Home Page Forums MAME MAME Requirements?

You could just connect it up to an Arduino to drive the address lines and read the data lines.

This happens if you hook up bsdimp's ROMs WITHOUT any patches ... unfortunately it won't boot either.



Sorry about the aspect ratio ..

Yea. that's my plan of last resort... I have a nodemcu ESP8622 module that has 13 GPIO pins... I need 16 for this . Time to do a deep dive into my collection of random boards to see what else might work... The BPM Micro stuff referenced earlier are about 50x-100x more than I want to spend on this project .

BTW, I think I was just able to recover the driver for the Univation Disk controller for MS-DOS 2.11 from a univation controller and semi-crashed disk I bought on ebay two years ago... Now to reassemble the drive enclosure to see if I can get it working on real hardware. Looks like there's only 3 registers and maybe an interrupt handler to this driver.... I'll let people know what's up after some testing and fiddling to figure where to put one of the two univation jumpers... If I get it working, is there any interest here in this stuff?

There's absolutely interest in whatever peripherals you can dig up materials for. Good photos of the controller card where you can read the chip labels/types are important too.

I'll give the card modeling lessons then... I'll scan the docs I have (again?) and post what I've puzzled out along with pictures. It looks like there's an I/O port interface that's small. A magical process happens (more later when I figure that out) and the data goes out a 34 pin cable that's converted to a 50pin SCSI cable inside the disk enclosure. The enclosure has an adaptec ACB-4000 that rejiggers the SCSI commands into Winchester. The disk's on-disk structure differs from a standard rainbow in ways I've not puzzled out. The card doubles as a memory card, and there's a weird 64k that's on the 'adapter' card, unsure if it is cache or extra memory for the system. The card plugs into the memory adapter, but has two 'vampire' clips that connect to the 74LS138 E51 over the BDL Select 1, 2, 3 and 4. The second connector connects somewhere else. I'm unsure if this is for the 100A to use all the memory, of if it's an interrupt line, more when I puzzle that out (my current favorite guess is E80 74LS373 which is fans the interrupt sources in)... I'm also trying to recover files from the badly non-functional drive that came with the controller....

And if I'm really really lucky, a machine with a second one (but no disk) arrives tomorrow hopefully with the vampire connectors intact. I have docs, but it only talks about one of the vampire connections.

I was both lucky and unlucky.

The new Rainbow did not have a Univation HBC. Alas. I was wrong in my guess.

However, there was a C.H.S. Dual Winchester Controller that supported 2 winnies. So that might be interesting to emulate. I'll see if I can add emulation for this. I know Venix supports it, and IIRC, the DEC driver in MS-DOS 3.1 does too (but I've not confirmed this). This card is based on the WD1010 as well, and I believe just uses the 'drive' field (bits 3 and 4) in SDH register. I'll try to confirm that. I've not traced it out, but since it was supported in Venix I think it's only a small delta from the standard DEC card.

There was also 8087 card with memory and ???? things with a ROM socket???? Weird. I'll have to look at this and maybe add support (if there's 8087 support in base mame).
And some Dallas semi RTC chip (my pics were too blurry to know which one)
Oh, and if that weren't enough, a graphics card!
Now, to see if all this stuff works or not...

Question: The Rainbow's Hard Disk controller is based on the WD1010AC. However, src/drivers/rainbow.cpp uses wd2010_device. There's now a wd1010_device, which supports multiple drives, while the wd2010 one does not. Is there some reason why the WD2010 was used? While I've thought over the years of swapping a WD2010 in for the WD1010 because it supports the same drives, but supports up to 2048 cylinders instead of 1024 like the WD1010, it seems like the Rainbow should be using the WD1010 device. So I thought I'd ask: would it make sense to do the changes needed to enable using wd1010, or is there some known issue that would block that?

The wd1010_device didn't exist at the time the Rainbow was brought up. It's not uncommon to do a workaround in MAME and then have to go back and do it correctly.

I'll have to give it a shot when I find some time...