• Please review our updated Terms and Rules here

MTI MXV22M to MXV22 (RX02 on Shugart Comaptible) Conversion

glitch

Veteran Member
Joined
Jan 31, 2010
Messages
5,439
Location
Central VA
Remember a number of years ago when Al found that big lot of MXV22 controllers, I bought them to split up, and they turned out to be MXV22Ms? Well, I finally got around to trying to change one over to a MXV22, and it works!

The MXV22M is a QBus controller made by MTI, which talks to 5.25" SA460 drives but presents as a RXV21 compatible interface; that is, RX02 compatible. Bit of a strange choice/format, but I guess it was a way to keep RX02 software compatibility without depending on 8" diskettes.

MTI also made the MXV22, which appears to basically be a 22-bit capable version of their MXV21. Both of these controllers will cable to a Shugart-compatible 8" floppy drive, such as the Shugart SA-800 or Tandon TM-848, and present a RXV21 compatible interface to the QBus. Unlike the MXV22M, the MXV22 and MXV21 are both 100% RX02 compatible: you can read and write diskettes that will interchange with real DEC RX02 subsystems.

The MXV22M purchase was something of an accident, none of us involved realized they were -M suffix variants. Having something that's RX02-alike but doesn't talk to actual 8" diskette drives is far less useful than actual RX01/RX02 media interchange compatibility. Fortunately, NeXT had a MXV22 and was willing to loan it to me (four years ago!) to see about converting the -M suffix modules.

Well, today I had time to set up and test everything. First, I tested NeXT's MXV22 with some RX02 diskettes I'd just prepared yesterday, following my repair of my (at least 12 years dead) RXV21 from my MINC-23. NeXT's controller works fine, it'll even boot from the KDF11-B ROM menu as it supports 22-bit addressing. Now that I knew it worked, I copied all of the jumper changes from the MXV22 to a random MXV22M from the lot I'd purchased. Then I moved the fuse PROMs from NeXT's board to the one I restrapped. There was one jumper I missed at first, and I forgot to close the 22-bit addressing jumper, but after I corrected those, it boots my XXDP 2.5 RX02 test diskette!

Now, the only trouble is, just converting these boards using vintage fuse PROMs is like...$70/ea from Unicorn Electronics, or roll the dice on maybe already programmed fuse PROMs from China. I went ahead and ordered vintage fuse PROMs just to make sure I can really actually copy them (I don't see why I shouldn't be able to, the Data I/O 29B with UniPak 2B supports it), but ultimately I'd like to design a little mezzanine board that uses modern EEPROM or Flash chips to replace the bulk of the fuse PROMs. This may require desoldering the fuse PROM sockets and permanently mounting the mezzanine due to height restrictions, but I don't see that as much of a loss.

As an added bonus, it looks like the MXV22 should be able to format RX02 media directly, without having to use the XXDP or 2.11BSD upconversion utilities!
 
That's great that you finally found time for this. I recall speaking to you about those boards when you first got them! What part numbers are the PROMs? Quantities needed?
 
Started dumping the fuse PROMs and I think my 18-pin Textool ZIF on the Data I/O 29B's UniPak 2B is going bad. I was getting intermittent changing checksums. So I've ordered a replacement ZIF socket.
 
I finally got some actually-blank 82S185 compatible fuse PROMs! I replaced the ZIF socket on the UniPak 2B and found a little green death inside two of the contact pairs, this was the cause of erratic bit flips in reads. I read everything in again, then loaded the files into the Data I/O and ran a verify...twice, in different orders! I'm pretty sure I have a good set of images now.

I blew a set of 7 PROMs with the MXV22 images, plugged them into my restrapped MXV22M, and they work! shirsch had bought some fuse PROMs too but it turns out his programmer can't do them, so he's sending them to me to program on the Data I/O. Once we've definitely confirmed that we have a good set of images, I'll make them publicly available.

I need to do more thorough testing on the MXV22 on loan from NeXT, and my converted MXV22M, w.r.t. formatting and writing. I have clobbered a couple diskettes trying to format and write to them, but it may be a media issue. I'm pretty sure it's not my drives as I'm using the pair of Shugart SA-801s I use all the time for testing with various (mostly S-100) systems. I'll clean the heads just to be sure, get a fresh pack of "last 8 inch diskettes manufactured" Athanas out, and make sure everything is working properly before having other folks blow fuse PROMs.

Haven't started on the "replace 7x fuse PROMs with some quantity of EEPROM/Flash" daughterboard yet.
 
Finally did a writeup on the process, complete with jumper diagrams:


The fuse PROM files are linked in the writeup. If you have blank fuse PROMs but no way to program them, contact me and I can do it on the Data I/O.

I have a number of MXV22Ms that I will go through as time permits and make available. If you want an untested, unknown board to convert yourself (cheaper if I don't have to!), contact me.
 
I went to get some more fuse PROMs from Twisty Wrist Arcade, which is where the 82S185A compatibles came from, but their site was down! Then I remembered I never checked with any of our usual suppliers for out-of-production ICs -- we have many such contacts since that's pretty essential to the actual pays-the-bills part of Glitch Works, LLC.

Turns out one of them has thousands of fuse PROMs at around $3/ea. I don't think I can build a EEPROM/Flash replacement for that cheap, so I ordered enough to do all of the MXV22Ms I have.
 
Nice work. I might have to keep an eye out for one of those cards and convert it
 
Fuse PROMs arrived and spot testing shows they're blank! I won't have time to blow PROMs this week due to work schedule, but now I have them in hand.
 
Really nice job! I'm not sure if I have a MXV11-21 or 22, will have to check. Any thoughts on enabling 22 bit mode if it's a 21?

Actually does it matter as a real RXV21 is 18 bit mode anyway.
 
More testing is required -- looking like 22-bit mode may be a DMA extension that requires patched/nonstandard drivers. The manual is super unclear.
 
I think I remember that. Probably similar to the trick used for SC03/B1 controllers that required patching at RSX compile time to do 22 bit DMA transfers.
 
Back
Top