Feels cmos (which makes sense for 1990), single-metal (more surprising, cost I guess). Active is barely visible, so it *could* be mask at the active level. But tbh it's probably implant.
McDonalds LCD games often even lack a score display, which made them quickly become boring after having won. But a few were surprisingly nice. E.g. this "Spyro" is a 3D maze game with inverted see-through segment graphics display.
As a kid I watched the school TV series "Einführung in die Digitaltechnik" (Introduction into digital technology) and always mused if LCD games consist mainly of shift registers. The way objects shift stepwise in a row in many games look like that. When multiple groups of objects step alternately (triggered by phase shifted slow clocks) in later levels, the outputs of multiple identical shift register flipflop rows were supposed to be connected through logical AND. Later I read that most LCD games had 4-bit CPUs based on calculator hardware.
But these MT3215B and MT3238 "McDonald's LCD game" chips seem to be indeed made of pure gate logics (no RAM/ROM). Isn't it? (Or are there hidden layers underneath?)
The McDonalds games are fun, but they’re dedicated logic, not microcontrollers. Not something that fits well in MAME, you’d need to trace the logic to simulate them.
If McDonalds games are all in gate logic (tough to decipher) I did not expect them to be emulated soon. (IMO the many Pong-age single-chip TV games are more interesting to spend time with emulating them.)
It is only surprising that such implementations were still in use long after 1970th to early 80th. Of course special realtime things like simple music keyboards had such chips, but instead of using slow MCU it can be cheaper and more effective (i.e. high mixing frequencies make less aliasing noise and lower key latency) to use a gate logic LSI than implementing them in software. LCD games need very little CPU power, so it may have been a patent fee issue or simply some old chip designers still had the tools to build them this way.
When chips barely freak out in complex ways by shitshot attempts (i.e. stuttering the supply voltage or clock pin connection etc. to provoke software crash) it can be a hint that they are pure gate logic instead of CPU based. Likely also the HF emissions differ between both kinds, which can help to do a preselection whether they are worth to search for rom dump test pins or decapping them.
Just to add a bit to the TV detector van stuff in the UK, the method they use from what I recall is a lot more mundane and simple; they look for the TV's local oscillator radiating from the tuner/antenna. They can tell if you have TV and what channel it's tuned to based on the frequency of its local oscillator which will be off of the channel of interest by the IF frequency (i.e. +45.75MHz or so).
The other theory is the vans are just for show, and using sales records vs. who bought a license is enough to catch the dodgers.
And for something on topic, I will try to photograph all the VFDs and pin them out on these games next week; I got the VFDs connected up so that all their segments light now.
Artwork is up for Biathlon in the "Non Nintendo" folder in the link below. The background is printed on paper so it doesn't have the typical translucent backdrop with gray background. The game is an......interesting idea using the same ROM as G&W Egg and countless clones.
The games loaned out by alexb has now been dumped and added to MAME. With this, all known G&W games and variants has now been preserved Thanks Alex for providing the units!
![[Linked Image from i.ibb.co]](https://i.ibb.co/bgGJ6Rd/Tayny-Okeana.png)
![[Linked Image from i.ibb.co]](https://i.ibb.co/W5D5p7G/biathlon.png)
Thanks Alex for providing the units!