About the reason why you need to use TI memory chips: the GA has one (well, several) bug that is triggered when you want to write a value to a I/O port managed by t he GA (that is, port $FE, write). To put it in a few words: when you write a value to port $FE, the GA also thinks you want to read from memory from the same address it has been used as I/O port address. The value that gets actually written to the GA is, then, the bitwise AND operation between the value coming from the Z80 and the value coming from memory. In the ROM, all accesses to port $FE use the OUT ($FE),A instruction, so the 16-bit address it is actually used is A*256+254. As bits 7 and 6 of port $FE are not used in a write operation, the ROM always set them to 0, so the 16-bit address is always in the range 0000-3FFF. Now that's the interesting part: there is 64KB of actual RAM in the Inves. the mechanism I've just described works the same no matter if it is RAM that can be addressed by the Z80, or RAM that lies behind the ROM (in fact, this is the only moment in which that RAM, and not ROM, is read). Such RAM is never written to by the ROM, even by the initialization routine, so this part of the memory keep its original content, the one it had when it was switched on. Due to how internals in a DRAM chip work, discharged cells in an even address are read as 1, while discharged cells in odd addresses are read as 0 (you can notice this if you press and hold reset while you switch on the machine and see which pattern is built on the screen). So, as long as all cells in RAM at addresses i*256+254 (where i ranges from 0 to 255) contains $FF, writes to port $FE will work as usual. Writes performed using OUT ($FE),A which A

Oh, Flare Technology (not Technologies) - figures as they were Ex-Sinclair people that were heavy in custom chip design (Flare One -> Flare II -> Atari Jaguar) en.wikipedia.org/wiki/Flare_Technology

Right. I figured that afterwards when I realized some of the functionality of the original ULA was implemented externally (interfacing with EAR/MIC and the joystick port). Very interesting. I wonder how much access to the original ULA (if any) they had.

@@NoelsRetroLab Well, AFAIK, Investronica wroked side by side with Sinclair in the 128K design, so I guess they had plenty of access. Unfortunately, the Investronica person involved in all the technical stuff passed away some time ago. I will try to find out which british company was involved in the design of the GA circuit.

Thanks! Hopefully part 2 will be the resolution :-)

The goal was to play Sabre Wulf on that machine (since joystick is all but mandatory) but didn't get as far. Next time!

Thanks! I was told the same thing in some other forums so... I guess I'm going back and researching what's going on! :-) Stay tuned.

Glad you liked it!

Glad you enjoyed it! Did you see the series I did on the QL specifically?

Thank you! Glad you're enjoying it. And no, don't worry, I'm not forgetting about the SVI 328! As a matter of fact, here's a minor spoiler: I started shooting part 3, put in the new VDP, saw the clock signal working again, but the video was all colorless and weird... It turns out I had bought an NSTC VDP!! :-( So waiting for the PAL one to arrive to wrap up that series.

Thanks! Glad you enjoyed it. Maybe one day I'll get my hands on a Timex and I'll make a video about it.

You need to make sure they all have the same reference ground, but yeah you could. Are you suggesting that to avoid noise artifacts in the image? To be honest, the bench power supply is somewhat noisy as well the times I've looked at it with the oscilloscope.

Yes, it wasn't very popular. I'm guessing it was sold mostly to new users, so if your friends started with the original Spectrum, they weren't the target audience.

I would check the resistors used in the voltage dividers as well. They may be passive but the values may have changed over time.

This particular computer only had 64KB (only 48KB usable) so it wasn't trying to do the same thing as the Amstrad CPC 472.

@@NoelsRetroLab It's the same just to avoid taxes. cpc 472 never had usable more than 64 (they were disconnected!!)

🤣🤣

pronounced "mah cloud". dont know where you got the "mer" from. Im Scottish BTW, and had a MacCleod great-grandparent

Oh, they made sure that they looked a LOT like a Sinclair machine. I bet most kids getting one didn't realize it wasn't one.

