Magnavox Odyssey 2 RGB Mod. Anyone care to try this ?

[the_crayon_king]

Well I have the mobo to a odyssey 2 but not much else.

 

I think you can get rgb just with 75 ohms out on the following pins on the 612160

 

01- 5V

08 - Sync (Hsync I think)

10 - Red

11 - Luma*

12 - Blue

13 - Green

15/16 - GND

 

 

Those are inputs from the 612130-1

So Optimally I would desolder the 612160 and use the holes so as not to have interference from RF

 

*Luma is defined further in the PDF for the 612130 as:

The Luminance output represents the ORed result of active patterns in the minor system, the major system, and the grid (if set grid bright is active)

8245.pdf

Edited May 28, 2017 by the_crayon_king

[the_crayon_king]

So apparently it has been done successfully before. http://atariage.com/forums/topic/260238-o2-rgb-mod/

I gleaned info form there; this is my spin on it: https://easyeda.com/hotdog6394/Magnavox_Odyssey_2_RGB-45247cce47a741fc80aaa742a76160a9

credit goes to grips03 for the info. This is basically a SMD version of his board.

 

 

I'd really like to do away with U1 which is the 7407N

I did away with inverting the sync signal, I don't think it needs to be done but I could be wrong. That being said I used a single channel inverter (U4). If sync needs to be inverted Ill use a dual channel inverter.

 

If I could get proper impedance matching out of this by finding an alternative to the 7407N I would be really happy.

Kinda holding off on making markings until I can (hopefully) find a replacement for the 7407N.

 

 

 

[the_crayon_king]

Possible improvement ?

 

LRGB to Hex 3-State non-Inverting buffer,

Non-Inverted [blank] to enable pin.

L to 3 input pins,

RGB to other 3 input pins,

 

Sum L+R, L+G, L+B

 

Voltage divide RGBS

Amp RGBS with THS7374

outputs should be RGB 1.4v

75 ohm series on all RGB lines to divide to 0.7v

 

It could be better to sum L(uma) in the beginning.

 

Im not sure if any of this is correct thoughts ?

Edited May 29, 2017 by the_crayon_king

[the_crayon_king]

This should work...Ill have some demo boards out soon.

[the_crayon_king]

Im just going to leave this here:

https://easyeda.com/hotdog6394/Magnavox_Odyssey_2_RGB-45247cce47a741fc80aaa742a76160a9

 

Jaggies and ghosting in screen capture seem to be an artifact of the capture devise cannot see them on my TV.

No blanking has been applied (on this capture). Not too sure if the method I used to apply blanking is even a correct way to do it,

Currently has color bleeding. Suspect over-saturation from luma being too high.

No sense in posting a schematic yet as Im still tweaking.

This board should be socketable in most NTSC O2 Models top or bottom after removing the 612160 encoder. (these screen still have the encoder and instead are wired into the holes)

Final design will have audio components.

Let me know if anyone has any input/questions.

 

 

[Ikrananka]

Any update on this?

[the_crayon_king]

Any update on this?

 

https://easyeda.com/hotdog6394/magnavox-odyssey-2-rgb-final-version

I tend to jump around from one project to another project.

This way would require 3 ics. one inverter, one triple op amp (with high slew), one buffer with OE set active low.

 

Im pretty confident that the output section will work fine eg. http://tinyurl.com/y3dm5bah

There are only 16 possible colors on the O2 and it may not be correct but I derived a formula from here http://videopac.nl/forum/index.php?topic=1631.0

255* X = 714; x=2.8

182* 2.8 = 509.6 mv

73* 2.8 = 204.4 mv

Ill have to do more research on the input section; I may end up using everything before the diodes and pull-ups from the existing mod.

I don't know if the end result will be better or more correct. I just do things my own way.

Edited April 28, 2019 by the_crayon_king

[the_crayon_king]

 

Thats the colors I have so far. Keep in mind its going through a XRGB and a screen capture program.
https://i.imgur.com/Dzdrjap.jpg

https://i.imgur.com/rXtruhb.jpg

https://i.imgur.com/aJyYNRr.jpg

https://i.imgur.com/OpFr1u6.png

 

What I need is more specific value for everything this just shows it can be done with just a few ics.

https://i.imgur.com/6zZejv3.png

Edited April 30, 2019 by the_crayon_king

[the_crayon_king]

I really dont want to keep double posting. Feel like im speaking into the wind.

