CRT to RGB Conversion (help needed)

[EDIT: need help still for this TV, but see Post #18 for the mod on a different 9" CRT) ]

Hey all,
So I've converted a standard US consumer-grade CRT to RGB before many years ago and am trying to do it again on a cute little 9" CRT I found this week. The small ones are hard to find (with color tubes) and this one is old enough that the OSD input (RGB) is going to still be analog. It's also old enough that the main jungle IC doesn't use I2C programming. It's a 1993 Zenith TV.

I followed the excellent guide Viletim posted here:
http://nfggames.com/forum2/index.php?topic=4159.msg42011
(post reply #2)

only difference my RGB inputs appear to be DC-coupled with no caps going in. I made a voltage divider for the sync (pin 45) which works perfect, and the RGB shows up on screen OK it just lacks a lot of color gradients, especially pink and skin tones (see attached pic)

I'm not sure what would cause this...it's almost like the jungle IC doesn't process the signal enough because it's only using it for OSD. Does anyone have any ideas? I've tried variable resistors on the RGB signals and various values of caps in case it was an AC-coupling issue. The datasheet seems to show it as DC-coupling as well in the pin description. The OSD microcontroller is unknown and The jungle IC is a TA8879N and the datasheet is over here:
http://www.alldatasheet.com/datasheet-pdf/pdf/31572/TOSHIBA/TA8879N.html

 

Well, I no nothing about this mod, but you might get some interesting here or able to ask Shadow directly.
http://www.psxdev.net/forum/viewtopic.php?f=62&t=723

 

Well, I found a service manual! (attached) It's not the best but it has some useful info.. Hard to track down since there is no model number written on this thing, I just searched some markings off the chassis (like the flyback). It turns out this is a Zenith SMS2049.

There are 47pf coupling caps for all the RGB and Ys lines as well as 2k7 ohm in series resistors.

The datasheet for the TA8879 also suggest applying voltage to pin 33 (video mode select) when measuring the OSD output. I'm wondering if perhaps my problem is from this chip being in TV Tuner mode instead of Ext Video In mode? I was thinking with Ys tied high (fast blanking) that this would be irrelevant, but maybe it still is.

On page 40 it suggests applying 4.3v to pin 33. This is confusing though since on page 9 it shows 3.25v is TV mode and 2.15v is external mode...

 

That jungle I/C isn't configured to accept .7vPP analog RGB inputs. It has quite a large swing, let's take a look at the data-sheet.



I believe it's also missing a clamp on those inputs, too. Are you matching impedence? (Terminating with 75 ohm resistors to ground to match 75 ohm load of signal)

That's quite a swing, you have a gain problem. Shouldn't be too difficult to dial that in.

 

Thanks so much for your help. Any idea on the best way to clamp the signals? I dont know much about it. I can just up the gain with a TH7314 or similar, yeah? or a quad opamp. Im just surprised that would cause missing gradients instead of just a low-contrast image. And yep, I did 75-ohm termination. Since its DC-coupled I should not use caps for the input, right?

 

"Threshold voltage" in the data sheet snippet seems to suggest to me that the pins may be digital inputs instead of analog - this would allow only 8 colors in total. Good enough for an OSD, but not good enough to feed a console's RGB output into it.

 

Unseen is correct.

That's TTL RGB. I failed to see the truth table for color output combinations in the datasheet.



You have a 3 bit color palette there.

You'll need to take your RGB source directly to the guns and drive it there.

 

Bummmmmmmmer. Nuts. Well, I'll have a look at giving it to the guns then, thanks to Armorant's link. Looks iffy and a bit tricky, since service manual says is sends 5.1v signals to the tube. Blagh, these jungle ICs are so hit and miss.

-Segasonicfan

 

That partial schematic printed next to the pin description makes it pretty clear, too - you basically have a long-tailed pair differential amplifier with a reference voltage on the other input, which is a classic comparator setup.

On top of that, looking at the service manual it's a hot-chassis design, so potentially hazardous to try and directly feed RGB into anyway.

 

Could you say a little more about this? What makes it a "hot" chassis?

 

Something has a hot chassis if there is no electrical isolation between the circuit common and the power line. If you look at the first page of the schematic, you can see that the power line comes in, goes through some EMI filtering components and a relay then into a bridge rectifier (comprised of CR3406 to CR3409) - the -ve output of that rectifier is then connected to the circuit common - so there is a low-impedance path between the power line and the circuit common.

If you look on the part of the schematic that deals with the AV inputs, you will also notice that they are fed through optocouplers - this is another strong indication that it's a hot chassis design. In fact, the whole input circuit has it's own separate common and power rail that's derived indirectly from the flyback transformer through an isolation transformer (TX9300 in this case). The only direct electrical connection between the AV input circuits and the rest of the TV is high-impedance - consisting of a 5M6 resistor and a 4.7nF cap in parallel between the AV ground and the circuit common.

