ACTIVE SCART BUFFER V3 from JS Technology issues

[Strider77]

I bought an Active Scart Buffer (ver 3) from JS Technology to run my RGB sources to both my PVM and Retrotink 5X.

It's giving me issues though. You can see in the links/pics below; black areas in the screen have some sort of blow out at the top of the screen. It's white-ish... but with Starfox I can actually see some color in there like green. It's also causing some sync drops when there 's a bright flash on the screen... like dropping a bomb in Starfox (it drops sync every time I do that). On the Tink 5X I can solve the sync drops by turning SoG Tresh setting up to 200.

I don't understand what's going on with the black areas though. My Super Famicom (1 chip) and Saturn both do this... my PCE and Famicom don't seem to do this. I have alot of other consoles but I haven't tried them all.

https://flic.kr/p/2pKdSof
https://flic.kr/ps/449t6n

[Guspaz]

All of your PVM's inputs have corresponding outputs, why not use the PVM's outputs to loop the signal through to your RT5X instead of using an external splitter?

[Strider77]

Because I would have to needlessly route the cabling twice. As is I have two "zones" that my reciever is powering for audio. I have two sitting areas and all the cabling is already set up and hidden (actually the PVM set up is completelyhidden until I slide a curtain that reveals the second sitting area. I dread the thought of having to do all that cabling and hide it a second time. That's like 10 feet of cabling again and I don't know if there is enough space to cram a second set between the trim and carpeting (again).

Plus wouldn't the PVM have to be turned on to pass it through? I don't want to have to do that everytime I sit down to use my main set up. The PVM is really just for tate stuff.

Before I got my OLED and Retrotink 5X Pro everything was good. But after upgrading display/scalers some things changed in my chain (alot actually). So I've been dragging feet on the PVM. If Retrotink 4Ks weren't so hard to get ahold off I probably would hold off even longer with it's "rotate" feature. But I don't think I'll be able to get one anytime soon. With work I can't camp at a keyboard and my phone is to slow (at least it was with the PS5). So the PVM needs to be sorted out after dragging feet for 2 years.

[Guspaz]

It's been years since I've used it for that, but I'm 90% certain that you don't need to power on the PVM for the video outputs to work, though that doesn't help with your other concerns.

[kitty666cats]

Did you buy the regular one or the sync separator model? If the latter, perhaps try different configurations with the internal jumpers.

Otherwise, contact John! He’s as quick to respond as humanly possible (he’s had some health issues lately, but even with them he’s putting in lots of good work when he’s got time in the workshop!)

[orange808]

You might drop a sync stripper after the SCART switch and get a Kramer or Extron distribution amp. I haven't shopped them in a while, but they used to be very inexpensive--and they do what they're supposed to do.

[Gunstar]

Yeah I see it too (SNES '2Chip' C-Sync > OSSC > PC CRT)

I don't know if the artifact also appears on a consumer CRT set though, I couldn't see it in my tests. I have the version with Sync stripper function but it's set to stock configuration (stripping is off). Didn't have any de-syncs when firing off bombs.

[VEGETA]

isn't these based on rgb amplifier like ths7374? then why it suffers from sync problems?

[DrJohn]

I was given the link to this page by someone else claiming to have a similar problem, where the colour seems to be persistent outside the picture area. It's not something that I've come across or had reported to me before this.

I get asked about the design of the Active SCART Buffer and it's based upon video amplifiers. The ICs are rated for 150MHz, but because you're using them as buffers you have a gain of 2, which then reduces bandwidth by 1/2. 75MHz, but there's a tiny capacitor in there too to ensure the opamps work correctly, so it's slightly down on 75MHz but compare this with regular video where you only need 6MHz and you'll see that it's very much able to do this. Why did I design with such a large bandwidth? Because bandwidth is where you get to the -3dB point, which is typically considered an acceptable reduction in signal strength. But that area is where all the detail in your picture is, so to me it's not acceptable to reduce this and hence I designed with a much larger bandwidth to preserve picture detail and sharpness.

What about the loading of the source? Video is a 75ohm system. This is why with a buffer you need a gain of 2, a doubling. The buffer correctly loads the video lines with a 75ohm resistor. If it didn't, you'd get ringing in the picture and that would appear as a ghost around the detail in your image. It's also why you need to use good quality cables with video, but most are these days.

