

This page is where I found the statement about the pin 8 of the cartridge port (external video) being  genlocked (it's near the very bottom) - http://spatula-city....ech/master.html


Edited by hitokage, Mon May 13, 2019 1:05 AM.


#58 hitokage OFFLINE  

hitokage

    Combat Commando

  • 4 posts
  • Location:Frederick, MD

Posted Mon May 13, 2019 1:59 AM

To expand a bit more about what I said with the ECS working with any video mod - I think it works like a cartridge, with extra some hardware attached to it - like the Intellivoice. From what I understand of the ECS is that it doesn't have any built-in video chips, so any video/graphic type stuff is generated by the STIC, and doesn't use the external video pin.

 

The system changer on the other hand does have its own video generation chip and outputs a composite video signal, as it's almost a complete 2600, and that does use the external video pin.

 

To try to explain this better/clearer because sometimes I'm not very good explaining things - anything using the external video pin on the cartridge port (8 or 2) is sending composite video on that pin. The Intellivision just switches to it and doesn't do any processing on it at all.



#59 mr_me OFFLINE  

mr_me

    River Patroller

  • 4,187 posts
  • Location:Ontario

Posted Mon May 13, 2019 3:44 AM

Thanks for the info hitokage. So the AY-3-8915 does convert an analog composite video signal to the 4bit digital color code? I wouldn't have guessed that.

The ay-3-8915 doesn't ouput the 4-bit digital colour code, it receives it as input from the STIC. The ay-3-8915 converts that digital code, which has a fifth bit for sync, blank, colour burst to a semi analog composite signal.

#60 emerson OFFLINE  

emerson

    Chopper Commander

  • Topic Starter
  • 193 posts
  • Location:Northeast Ohio

Posted Tue May 14, 2019 4:58 AM

Wow, I'm not sure how I got stuck on that idea. It happens sometimes I guess...







Also tagged with one or more of these keywords: intellivision

0 user(s) are browsing this forum

0 members, 0 guests, 0 anonymous users