Home Page Forums MAME MAME Sanyo MBC 550 or 555 emulator

This was my computer growing up & I would love to see it emulated. I am interested in attempting to program the mess version of this computer if no one else is working on it. Any suggestions of where to start would be welcome.

I have a memory map and tecnical manuals at home that will help.

Sanyo mbc 550 is a Ms-dos compatible computer from 1983

came with 64 up to 256 kb of ram
5.25 inch disk drive (MBC 555 had two drives)

parallel port

ran ms dos 2.11 and ibm programs that didn't use direct video memory access. Had a non standard video format of 640 x 200 by 8 colours.

Sounds like a varient of the normal 808(6/8) PC with custom video hardware. Probably it'd be just a few tweeks to the existing MESS PC driver.

that's correct uses a 8088 processor at 3.77 mhz

I expect it would not have to be done from scratch just a few tweeks. So I might be able to manage it.

You mean 4.77 MHz, right? That's the same speed as the original IBM PC and most early PC compatibles. 3.77 MHz would be a very strange speed to use.

Nope, 3.77 mhz is correct according to the original documentation and I believe I have read this elsewhere as well. Here is a sample: I thought it was strange at the time also.

http://www.qsl.net/k6fv/sanyo.html

3.58 mhz according to another source I looked up.

Be careful, even original documentation can have errors. (One of the reasons I'm skeptical of 3.77 is because it looks so much like a typo, where the typist hit "3" instead of "4".)

However, that link you provided actually makes things clearer. If you read it carefully, the guy says the clock speed of his MBC-555 is 3.58 MHz. That value is actually much more believable. Here's why.

Computers designed to produce composite video signals compatible with NTSC television are generally based on a 14.31818 MHz fundamental clock, because the frequencies used in composite video are all submultiples of this frequency. Thus all the frequencies needed in the system can be generated by dividing this fundamental clock signal. For example, the IBM PC's 4.77 MHz clock speed is exactly 1/3 of this frequency. (IBM used a CPU chip with a 5 MHz spec, but clocked it at a slightly lower rate for video compatibility.)

Well, 3.58 MHz is exactly 1/4 of 14.31818 MHz. (OK, you'll be off in the 3rd and 4th decimal places, but these clock rates are rounded to begin with.) So Sanyo was clearly dividing their fundamental clock by 4 to generate the processor clock.

3.77 MHz, on the other hand, has no simple relationship with 14.31818 MHz, which is the other reason I was skeptical of that figure.

[end pedantry]

Oops, sorry, our posts crossed. Anyway, I believe the 3.58 MHz value.

Properly emulating this system might be trickier than you think - not only is the video weird but everything else is weird in some way too (only exception being the add-in CGA card, but those can have an extra 256KB RAM so in a way they can be weird as well!)

Some of the quirks that may be a little tricky:

Reading non-existant memory returns the last byte read by the video circuitry due to bus capacitance. This is the only practical way to achieve vsync feedback without slaving the timer chip to count (with 100% accuracy to prevent drift) video position.

There are multiple dip switch settings to control how the composite monochrome signal is generated from the 3 color planes.

The speaker is wired to the transmit pin of the keyboard's serial UART. Emulating sanyo's buzz is easy, but emulating a 1 bit DAC running at 78.6KHz (approx 3.5 bit 6600Hz PCM if using 12 bits per sample) that can be achieved with a low pass filter right before the speaker is probably not so easy. Yes, you heard it right, Sanyo apparently unintentially designed this machine to have a pretty good sound capability by using synchronous mode on the 8251. Up to 1.33 samples can be buffered at a time in that scheme (compare to buffering 1/2 of a sample in a PC speaker driver using PWM, or 1/12 of a sample for PC speaker using an exact 1 bit DAC equivalent), and a RAM upgraded machine with 512KB has enough memory to play a recognizable song of a bit over 3 minutes length! Also emulating the keyboard receive when the 8251 is set in that mode would be a minor complication.

The memory bus is available only once every 4 cycles - I don't think this is the same as 3 wait states, but that would be a close if not exact approximation, since the system is so bus starved most of the time.

If anyone can make an emulator that properly handles these things and the rest of the hardware, I've got a $100 bill for you at least. (I'll accept some sloppiness in the floppy emulation, but it should at least handle the actual data tracks and not just sectors - that chip is a nightmare to program for, I can only imagine emulating)

I've documented what I was able to find out about the MBC at http://www.seasip.info/VintagePC/sanyo.html . The hardware's very dissimilar from IBM's: the floppy controller's a WD1793, for example.