Anyway I had a broken NES RGB board laying around so I was able to use the BH7236AF to get Y/C and CV out.

 

For my uses this means not only can you get RGB, S-Video, and CV. Having Y (from s-video) means you can easily get YPbPR (component) from mixing the RGB out of the BH7236AF encoder.

Ill make two versions one that uses the BH7236AF to get all the video formats out and another that has RGB and Component only but uses newer more reliable to source components in case those BH7236AF encoders dry up.

 

One specific thing I noticed was a noticeable difference in the color output compared to RGB. Just a note. Sometime after the 8th I will have a batch of prototype boards.

This abomination is what I've been testing on https://i.imgur.com/jJRERMv.jpg << That is not what the end board will look like at all., the end PCB will only be around 2x2 inches or smaller.

 

Image of this mod through component :

https://i.imgur.com/9U9hKsL.jpg

https://i.imgur.com/Rjk75Iz.jpg

https://i.imgur.com/WupmnLS.jpg

 

Keep an eye on

https://easyeda.com/hotdog6394/magnavox-odyssey-2-rgb-final-version

and

https://easyeda.com/hotdog6394/magnavox-odyssey-2-rgb-unclked

Edited May 3, 2019 by the_crayon_king

[Mikebloke]

I'm a PAL user, so I have a Videopac instead of a Odyssey 2.

 

I'd be interested in a RGB/Component mod to future proof the machine as its something I really enjoy.

 

I have a feeling though from composite mods that is out there that it would require slightly different configurations than the US version (however, my knowledge on what that is, is poor).

[the_crayon_king]

I'm a PAL user, so I have a Videopac instead of a Odyssey 2.

 

I'd be interested in a RGB/Component mod to future proof the machine as its something I really enjoy.

 

I have a feeling though from composite mods that is out there that it would require slightly different configurations than the US version (however, my knowledge on what that is, is poor).

 

I'm like 80 percent sure that the RGB portion will work on both PAL or NTSC since those outputs are already clocked by the O2.

The expansion board shown below needs configured a little bit differently for PAL or NTSC. I did not make toggles or anything like that. I will in the future.

I'm not up on my PAL knowledge but if its just a matter of clock speed then everything should be the same (except for a capacitor and an inductor).

 

These boards are much smaller than they look.

 

Main board: 1.9x1.1 inches

 

expansion board: 2x1.4 inches.

Edited May 9, 2019 by the_crayon_king

[the_crayon_king]

I made a video but showing three video outputs. My setup is a pain to use so don't hate.

Also I SUCK at this game. There are vertical jaggies on the video that I can't see with my eyes when playing on the TV but show up in the recording.

There is also some haloing with blues:

 

 

Installed:

 

https://i.imgur.com/2BO0juK.jpg

Low Saturation

  Medium Saturation

Full Saturation

 

RANDOM NOTES:

The video is currently over saturated which is what these scope readings are showing.

The ratio between low sat and medium sat is what matters most. The sum can be easily adjusted with the pulldown being higher/lower.

It may be best to tie a pot between the pulldowns and ground so saturation can be adjusted manually.

The ratio I am using is supposed to be 2.5. That is LS*2.5=MS., and MS/2.5=LS. MS+LS=FS

LS should be around 408

MS should be around 1.020

FS should be 1.428

The reason I don't have closer numbers is because I do not actually have the specific resistors I need and had to daisy chain 5% tolerance ones.

 

I need to change the capacitors and voltage regulator so some video aberrations could be from that.

I played this on my CRT through composite and S-Video and it looks a 100 times better than through my scaler (XRGB).

I will try component later but I would guess it will only give a slight improvement over S-Video.

Here is showing my CRT:

 

 

[Mikebloke]

Looks good! I think for me component is slightly more future proof, although even that is appearing less and less on new tvs. S video never caught on here in Europe and generally RGB scart was preferred, but again, find a new tv with a scart port. 

 

Also, the C7051? That's another thing on my wish list! 

