A new powerful MSX videochip...

Interesting: i have seen a lot of people that complained about vdp limits, but no one will buy a new video card.....

Screen2 limits just gives MSX its personality.

The TMS wasn't really an MSX chip. So, to say that there is MSX personality would be a bit too much honor.. The interesting question is whether the MSX2 VDP (designed for the MSX2) defined the MSX2 personality. Same question can be asked regarding the MSX2+ VDP. Simply "what" defines MSX/2/+... "what"?

I'm all for developing new hardware, even for old platforms like MSX. And I'd be the last to try and stop you from developing something that only you and 2 other people might use. But:

History has shown that to be useful, you have to have 'content' for it, and small incremental improvements don't do it, when you're trying to put a new 'product' in the market. So if you want to create a new VDP for MSX, I'd try and think of ways that could help it become popular:

a) Make it compatible with what exists. For example, a replacement for a MSX2+ videochip, 100% or 'close enough' compatible, only way faster, more video memory, sprites per line, higher resolution / color depth combinations etc. If compatible, you have an existing software base, and people that know how to program it. And thus, USE it and create new things for it easily.

b) Or use something known from other systems. If you create a new VDP add-on for MSX, but the video hardware used is known from other (game?) platforms, that may enable easy porting of games that exist for *those* platforms, to MSX.

c) Integrate it in a bigger project. Think 1chipMSX, but even better (and lower priced). That way, you'd have -potentially- a bigger audience, more users, more interest, more developers, = more applications.

d) Or make something that is easy to grasp mentally, provide excellent documentation, open source design files, cheap and ready-to-buy development boards etc. Then it just *might* be used for educational purposes. Read: more users, more people interested in it, ... (well you get the idea)

But really, whatever you do, PLEASE aim your efforts at something that has the potential to grow a large user base. Otherwise you're doing it just for yourself. Which is fine by me, if you just do it 'because you can', for the fun of it, or to learn things along the way. But don't kid yourself into thinking you're supporting MSX or advancing MSX as a platform, when in reality only you and 2 others are using it. Personally, I'd pick from option c) or d), depending on the reasons for starting a project.

Maintaining the compatibility also maintain architectural limitations. Another example:
MSX2 sprite system use the special Y value to hide sprites (the same as TMS) but this in combination with vert scrollregs forces the programmer to watch for this value when scrolling, and do not update this. Result:
1) non smooth sprite movement
2) another work to do for the programmer and the CPU, especially considering that usually the SAT is written on VBLANK, a very limited timespan, often also used for other activities.

That's the result of maintaining compatibility.... If the V9938 does not had to be compatible with TMS but an independent chip, we surely got a better chip. Now, maintaining compatibility with all video chips means to inherit all the dirty things (and there are a lot of dirty things on msx video chips) on the new hw.

Best to forget all, and start from a clean prj imho.

to pinpong:
you can extend the v9938 by adding a bit in an additional register
to switch the sprite switch off system with a more convienient one,
0: standard mode
1: extended mode , the one you intend to implement in the new hw
after reset default is standard mode.