Derek Andrews

You only have to look at the datasheet for the Signetics 2636 and the included schematic for a games console to see that they created an entire chipset, processor, video chip, sync generator and video encoder that forms the heart of these systems. It is of course possible that they worked with feedback from a customer, or customers, to come up with the central design, but most of the work so obviously comes down to the chip manufacturer. The main thing missing from that schematic is the external audio circuitry that controls volume and white noise, and specifically the assignment of the universally adopted choice of data bits that drive them. Discovering how that was set in stone would be enlightening.

I think 'processor' is not a good term to use; I've even come across people calling it a co-processor. In reality it is just a bunch of registers, counters and logic circuits and should really be considered as the same type of chip as say the 6521 Peripheral Interface Adapter, except that it outputs video rather than handling parallel I/O. Its full title is Programmable Video Interface and is totally controlled by setting values into its registers. It just chugs away counting horizontal and vertical lines and using data from the registers at appropriate times to output RGB values.

It is a PAL encoder, probably a prototype number for the TEA1002. And yes, it takes the crystal oscillator input and divides it by 2.5 for the 2621 Sync Generator, which in turn divides by 4 for the processor clock.

The 2650 is capable of addressing up to 32 kilobytes of memory, in four pages of 8k. However these consoles do not take advantage of page addressing so the processor can only address the first 8k. Included within this 8k memory map are the PVI and peripheral chips for reading the keypads and controlling audio circuits etc, so not all 8k can be used for ROM or RAM. I am currently writing a Wikibook, Signetics 2650 & 2636 programming, which might provide you with more technical information if you need it. There has been a lot of misinformation published and propagated on the web over the years.

The purpose of the adapter was to enable sales of Munch&Crunch and Leapfrog cartridges to Acetronic owners.

I would be interested in trying it on YPbPr. Are you in North America or elsewhere?

While its not exactly the same console, I did just get my Voltmace Database running with an AV mod into a scaler and thence to HDMI to tv. https://sig2650.wordpress.com/2021/05/26/a-v-mod-for-the-voltmace-database/

How important is a two-player option for a homebrew game for a retro console?

Thanks, that is interesting but I honestly don't know if they ever actually sold any of those. But yes, I did all the design work for DAZRAM. Given that it was nearly forty years ago, I don't think I can add much beyond what is in the advert. One end plugged into the ZX81, presumably an expansion port, and it had a ROM with Z80 code that I wrote to allow the user to input 2650 machine code. The code was stored in a RAM and connected to the Database cartridge. Given that the user was going to have to write their program in assembler and convert it to machine code themselves, and the complexity of programming the console to do very much at all, it would have taken someone with a lot of perseverance to get very far. This was probably one of the crazier things I have ever been involved in, but I was getting paid to do it.

The joystick circuitry is odd indeed, and I can't say that I fully understand it. The 2637 pins 8 to 11 that read the joystick switches are analog inputs that are measured by an analog to digital converter on the chip. The asymmetric values of the resistors and capacitors in the circuit (10k/470k ohm and 0.1/0.047 (uF?) ) must be significant, perhaps forming some sort of timing circuit, but that is just a guess. I have no idea what the jack plugs J1 and J2 are connected to or what role they play.

Yes. The 74LS145 is used to output a logic 0 to one of the six sets of four keys each. The 74LS258 is used to decide which of the two handsets to read, and outputs the state of the four keys selected by the '145 to the low bits of the databus. http://www.ti.com/lit/ds/symlink/sn74ls145.pdf http://www.ti.com/lit/ds/symlink/sn74ls258b.pdf https://console5.com/techwiki/images/f/f0/Emerson-Arcadia-Schematic.png

There are large solid areas for the joystick switches.

I wonder why they leave the gaps in the contacts Why not solid round pads?

A bit late replying on this, but yes, and for the record: The 'speed' of these memory chips is usually quoted as an access time. A -150 signifies a memory device whose contents can be accessed in 150ns. A -200 takes 200ns for the data output to be valid. Conversely, microprocessors are specified to work at a particular clock rate, for example the Z80 ran with a clock rate of 2.5 MHz, the Z80A runs at 4MHz, the Z80B at 6MHz.

I finally broke down and bought myself not one, but two Voltmace Databases. It will be a while before I have the time to get them powered up on 110v and displaying their video, but in the meantime I plan to open them up and document their construction. I have a list of things I want to record (below), but wonder what others might like to see recorded for posterity. Photos of the original box artwork. Photos of the polystyrene packaging. Photos of the mouldings with dimensions. Photos of the circuit board, top, bottom and one superimposed on the other, from which a circuit diagram could be constructed. Parts list. Manufacturer and date code of all the IC's which may give a clue as to the approximate vintage of these two consoles. Ditto the cartridge wiring and construction. (My regional library has a 3D scanner and printers, so I will see if it is possible to scan the two parts of the cartridge mouldings and print new ones).