Output of the buffer? We have the input loaded correctly, so we know that the input levels are correct. It's all standard, nothing surprising. We have buffers and they're set to double the input voltage. Now the question is why do you need to double the input if it was already at the correct level? It's a good question, and it's because when you put video down a wire it is expecting a 75ohm load. That cable has to appear as a 75ohm resistor to the source as it's a "transmission line". To ensure there's no reflections down the cable too you then have to ensure that the source has the correct impedance as well. That means your source has to go through a 75ohm resistor before it goes onto the cable, which will appear as a 75ohm resistor. Mid point of this is the signal at the correct value, but your video is being dropped across two 75ohm resistors and hence why you need to double up the level of the signal because it's getting divided by those resistors! At the load, it will appear as a 75ohm resistor and looking back at the cable see a 75ohm resistor. This prevents any ringing of the signals along the cable.

Video is also an AC signal, not DC. Yet, you have a point in the video signal that is the black level and that is referenced to 0V. At the start of every line is the back porch, just after the sync pulse, that the system resets itself to the input signal's black level. If you put a video signal on an oscilloscope you'll see it varying about zero volts. It should vary slowly, but here's the problem. Some systems vary far to quickly. Often it's because the manufacturer has cut costs by reducing the output capacitors in their equipment. There's even techniques for using smaller capacitor that I've seen by having a small feedback look back into the output IC. Those are horrible.

In some systems where you have a fast varying DC level in the picture that can cause other equipment problems with the sync. I've seen this with some of the newer capture cards that can't keep up with a fast varying sync, and it comes down to the source really. It seems that some of these devices are just able to work with these difficult sources, but they're on the margin.

In this case, it's hard to see why it's giving more colour around the edges. It does look that the display / scalar isn't able to keep up with the varying DC level as quick as it should. With a CRT you tend to find that colour does stay around the edges and you will colourisation of the surrounding background. Especially when you're using high contrast computer graphics that have black borders which you'd not normally see on regular TV picture.

With the buffer the output is using industry standard capacitance and resistance. It will have an output which the DC level will vary slowly. The input picture image, if it's from a poor quality source, it will have a rapidly moving DC value, and the buffer will capture that. It will output the picture faithfully with that variation, plus the tiny amount already because it is a buffer with the right type of output. This can in some systems, not many, make it hard for a display to keep track of that input signal. Those that have reported this back to me will say that one display works fine, but the other doesn't. Or the capture card works good but the display doesn't or vice versa.

I'm trying to add two images off an oscilloscope that should also help explain this. Two different DVD sources, both on the same paused bit of video. Yet the DC level is varying rapidly on one but not on the other. This is a full frame of video (two fields) and you can see where there's a lot of picture information (the busy bit of the scope grab). The DC level should vary nice and slowly, but in this second one it's varying rapidly within a signal frame of picture. That's a bit quick.

In systems where this is a problem and the addition of the Active SCART Buffer is pushing it over the edge, I offer a couple of solutions. Firstly, to make the input DC coupled. As per engineering convention it's AC coupled, and it's rare to see a video input that's DC offset, but not impossible (a popular set top box in the UK had a massive offset years ago). It's something that can be done by anybody that's competent with a soldering iron as it simply involves shorting out the input capacitors. Second option is to add more capacitance to the output, which will slow the varying of the DC level further. It's a little bit more tricky to do but those a bit more advanced with soldering could do it. Finally, if you have problems and you don't want to do the modification yourself then you can return the Active SCART Buffer and ask for these modifications if you need them; it's a free service because I love my A/V too.

Does that help?

Now I hope those images attached correctly...

panny full frame 33-59 01 by docsymanski, on Flickr[/img]
pioneer full frame 39-59 01 by docsymanski, on Flickr[/img]

[DrJohn]

Just to add about the images. If you notice the V/division on the two traces are different. Yet they look the same - identical. Why? Yellow trace is 500mV/Div and Green 1V/Div which means the scaling of the traces are different on the oscilloscope. Green is the output from the buffer, because it's larger, and Yellow the input. This is shows how the output is doubled up as you need to do with a buffer.

What is really important, however, is how both top and bottom traces look identical. The buffer is faithfully reproducing the input to it.