Unable to get RGB image from MSX2

Yeah you are correct, you can wire all the grounds to a single point on the connector. With return I actually meant the video and audio signals coming from the other device (as both SCART as well as RGB21 are 2 way cables for the composite video and left and right audio).

Not sure if you missed my message on the previous page, but it still looks like you've got pin 16 on the SCART side wired to "use CVBS" rather than "use RGB." (Note that there's a bit of confusion in that message; I mentioned ja Wikipedia when I was talking about the SCART function of pin 16, but the idea there is correct: you should have a 1-3V signal on pin 16 of the SCART input, and you can probably get that from pin 16 of the JP-21 output.)

Can you check, with the computer and the OSSC connected together and both turned on, what the voltages are on pins 8 and 16 of the SCART connector? 8 should be above 5 V (preferably above 9.5), and 16 should be above 1 V. Those are the only two signals I can find that might select the SCART display not to display the incoming RGB signal. If those are both confirmed high while you're operating, there's probably some other issue that may be difficult to debug without a 'scope.

Yeah, so that looks fine. It seems unlikely your problem is due to incorrect wiring of the SCART control signals, unless your monitor is particularly sensitive to pin 8 being just slightly under the 5 V threshold for on/16:9 aspect ratio. (The high threshold of "off" is 2 V, so it's well above that.)

That's not strange at all. You may be familiar with the waveform for a CVBS signal (often called "composite video," or even just "composite," though the latter can also refer to just "composite sync" with no video signal overlaid on it). RGB is pretty much the same thing, except that there are three channels with just the video levels (R, G and B) and one more with just the sync signal. In some cases, the sync may be alone, in others it may be a full CVBS signal (monitors usually find it easy to extract the sync from this). A third option is that the sync is carried on the green channel, making G be a complete CVBS signal though the video information is just the green colour.

A multimeter may see an AC voltage on any of these (sync always, R, G or B if there's changing video information on them, such as alternating colour and black on each scan line), but it may not. The issue there is that many multimeters are fine with very low frequency AC signals (50-60 Hz, as mains power uses) but don't have the frequency response to detect the much, much higher frequencies of AC video signals: a sync signal is roughly a 15.7 kHz pulse wave with a very narrow pulse width.

One thing you can try to test out your multimeter is to clear the screen to black and try to measure the AC voltage on your known-working composite video output. Whatever voltage you see there is about what you should see on the composite sync line of the SCART connection.

Another is to use a frequency counter on the sync and R, G and B lines (the latter three with some sort of pattern on the screen: just filling it with a letter should do) to see if that gives you a number. I'm not sure how well that would work, it depends on the frequency counter. Many logic probes also will show if a signal is present and changing.

In the case where your computer is supplying sync on green, you can wire the green output directly into the CVBS input of a monitor and you should see just the green component of the RGB signal.

If you are lucky enough to have a PVM with an "H/V delay" switch (Sony PVMs usually have this) you can press that which will shift the entire image down and to the right and add some luminance during the horizontal and vertical blank intervals, giving you a "cross" pattern on the screen showing the sync and borders around it in a lighter colour. If you have this, try it out on your working composite output to see what it looks like, and then try feeding your SCART sync signal into that same input with H/V delay set and see if you see the same pattern. That will tell you if you're getting sync on the SCART composite line. If not, try that with the green signal in case it's using sync on green (though if it is, the test above with green into CVBS should work).

Failing all of this, my best suggestion is to buy a cheap oscilloscope so you can get some idea of the waveforms on the lines. Even a little $30 one from AliExpress should do fine for video frequencies, since these are very low compared to what 'scopes are normally used for. If you get one, start with analysing the composite output (triggering on the sync pulses as best you can, though this will probably not produce a perfect display) since you know that works and it will help you learn to use the 'scope settings and give you an idea of what to look for on the other lines.

Sorry, browser problem... Now I can see it.

I don't know if you have updated your wiring, but instead of:

To me it looks like pin 20 (SCART plug) goes to pin 11 (Socket) or 12 (don't know if you are using an additional SCART extension with it)