[SSG]

16 hours ago, Mikebloke said:

Looks good! I think for me component is slightly more future proof, although even that is appearing less and less on new tvs. S video never caught on here in Europe and generally RGB scart was preferred, but again, find a new tv with a scart port. 

 

Also, the C7051? That's another thing on my wish list! 

S-Video never fully caught on here in USA either. I have never seen a S-Video capable TV or VCR. Only device i saw with S-Video was a input switcher.

[the_crayon_king]

On 7/2/2019 at 7:24 AM, Mikebloke said:

Looks good! I think for me component is slightly more future proof, although even that is appearing less and less on new tvs. S video never caught on here in Europe and generally RGB scart was preferred, but again, find a new tv with a scart port. 

 

Also, the C7051? That's another thing on my wish list! 

 

I can't get this to work on a modern TV without going through a scaler. Even if you could get straight component to work on a newer LCD TV it would look bad because the TV can't scale cleanly. I did get composite to work on a new TV but it looks terrible (poor scaling).

 

The issue I was seeing is pixels tearing left or right from where they are supposed to be., but only on specific colors.

There is also a sawtooth type pattern that was reduced by replacing either C49 or C28. I am not sure which did it.

However I still see the remnants of that pattern on my no name video scaler VS my XRGB looks fine besides that tearing.

I tried a mixture of my mod and the G7200 mod. I tried different setups for the THS7374. I tried adjusting the clock of the console.

All methods seemed to be the same on my TV.

(It's important to note I also have a different console and it had the same issues.)

 

It's some sort of delay or timing issue. Also its interesting that the tearing and sawtooth pattern do not show up on my CRT even right next to the screen.

Some improvement may be had with using a different method to sum the colors. The way I am doing is just cheap and easy but by no means perfect.

I am curious if using a D-type flip flop with the NSTC clock would repair any of these issues. The other idea would be using an normal op amp instead of the THS7374.

I am open to ideas.

 

What I have laid out is basically a 'passable' way to get video out to a CRT in one form or another. I can button this up inside my console and live with it as is for my uses.

 

I have written the code to do this same kind of mod but through a FPGA the problem there is I need a few things I don't currently have on hand.

I will continue to tweak what I have but the end goal is a FPGA based mod which in alot of ways is easier. With a FPGA I  could scale or re time the signals easily.

Also it could be sent through HDMI. 

 

I also did some other things to get a nes controller and atari controllers to work on this console. You have to mod the nes controller but the atari can be used as is.

Edited July 4, 2019 by the_crayon_king

[the_crayon_king]

Not to necro bump but there has been significant changes how I am doing this.

For one this is using a CPLD and DAC instead of a buffer, inverter, and opamp.

The problem is that the colors can now be whatever I want so that leaves a great deal of indecision. I used two completely different palettes but besides yellow/orange the difference is subtle. The settings on my datapath (capture card) were not 100 percent correct; thus cyan looking so weird. If someone knows of better palettes let me know.

 

Another weirdness is the video gets glitchy if you use the 3.58MHZ clock instead of the 7.19mhz clock (I had issues with the Aquarius in the same way).

I may attempt to buffer the 7.19MHZ clock since it is at 600mv low and 2.4V for high. I need TTL levels of below 800mv low and above 2V for high but cutting it so close makes me nervous.

The video is still not 100% perfect but it might be as good as it can get without taking the inputs from a source further up (the ppu inputs). I will learn more after I look at this closely under an oscope. This was just the initial installation I will hopefully be able to figure out any weirdness and be done with this.

 

On another note the CPLD version is 20-30mv lower (darker) than it will be in the end as I am currently using the wrong resistor.

 

Picture below:

 

 

[the_crayon_king]

Ok, so the newer mod kit has finally been tested against my older mod.

The sawtooth pattern doesn't seem present (at least on the OSSC). It might still be there.

I don't anticipate making any changes unless someone comes forward with some specific color pallets.

I will put these up for sale soon. I don't know how these will stack up against any of the other mods out there.

 

Image looks a little soft. There is nothing I can do besides tweaking the OSSC settings to make it look any better.

It should be noted those red specks in the gameplay are in the game not some kind of glitch on the part of the kit.