I am going to need to low level format it. Please correct me if I am wrong....

Yes you will have to Low Level format with the controller IF it is MFM or RLL.
Ok to better help: need the original JVC HD part number (and the new one), a picture of the drive with any labels be best. Also need a picture of the controller card.
You included two files, both only are 1/2 full of their 8k which seems a bit odd, how did you retrieve them ? -->> need a picture of the controller. Thanks

Horun wrote on 2023-09-16, 01:25:

Yes you will have to Low Level format with the controller IF it is MFM or RLL. Ok to better help: need the original JVC HD part […]
Show full quote

I am going to need to low level format it. Please correct me if I am wrong....

Yes you will have to Low Level format with the controller IF it is MFM or RLL.
Ok to better help: need the original JVC HD part number (and the new one), a picture of the drive with any labels be best. Also need a picture of the controller card.
You included two files, both only are 1/2 full of their 8k which seems a bit odd, how did you retrieve them ? -->> need a picture of the controller. Thanks

The controller board:

The eeproms.

The eeproms are both 27C64s and I pulled them and dumped them with my eprom programmer. I think one is the BIOS and the other is code for an on-board micro.

Unfortunately the original drive was relabeled and I did not see a label underneath it.

This is the new drive, removed from a Toshiba T1200.

Here is a link to all of the pictures I have taken:
https://imgur.com/gallery/o1hAdmD

I just stumbled on the fact that this 20MB drive is also the same model used in the Zenith Z-180 laptop. I could maybe try to find tools for that, but I am not sure they would work since the controller is likely different as this was a third party upgrade.

If it is helpful, HD-DIAG reports 2 heads and 615 cylinders. But I think that it said this for the 10MB drive as well. That is what makes me think the CHS is hardcoded in the BIOS. The SEEK test passes with HD-DIAG, but the head select test just reports "Address Mark Not Found" for every other test. Not unexpected if I had done a low level format.

I honestly have no idea what I am doing when it comes to MFM or RLL drives, so what I think I know has come from research online. It appears that LLF occurs by accessing some routine built into the BIOS of the controller card. Many mainstream controller cards have been documented, but unless this is a clone of some controller board, I am in the dark where that routine would be located within it. In the ROM, I do see what looks like possibly a table of entry points at the beginning. I don't see any strings I would expect to be generated during a low level format, though. And yes, I also found it unusual that only half the ROMs seem to be filled. I was expecting a lot more.

EDIT: I found this page which appears to confirm that the 20MB JVC drive geometry is 614/2/34. I attempted to do a low level format by writing some assembly using debug using an interleave of 3 and it appeared to be working. Going back to hd-diag, the head select errors now read "ID NOT FOUND" instead of "ADDRESS MARK NOT FOUND" at least until cylinder 160 and then it reverted to the address mark error. Reading around, apparently an Address Mark error is better than an ID Not Found error. Hmm...

https://knm.org.uk/blog/2017/04/the-jvc-26-pi … terface-part-1/

I am suspecting that the power supply in my Zenith portable might be at least part of the issue. When the disk is spinning up, my LCD is darker than normal and flickering. Perhaps I am not able to give the drive enough current?

The BIOS ROM is located at C000:8000 as it should be. There is a jump table at the beginning, but it appears to require loading some registers beforehand. They are not "user friendly" routines. But this will hopefully assist me in disassembling and understanding the BIOS better.

This 20MB drive is 614/2/34 which does translate to 20MB. I suspect the 10MB drive has only one head. Reading through some confusing online information, it seems that the track 0 is special and contains the disk geometry. I used debug to unassemble the BIOS and looked through it and I do not see anything that looks to me like a low level format routine. Lots of INT 13 calls which I will need to look up, but I didn't see any mode 07 in there unless I missed it.

I did stumble on several pieces of assembly code to run in debug which should low level format and running them does seem to step through the disk like I would expect. This does not seem to allow fdisk to see the drive, regardless. The drive could be defective, of course. But I have no real reason to suspect that it is. I am definitely missing something that is likely common knowledge for RLL or MFM drives for which I have no prior experience.

HELP, PLEASE! I am not a complete noob and don't need full hand-holding. I just don't have experience this far back and need some direction pointing. Should I expect an INT 13h function of 08h to give me useful drive parameters?

I never did get this sorted out. Another person had created an ISA adapter and an XT-CF board for this, which met my needs for storage. I also was able to fit a gotek drive. I was recently revisiting the BIOS and internal SRAM storage of this computer with some thoughts on more upgrades.

I know have a few of these computers and they are hands down some of my favorites to play around with, next to the Sharp PC-4502.

Just for the sake of archiving, I have dumped the ROMs for this rare upgrade and would absolutely love to revisit it and get a compatible drive working again to pair with it. Unfortunately throwing money at more 26 pin JVC hard drives is expensive and with me not knowing enough about MFM/RLL hard drives, I have conceded for now that my chance of success is low.

I'd have tried an older version of ontrack disk manager, myself-- or a copy of spinrite.

Either of those can do a low-level format without the controller bios, since in this case, the onboard controller's LLF routine is hard to access/obtuse.

Ontrack's DDO might actually be useful here as well, if the upgrade's disk bios cannot handle multiple heads. Head 0, track 0, is all you need to load the DDO from the disk, and then it can handle the disk IO, at the expense of some of the system's low memory. (Most people use a DDO to overcome the 504mb disk restriction imposed by AT Bios, but this thing apparently has different problems with geometry. It being MFM/RLL instead of IDE is not that important, since we are not activating LBA or anything-- Manually disabling those features after installing the DDO might be needed.)

I believe that I had tried both of those programs. I was also likely battling the fact that the 26 pin JVC drives are typically all bad by now and both of the drives I tried were actually just bad. Dealing with the question of whether it is a bad drive or incompatible controller and no original software utilities for the upgrade made this project quite difficult. I could toss another 26 pin drive at it, but it gets expensive to keep trying these old and rare drives.

Retroplayer wrote on 2025-09-27, 15:38:

I believe that I had tried both of those programs. I was also likely battling the fact that the 26 pin JVC drives are typically all bad by now and both of the drives I tried were actually just bad. Dealing with the question of whether it is a bad drive or incompatible controller and no original software utilities for the upgrade made this project quite difficult. I could toss another 26 pin drive at it, but it gets expensive to keep trying these old and rare drives.

I saw JVC drive, 26 pin connector, ... and thought, hey maybe this is relevant ?
Re: Unusual MFM drive - how to find the parameters?

Thanks to Matth97 for the information.

https://github.com/kuba2k2/pico506
https://kuba.szczodrzynski.pl/posts/toshiba-t … l-hdd-emulator/