Looks like the site MK already showed me... at first I dismissed it because it looked like to me like it was only about enabling a very specific feature, 240 pixel vertical resolution, and that was it. It looked like neither the S/W / project nor the board, were capable of modifying every aspect of the video signal. Plus, I am not a F/W engineer, nor a video protocol guru, so even if the board were capable of arbitrary timings, I would not be competent to make use of it...
But I looked at it again and it looks like I was wrong... there seems to be a user friendly interface to let you repurpose the video chip, and it has presets for NTSC and PAL signals.... i.e stuff that's even farther from CGA than my signal is.... so maybe there is hope for me as well. OK OK.... I will keep that in mind, will look at it in more detail at some point...
Anyway, some good news : I did what I said I would do... I added a XOR gate on the breadboard to create a composite SYNC signal, as per what I saw on YT.
Well, my old book on logic gates says there are somehow only two TTL chips that do XOR ... 7486 and 74136... with my usual luck, despite having a broad collection of crusty TTL chips.. I didn't have any of these two !
I didn't feel like using multiple chips to cobble together my own XOR gate either, too much mess on the bread board and not enough Dupont wires.
So instead I went for the next candidate : XNOR gates. Book says there is only ONE chip that does it... 74266. Well I DO have a few of those !
Problem.... it's got an open collector output !
Would that output be good enough to handle my sync signals without "deteriorating" them to the point that it might upset the video chip... and what resistor to use to get optimal signal shape ?! More headaches in perspective, grrr...
I let destiny do the math for me : a 3.9K resistor happened to be on the breadboard from a past experience I guess. Good enough I thought, so I used that and scoped the output... perfectly square signals I was greeted with, what a relief !
So I connected that to the video board, and hey presto, much better now !!!
As I thought, it now works well enough that it's usable, but not quite perfect because it's not CGA, only close to it.
So, what do we get in exchange for our XNOR gate ?
1) The board is now reliable. It never freezes, not even once. So the F/W must have been upset by my separate sync signals I guess. Poor F/W it is.
2) The AUTO push-button now works : as soon as I press it, the red LED next to the MCU, blinks for 3 or 5 seconds while it's trying to figure the sync signals out.
3) Once that's done, I get a meaningful and stable picture, woohoo...
The horizontal side of things is working well, as I hoped, because at 15.4kHz it's close enough to CGA for the chip to make the effort.
So I was able to adjust the horizontal position of the picture to center it on the screen, and adjust the size.
However as I anticipated, the vertical sync is too far from CGA, so it kinda works but not perfectly.
The vertical position control behaves weirdly. The best / most useful I could get for a picture, is to get the bottom area of the screen, where the menu bar is, to roll over and display instead at the top of the screen. If I try to put it at the bottom, and I can't achieve it anyway, it destroys the picture.
So now I have a stable and usable picture. Menu bar is at the top instead of the bottom but who cares... let's just say I customized the F/W in the programmer to suit my particular taste !
So now I can watch the entire boot process and see exactly what's going on. It's just... so cool.
This is what we see :
1) At power up, displays logo and "Insert system disk and press F1", and the leftmost "button" in the menu bar is active and shows "BOOT", which ads up.
2) That lasts only a split second, so I had to remove the system disk from the floppy drive to force the machine to stay in this state so I would have time to take a pic of it and a video clip. That's why it says "Failed" with some error codes... it can't find the floppy in the drive.
3) There is line of text in reverse video that blinks, this is what I thought might be a blinking cursor a weeks back when I was first watching at the signal in the time domain on my scope. It's more than just a cursor then.
4) I was right : the programmer thinks it is a full fledged model 5000 rather than the humble deaf and mute model 3000 that it is ! Really interesting, that.
5) OK so then I insert the system disk and reboot it so it can actually boot from the drive. It reads the disk for about a minute, during which the blinking text reads " REVISION 5.40AP ". Also, the menu bar contents change : the "BOOT" option on the left is gone. Now the only button available is the rightmost one, which reads "ABORT". Ads up... first you press F1 to boot, then while booting you can abort the boot process.
Then in the video, at T = 1min14s or so, the drive stops working (I turned off the sound in the video, sorry you can't hear the drive working to make the video less boring)) , boot process is complete. At that point the blinking text now VERY BRIEFLY says " ERASING MEMORY ".
6) Then the logo and text disappear, screen is cleared and we get just the menu bar at the bottom, well the top here sorry. Again only available menu option, still the rightmost, that says " END ". Underneath the menu bar, it reads " REMOTE CONTROL ".
Now in video :
So how cool is that ? It is SUPER cool !
Even cooler : that French forum is cool too, lots more people active in there than I would have imagined !
I have had already two people reply, and those two actually have a 5000 model and both live not too far from me (100 / 150kms away), and we are thinking of setting up a gathering at my place with their 5000 so I can see this beast in the flesh and do some reverse engineering in my lab
Maybe this spring. Some time anyway.