PlayStation 2 RGB

[RARusk]

If your VGA monitor is a Sync-On-Green type then it should work with Progressive Scan games.

Right now I am writing a guide for "Grand Theft Auto: Vice City Stories" (PS2) for GameFAQs. When I am done with that then I will update my PS2 wiki page (on the GamesX Wiki) with new information gleaned from the service manual and my latest modification work. I am still waiting to see if someone might post the service manuals for the SCPH-50000 and SCPH-70000 models.

[anti_climax]

I've pulled apart my v6 board and traced, with the help of the PS2 service guide wiring diagram, the signals for Red, Green, Blue, HSync, VSync and CSync. This is the top view of the board with the relevant traces highlighted. RGB are self-explanatory, Purple is VSync, Brown is HSync and Black is CSync. The vias above the pull-down resistors (marked 121) tap directly into the RGB signal coming from the GPU.  From there they go through PNP transistors (marked 5K and FQ13) the output of which is attached to the pull-up resistor bank (marked 102) and transferred to the other side of the board. Please be aware that this photo is showing an area of the board only about an inch wide, so finding these particular components on a different rev may prove difficult.

[anti_climax]

Below is the other side of the board showing the analog video encoder chip. The RGB signals are connected through caps (between the vias and the chip leads). Below that we see the vias for the sync signals. Contrary to what Rarusk posted above, there doesn't seem to be a sync mixing chip on this board. The vias for the sync signals are tapped off the lines running from the GPU to the Digital Video Encoder, and the GPU shows outputs for all 3 syncs. The service manual wiring diagram indicates that the chip between the vias and the Analog video encoder just runs each pin through a resistor and then couples each line, through separate capacitors, to ground. Using these features it's likely that you can find these lines on any of the 3xx model PS2s. I would imagine it would be best to grab the syncs right at the vias, and pull the RGB from the vias just before the caps leading into the analog video chip. I'm not sure, but you may need to put caps in line with the wires you run off the board. Perhaps someone else can speculate on this.

[RARusk]

I did find out, thanks to the service manual, that the "sync mixing chip" wasn't a mixing chip at all but appears to clean up the signals a little.

A funny thing about the manual: there is a spot on one of the circuit trace pages where it explicitly marks where you can tap into the R, G, and B signals. But when I tried to tap into them I got nothing, just a solid black screen.

I'm kind of burned out on writing right now but I am formulating how to update my PS2 page on the Wiki. I also need to look out for service manuals for the other motherboard revisions.

[anti_climax]

Feel free to use my images on your wiki entry. I'd have added them in myself if I could have figured out how to upload then to the wiki server. Now that I've ID'd the traces for these signals I'm having my friend solder up leads to each with wifi pigtail coax. that should keep the signals nice and clean. From there I'll likely run those to a DB-15 or RGBHV-BNCs so I can drive my CRT directly from the PS2. I'll likely add my own resistors and cap to the sync lines to clean them up like that final chip, but I'm wondering if I need or want to put caps on the RGB lines. Each is on a pull-up, so there will be a DC bias to them, but I'd guess that the inputs on the monitor would be high impedance and it wouldn't really matter. Any suggestions?

[RARusk]

I don't have any suggestions myself. However, I am currently making new pictures and writing up new text for the Wiki update. I hope to have this done soon.

[RARusk]

I also just noticed that on the picture for the A3525 that there are no traces leading from the north end of the chip. Makes it harder to get at the Sync-On-Green and Component Video/RGB control pins (unless the traces go underneath the chip and come out the bottom).

[anti_climax]

If it's the pin I think it is (Pin 32 "SONGEN") then it's connected to a pull up resistor, which would require it to surface somewhere before the chip connection. On the rear of the board there is a surface mount component that is tied into the same VCC bus as the pull-up resistors for the video signals, which agrees with the wiring diagram. I'll check with my multi-meter when I get the board back from my friend, but I'd bet that's where to grab it if you don't want to solder to the pin itself.

Of course I may not be identifying the the right pin anyway

Just so you know, I was careful to crop these photos so they showed the same area front and back. If you save them locally and reverse one horizontally, you can compare both sides of the board point for point. As for the component/RGB control pin (I'm guessing Pin 28 "RGBYC"), it doesn't have to connect to anything along the way to that pin. You can try the other vias near there but it may not even be accessible from that side. Worst case, someone that needs that signal from this rev would have to tap the pin directly.

I'm glad to have someone who solders for a living helping me with this mod. These pins and vias are actually a step or two up from some of the other things he's soldered. When I get it back from him I'll check with my multi-meter and see if that pin has continuity to any spot on the back of the board.

[RARusk]

SONGEN is indeed the correct pin. On the V4 board the trace is on top. It goes through the resistor and ends at a solder pad. No vias. However, there is a small bit of solder still on the pad leftover from the factory and I simply used it to help connect the wire to the pad. Then connect the other end of the wire to ground to disable Sync-On-Green.