Good suggestion. I'll make a short video on the basics of scope usage soon.

It helps that I'm from Spain :-)

@@NoelsRetroLab LOL, that explains it! 🤣🤣🤣

@@NoelsRetroLab Except for "Miguel Ángel Rodríguez" you have a very good English pronunciation. :P

That's really cool. There are some very interesting clones in East Europe. How could they get actual Ferranti ULAs though? There has to be an interesting story there.

@@NoelsRetroLab I think Ferranti was initially unable to keep up with demand from Sinclair. As a result, many chips were ordered and produced. But then, Sinclair's poor financial situation combined with the release of 128K meant that nobody wanted these chips. Yugoslavia was a bit of a gray market country and they somehow managed to acquire genuine Sinclair parts. How they did it is a mystery to me. Once behind the iron curtain, Czechoslovakia was able to buy a batch of ULAs from Yugoslavia. Deltas were not even officially classified as computers but rather as teaching aids. They were literally sold at stationery stores instead of consumer electronics stores so that high commie officials don't get the wrong idea. At one point, Slovakian Didaktik company even acted as a semi-official Sinclair "repair shop" which enabled them to order another batch of ULAs directly from Ferranti and to bypass import laws. Ferranti probably knew what's going on but it was all good on paper. Today these stories sound strange :)

Oh I'm afraid it isn't anything that interesting. The IC just takes care of converting the RGB signal into a PAL composite video signal. That's the IC that the QL and the Inves have in common. The ULA, which controls the video modes, is completely different. So no hidden video modes in the Inves I'm afraid :-) (although we could do something to use the full 64K of RAM!).

@@NoelsRetroLab thanks for clarification. The Timex 512x192 pixel mode was a kind of surprise to me. I think we could kill for such a resolution on Commodores :-) Imagine GEOS with that on C64 (yeah, I know C128 had a version with 640x200).

It's on my list of things to do. First I need to do some testing. Don't worry, if I come up with anything, I'll make a video about it.

Yes, it's a Kempston standard, so that's perfect for that time period.

But there's no chroma crystal that I can see! Just a master clock that then gets divided (I'll get into it next video). Also no trim capacitors that I can see. The other weird part is that on screens with other colors, everything looks great. Maybe it's just a slight green tint and it's only noticeable on a gray background.

@@NoelsRetroLab I think with "trim capacitors" the capacitors with fixed but specific values are meant. I'd check the clock frequency next. Also: I suspect that the first part of the name "mcleod_ideafix" is possibly based on the Scottish name. See Wikipedia for correct pronounciation (or the Highlander movies ;-)).

Ah, I thought he meant trim capacitors as in adjustable ones. I'm looking into the clock frequency right now. Unfortunately the division happens inside the ULA, so I suppose that could be bad (and no way that's easy to replace). I will look at the master clock and even try replacing the Xtal just in case. Oh and I'll check the pronounciation. It's been too long since I watched Highlander :-)

@@oldguy9051 Indeed! I chose that nickname because of the Highlander movie. Ideafix is/was the name of a dev group I involved into during the 90's (much like a demoscene group, but with much less glamour)

@@NoelsRetroLab Is it like the 128k Spectrum where there's a 17.7345 MHz crystal that is driven by a 74S04?

Oh wow, I didn't even think of that! That would be devious! :-) But I don't think that's the case because the diagnostics ROM (which I didn't show here yet) shows a greenish screen as well. I'll try that just case.

OK, it gets points for thinking outside of the box, but I just tried that and still the same :-)

@@NoelsRetroLab Uff, that's a shame; it could have saved you a lot of troubleshooting if that had been the case.

That'd be on the MC1377P, right? So the color components? I can give it a try, but I've never had a ceramic capacitor go bad. I'll let you know.

@@NoelsRetroLab I have seen several times changed the ceramic capacitor internal resistance up to short-circuit.