Also, looking at the datasheet, do each of the RGB pins that are tapped before going into the AD725 need to be buffered before going into the rest of the YPbPr circuitry?
“If a separate RGB monitor is also to be used, it is not possible to simply connect it to the R, G and B signals. The monitor provides a termination that would double terminate these signals. The R, G, and B signals should be buffered by three amplifiers with high input impedances. These should be configured for a gain of two, which is normalized by the divide by two termination scheme used for the RGB monitor.”
Nice find. I suspect we will be fine here, our op amps are providing the buffering and termination. But the configured for a gain of two part is useful - it tells me the gain of the three ampifiers to the right should remain a gain of 2 as in the schematic. I will add the s-video caps and resistors.
I am knee deep in trying to hand solder SMD LEDs to aluminium PCBs (a complete PITA I do not recommend, even with a microscope and a hotplate to help) but when I have my bench a little clearer I will get the breadboard out
The first one is three of the RCJ-04x’s in a custom holder with screw holes, so the extra Keystone mounts would be necessary. I have not looked at the height difference yet.
The second one is a little different than the RJ-04x’s but with an insertion switch. They are also a few mm higher.
Nice find! I nabbed a few off of eBay that are supposedly NOS but we’ll see when they show up. That IC seems to do a lot of other interesting things I’d like to poke at once I get YPbPr out working.
CSync can be pulled from the sync bus and buffered with inverters like in the Cadet Encoder. 2N3904s will probably be fine to replace the C945 transistors which seem to be obsolete as well.
I also found this PCB looking for more info on the BA7230LS which seems to mostly match the schematic from your link. Getting a few boards printed now to try it out.
Cool, let us know if the ebay parts work.
The C945 has internal resistors, so while a modern replacement should work, you’l need to bias it with external resistors.
In that same thread, there is a link to an OSHpark project another user uploaded.
The parts/PCBs showed up today but I couldn’t get anything other than faint purple nonsense on my CRT’s YPbPr input after wiring it up. I ended up getting some BA7230LS ICs and C945s from UT Source which looked legit to me.
RGB, Sync from U7.4, +5V, +/-12V, and GND are being pulled off of the Cadet II.
I found another thread that ran with this design a bit further which ditches the transistors for a THS7374. Also seems to handle sync in a different way as well.
I was taking a bunch of scope shots to reply to this and realized partway through that the PCB has the C945 transistor pinout wrong. I also wasn’t getting anything on pin 17 (B-Y) and 18 (R-Y) of the BA7230LS but pin 16 (Y) was putting out something so it didn’t seem like the IC was completely faulty.
I replaced the C945s with the leads swapped and that did it! Full color YPbPr output but with some unstable jitter sweeping through a PR’s full range. Looking at the sync implementation in the Aussie Arcade link (Sync injected right at the Y jack thru a 1.2k resistor), I just ran the output of Cadet Encoder’s U7.4 with a 499r resistor directly into the tip of the Y jack and the frame became completely stable.
I just happened to have a 499r resistor handy and I really have no idea about the correct calculations here. Using a lower value resistance just blew the frame out (full white) and higher resistances seemed to ruin sync. Can anyone point me in the direction of the magical resistor values I should use here?
Here’s a clip fading in the RGB channels on CC with vertical bars from PR. YPbPr output from BA7230LS → Intensity Shuttle TB:
On my CRT I can see obvious purple scrolling noise (overexposed example below) whenever my output is black and the capture card seems a bit purply with some random infrequent frame drops, too. I’m guessing I have the sync insertion resistor or mixing wrong.
I have the other PCBs I posted about on the way so I’ll just build one of those up next to try. It’s at least one less obsolete part and the evidence of success seems higher on that PCB than in the other threads. Thought this was promising to see the IC actually works and apart from the purple jitter lines / sync issue, looks pretty good so far!
A great start! Looks like you’re getting to your goal fast.
The purple is very interesting. What makes purple? Is it safe to assume there could be a tiny positive voltage offset on two or three of the outputs which causes black to appear purple?
It makes sense that an offset on both Pb or Pr could cause purple but the effect seemed not as apparent when I removed Sync from Y. I don’t know how that would affect Pb/Pr for where Sync was injected after the colorspace conversion.
I’m just going to wait for the new boards using the THS7374 to come in rather than debugging this one which already has layout errors.
I Like the idea of a replacement for those old transistors, especially since the pcb from Fabio uses unusual resistor values. (EDIT: I was looking at the inputs and outputs backwards but still uncommon values) The board just doesn’t align very closely to the example in the BA7230’s datasheet or other schematics I have looked at. From comments on that forum, it may be a good idea to include a dedicated voltage regulator just for the BA and THS chips; separate from the Cadet II’s 7805 for lower noise.
Yes, I’ve got some Extron VSC 700s which can output RGsB in addition to a few other color space/sync options. I’ll see if I can get anything out of it sometime this week.
I am up way too late, but I just noticed that S-video Y and Component Y are the same thing. Yes, no?
And the AD724 creates an acceptable Y signal which should have sync already. Definitely worth plugging into your tele.
Plugging only the Y (thru a 75r resistor and 220uF cap) into the YPbPr connector gives a stable B&W image as expected. Sorry, these pictures are a bit overexposed but black levels look right to me on my CRT. PR Horizontal bars -> CC -> RGB Encoder.
However, if I turn down the levels of any of the individual RGB channels on CC with only Y going out from the Encoder, I get a stable diagonal pattern that scrolls horizontally to the left overlaid on anything not at 0V. The checkerboard effect is just bad photography on my part–the diagonal lines point from upper left / lower right.
Plugging Pb and Pr from the BA7230LS in, everything looks stable when it’s a fully white signal as above. As soon as I start attenuating one of the RGB signals, the scrolling diagonal pattern appears again. The colors track properly.
If I add/remove the individual buffered RGB jumpers from the AD724 going into the BA7230LS while keeping the RGB knobs fully CW on CC, the colors look pretty close to what I’d expect–maybe a bit brighter but the patterns don’t appear.
When I have a completely black image, I can still see a little bit of purple with YPbPr all plugged in. If I remove Pb and Pr, I can see very faint scrolling horizontal bars. It’s not anywhere near as noticeable as when taking the Y with the mixed in sync from the BA7230LS and doesn’t have that wavy dot pattern at all. Could just be noise/offsets that’ll go away with not having this on a breadboard.
