Making one's own Supergun: NTSC Video Encoder

[Calpis]

I'm not sure what the NPN circuit looks like, I don't mod consoles and if I did I probably wouldn't use the "usual" circuits floating around. Neither datasheet's app note has an external filter on luma. The '1645 datasheet shows an internal filter on luma, but that's a typo, Y and C are swapped.

[l_oliveira]

What would be the proper way of filtering the luma output ?

By the way, Panasonic did manufacture an fine MSX computer back in 1990 (the FS-A1ST) and it uses an CXA1145. It's the best, crispiest S-Video output I ever seen on any machine. But it's S-Video circuitry is quite complex.

http://msxarchive.nl/pub/msx/mirrors...turborstsm.pdf <- service manual with schematics.

Does that circuit have an suitable filter ?
Do you have any other suggestions ?

[Calpis]

It has a filter, but I'm not sure of its corner frequency or roll off. You may need a steeper roll off since there's still crosstalk. Wait, are you using S-video or composite? You don't want Y filtered for S-video because S-video can take advantage of the extra bandwidth for added clarity. If you get crosstalk in S-video (you shouldn't on the '1645 at least), then the MSX circuit on the '1145 is causing it since Y and C outputs are both coupled to the audio (which I don't get).

[l_oliveira]

By what I understood, it is using the "Audio buffer" circuit (which is simply an generic audio amplifier integrated to the CXA1145 encoder die) to boost the chroma signal it "harvested" from the C loop that goes back to the Y+C mixer which later generates CVBS. Probably they are using that just to reduce the transistor count on the output circuit for C.

My circuit on CXA1145 uses just two copies of the same output circuit using one BC548 for each signal. for the CXA1645 it's the reference circuit from the datasheet.

Well the "wavy" effect was happening on S-Video. It happening on C-Video would be acceptable, while annoying.

Amazes me the fact that it disappears when certain machines connected to the input despites the fact I'm aways using the same encoder board.

I'll try to duplicate the Turbo-R circuitry and see how if affects the video output.

I also found another interesting chip to try, the MC13077, an enhanced version of the ancient MC1377 (it's at least good to run from 5v. MC1377 required 12v, yuck)

[Xian Xi]

I don't know if this will help but I found this image a while back, it's a circuit print for an AD725. You can print it on a transparency and then iron it on a blank board then soak in chemical and have yourself a PCB. Never tried it myself.

[Stone]

The said video encoder (exactly the same thing you saw on the picture) captured on a BT878 capture card (this thing has an better comb filter than my TV as you can see :P)
CPS2:


Do you see any degradation on the image ? The problem is that the TV is not smart enough to "filter the problem" I'm having with the chroma.

That looks fine, so the capture card must be filtering out some analogue crap on the video signal which is throwing off the TV - RGB video is basically just amplified and passed to the electron guns so it'll show up any flaws. A perfect signal will look perfect whatever you plug it into!

Dot crawl is usually worse in s-video than composite - the luminance signal has 4x the bandwidth of the chrominance so edges that don't get 'painted' look awful (and it changes from frame to frame so you get that annoying flicker). The Jaguar had particularly bad s-video output, it showed up most on polygon edges.

I stand by my comment about your construction method (not just the wiring!). Even if you're not going to put it in a box or case, at the very least put it all on one board and insulate it from the floor (!) so you don't get random wire clippings and solder splashes running along the bottom. You're dealing with relatively high frequencies here (as Calpis correctly pointed out - 640x480 has a 25MHz pixel clock!) and even with a properly constructed circuit you can throw it out very VERY easily. My AD724 board, which gives a pixel-perfect image, is so finely tuned that you can alter the colourburst by waving a screwdriver near it. Especially when they're of such a subjective 'it's not quite right' nature and as I've explained to dozens of people your problems will be a million times easier to fix if it's neat and tidy and repeatable. I don't know why I'm bothering tbh, none of the others listened either :DOH:

Get an AD724 and a SMD prototyping board and try it on that. Analog Devices give them away for free, for pity's sake. Sony encoders have always been an arseache, Sega probably had professional design engineers working on their PCB design and the composite output of a Mega Drive still looks gash. If there's a better alternative available, why not try it? I promise you it's worth it.

Stone