 

wow, thanks so much for explaining all that! Thats nuts...and sounds definitely dangerous for hacking in RGB to the neck board. sounds like I would be better off to test a signal from a system running off of batteries if I do it. Or invest in an isolation transformer.

 

Your main problem is the design of the jungle IC and the age of the TV. My thread demostrates a successful attempt with a 1999 Samsung with isolation transformer, switch based power supply, and analog RGB inputs as outlined in the service manual. This config hits paydirt.

http://assemblergames.com/l/threads/crt-tv-rgb-input-hack-attempt.62175/

 

Yeah, this CRT isn't so great for hacking, but I must say, I don't think the age has anything to do with it. I've successfully hacked a 1994 JVC CRT for RGB before because it met all the proper conditions (manufactured the exact same year as this Zenith TV). I think the problem lies in this being a portable 9" CRT. So it's less likely to support analog RGB because it will have a very cheap Jungle IC and it will also have a cheaper power supply.

Likewise, I have a Sony CRT from 2000 that is not hackable for RGB either...it has analog RGB inputs (digital) for OSD and even analog ones, but it can only switch to analog mode via I2C bus that the jungle IC is programmed to use. So it would take an awful lot of work to make it RGB (which is a shame, cause I love it).

So yeah, I think it's really hit and miss with these jungle ICs and other configurations..

-Segasonicfan

 

I've laid down quite a bit of information and instruction in a thread over at Shmups:

http://shmups.system11.org/viewtopic.php?f=6&t=56155

Could be helpful, I cover quite a bit there.

 

Not really "an awful lot of work". You could do it in an afternoon.

Capture the I2C packets with a logic analyser (24MHz 8-channel Saleae clone, $10 off eBay), program a small microcontroller (ATTiny or PIC) to replay the captured packets, done.

You'd essentially be building a mod chip for your TV!

 

Damn, what a freaking awesome idea! This stuff is a little over my head but I do grasp the concept. I've never messed with a logic analyzer or programmed chips before..

However, the GOOD NEWS is that the monitor actually doesn't need the I2C programmed differently for RGB support! The chip was in RGB mode all along (even though it is unused). I was confused because the OSD uses digital RGB and there are 3 blanking pins, but I figured it all out last night!

The only thing left is getting this darn thing to take my TTL C Sync from a Sega Genesis or other system. It works perfect with composite video on the input but my RGB cable doesn't support that and I use TTL Sync for everything else (plus, my Playchoice RGB NES doesn't have composite video). I've tried a bunch of circuits to no avail..... Here's the datasheet of the Jungle IC:
http://www.alldatasheet.com/datasheet-pdf/pdf/118636/SONY/CXA1464AS.html

Does anyone know how I could get 5v TTL Sync into this thing? I've tried a voltage divider to drop it down to the 2V p-p that the datasheet asks for on the Y input on pin 3 (using 10k in series and 10k to ground). Also tried various capacitances and nothing seems to work. I even tried adding C Sync to the junction point before the H and V inputs (see last page) with no luck. ...
EDIT: I figured it out. I had to find the right point to input the Sync by isolating the transistor base and feeding it only to the H+V circuits. I will post lots of pics and info soon!

-Segasonicfan

 

All done!!


PICTURES:
http://s18.photobucket.com/user/Segasonicfan/library/CRT Mods/KV-9PT60 RGB mod and better S Video?sort=3&page=1

It looks *fantastic* the video and pics dont do it justice.
I hacked the internal switching MUX for swapping S-Video and composite instead of Cable / Video. Then I wired the MUX (74053) select pin to pin 12 of a VGA plug I use for RGB so it grounds when connected (with Ys having a pull up resistor to +5v) That way the entire thing is switchless
I did have to add another MUX for the Sync though, since there was no easy way to inject it into the mixing circuits. So when Ys is pulled high it also disconnects the Y input to Sync circuits and instead grabs it from C Sync.

and for everyone out the using TTL Sync, a great simple voltage divider that has worked for my CRT hacking is:

TTL Sync -------4k7----4k7--------- Sync Out (0.7V p-p)
with a 1k to ground at the end of the 2x 4k7 output.

I also did a bunch of hacking to the audio since the built in amp was absolute trash. It now has a decent class D amp and the stereo audio is properly mixed with an opamp buffer stage as well. The headphone switching was a pain but I figured that out too.
This thing runs off 12V too. the perfect mini CRT I've always wanted

-Segasonicfan

 

Nice work Segasonicfan!

I have a JVC A140-PN that only accepts Composite and S-Video. I would like to see if it's possible to RGB mod it. I've opened up the case and found the "jungle chip" and it's a M37212M6-121sp and the pinout of that chip looks like the attached file.

Do you think it's possible to RGB mod it?

Thanks

 

Oh, I just got hold of the service manual and the Video processor is a Toshiba TB1226BN, see the attachement.

Analog RGB IN. Cool!