According to the Service Manual, RGBYC directly connects to Pin 267 (GCONT) on the Graphic Synthesizer chip. If you can find that pin, and it has a via, then you might be able to use that as a connection point.

I am currently working on adding highlights and arrows to the pictures I wil be adding to the Wiki update. Then I will finish the text and upload the changes.

[anti_climax]

For my machine, I'm only interested in running my RGBHV monitor off these video signals. Am I correct in assuming that all I need to do is tap these signals and drive them into my RGB monitor, or will I need additional connections to accomplish that (like disabling SOG, etc.)?

[RARusk]

It depends on where you get the video signals from. If you acquire them before the A3525 chip then you don't have to worry about anything else other than possible amplification. If you acquire them from the Audio/Video plug then you will have to worry about Component Video, Sync-On-Green, and MacroVision (and thus needing to disable them).

[anti_climax]

That's what I figured, I just wanted to be sure. I'd hate to have a bunch of fine soldering work done and then find out I forgot several things.

I am considering a switch that would toggle the RGB lines between the A3525 and my new outputs. Having only one set of caps on the line and driving only one display input per line should make the signals a bit more usable.

[RARusk]

I finally finished up my PS2 Wiki page update if you want to have a look.

[anti_climax]

That's a good write up. If you'd like me to snap shots of the relevant spots on my v6 board, let me know.

[anti_climax]

RARusk, I was wondering something. You had said the levels were kind of low tapping off those lines. Did you get a measurement of the actual peak to peak voltages or just a qualitative measure?

I ask because I'm going to be running them into my FW900 widescreen CRT and I know its nominal input is 0.7V p-p @ 75 Ohms.

I'm going to look at the outputs with my O-Scope once I get it back from my friend, but I figured I'd ask while I could.

[RARusk]

I don't have any sosphisticated equipment. I basically poke a wire at a given point and see what pops up on my monitor. Not exactly professional but it gets the job done. Which point were you talking about, Location 1 or 2?

[anti_climax]

Both points really. You had said they were a bit dim before the analog video encoder but perfectly fine coming out. Come to think of it, it may be an impedance mismatch. Being a chip to chip connection I'd bet it's not expecting the 75 Ohm impedance of your monitor inputs, but the output is expecting that. It's no biggie, I'll likely do the mod as you described it on the wiki page. I'll get proper signals, I won't have to load the line with additional caps and it'll make for a little less soldering work.

If I can get the system working and still get to the connections I'll compare the video encoder's RGB in and out and see how they compare on the O-Scope. If they have the same level without anything connected, I'll try running them to ground through a 75 ohm resistor and see if that pulls the inputs down.

[RARusk]

I'm curious about what needs to be done to get a picture from Location 1. I think that information may be quite useful for the other two PS2 models (50000 and 70000 units).

[anti_climax]

How about these points?

[RARusk]

Those points are just beyond the A3525 chip. Attaching wires to these points is the same as if you attached them to the Audio/Video plug.

[anti_climax]

I'd like you to try something if you would. Go to your local electronics supply store and grab three 0.1uf caps and three 250 ohm potentiometers (if you don't have them handy already). Connect the caps to the RGB lines that are going from the GS to the A3525R, preferably at the vias before the surface mount coupling caps, then run the other end of them into your CXA-1145. Then hook the pots to the outputs of that and dial them all to zero. Hook each monitor pin to the center pin of the appropriate pot and run them up until each color is the right brightness.

Since the CXA-1145 is over amplifying the signal and the impedance of the monitor is constant, you may be able to get away with just splitting the signal voltage between it and a serial resistance. While the voltage will be right, the amount of power going into the inputs of the monitor will be reduced. I'm not sure if that's going to be a problem. If that doesn't do it, pots on the inputs of your CXA-1145 should, but I'm not sure what they'd have to be rated at to work right.

[patch6]

anti_climax: You could try bridging the capacitors that are connected in serial to the RGB outputs on the A/V plug, if the image is too bright.

Now, has anyone found the RGB and Hsync/Vsync points for the V7-V16 models yet?
I have a gutted V7 and V4, with the GS chip desoldered on both to find out where an RGB signal can be tapped, but the pin designations seem to have been changed after V6 (which has the same designations as V4), so using traced points for those pins on the other side of a working v7 doesn't work.

A point that can be grounded to disable sync-on-green for the V7-V16 models would be good, too.

I might have to get an oscilloscope to figure this out.

[NightSprinter]

Sorry to do what's a necropost for my first time here, but I believe this would be pertinent for those of us that own SCPH-39000 series units.  I have a PDF of the service manual, which apparently has easier-to-access points for RGB and sync.  I sent a PM to RARusk already, but I can happily provide the file to those that would need it.