I suppose the drive has 4 heads only, so the setting in the SpeedStor looks more appropriate.
Head select lines are forming a binary address, in this case 0-3 are valid, which are placed on HS2(0) and HS2(1). HS2(2) is most likely not used, so the level on that line is irrelevant.
The drive probably has the configuration of 615/4/69 (CHS), which gives 82.8MB.

I was also confused by the first doc Franc linked, it has HS0, 1 and 3, no HS2, the 2nd one mentions 4 head select lines, but the ST506 spec has only 2, mentioning a 3rd one as reserved for future products.
http://minuszerodegrees.net/manuals/Sea ... Manual.pdf

I found one similar drive here: https://aukro.cz/hdd-3-5-seagate-st-110 ... 6992401046
can't see the config anywhere on the label but from the defect list i can only see H0-H3. What does the defect list of yours look like?

pepe

_________________
Adatmentés - Data recovery

The ST-1100 has 9 heads, but only 17 sectors/track. I haven't been using SpeedStor much except for a failed attempt to make it work properly on a Pentium III computer... I probably need something older than that.


Well, AFAIK the original specs were for 3 head select lines; some drives - like the ST-251 even - has 6 heads. I've used the Myarc HFDC for the TI-99 series computers, and at default it only supports 8 heads but the Write Precomp line can be altered so it gets the H3 signal on pin 2, which is what I have done... but it hasn't made any difference; probably because it doesn't give 3V on that line for some reason. I'm confused... c".)

The stason.org document has an error in the HS definitions. I explained this in my earlier post. "Head Select 2(3)" should be "Head Select 2(2)".

ST1100 is an old, dumb MFM drive. It does not do any translation. That came later with the arrival of IDE. I don't know where you got "615/4/69 (CHS)" from, but that stason.org document clearly states that this drive's config is 1072/9/17 native.

_________________
A backup a day keeps DR away.

well, if stason document has an obvious error i am doubtful about the rest of it too, however, it is probably right about the CHS indeed (didn't see that before :s) . Now i found some similar drives with better readibility of defect list, so i am ok with the higher number of heads. I just calculated that value and seemed viable, sorry (however, i never stated it as a fact).
For 9 heads it does indeed require 4 HS lines, so it is a question wether the controller supports that, as it seems like the standard have been developing over the years and early specs say nothing about HS3 line.

pepe

_________________
Adatmentés - Data recovery

By the way... are you from the Czech Republic and know that site, aukro? I can see they didn't manage to sell that particular drive, but I can't see anywhere how to write to them and ask...

As standard the controller doesn't; it uses pin 2 for precomp. It does have the logics to do it though, so a modification allows it to use it for H3 instead... and I made that modification according to the instructions at hand.

However, when I have a terminated 20 MB Tandon (CHS 615/4/17) at the end of the cable and the ST unterminated in between, the voltage on pin 2 reads only 0.16V - now it sits alone with termination and now it reads 3V as it should... I still can't make it format with heads 0 and 1 though, but the rest seems to format properly.

So it definitely seems to be caused by a faulty termination or something - or perhaps the cable, if the drive is more sensitive to fluctuations than the older ones I have.

Okay, so now i've measured the voltages on pins 14 and 18 during format. 14 sits at around 3.2V and drops to 0.3 at head shift, and 18 is a bit lower at around 2.9V and about the same as the other when low.

Could it be caused by the power supply? I do have some options to swap even the old TI-99 expansion box; perhaps I should try that and see how it turns out.

I could simply drop this drive, but except for really wanting it's higher capacity etc, I'm also very stubborn about such things, haha

Those voltages seem OK to me. The difference is probably due to tolerances in the resistor pack. Compare the voltages against the other HSx pins.

_________________
A backup a day keeps DR away.

Which IC is connected to HS0 and HS1 on the HDD? Perhaps this IC is faulty? If you can access this IC while the drive is powered, measure the corresponding output pins of this IC.

_________________
A backup a day keeps DR away.

I haven't been able to trace those two pins further than the termination connection. However, I have an identical PCB, and that one behaves in the same way. Hmm, it's very weird - and it drives me nuts! c".)

Anyway, I'll see if I can identify that IC and measure it. B-)

I've examined other circuits and they have a 16-pin 7445 BCD-to-decimal decoder IC, either directly connected to the HS pins, or with an additional logic gate in between. This chip decodes 4 inputs into 1 of 10 outputs. That's enough for 10 heads. Look for this IC, or something similar. If you upload a good photo or CCD scan (much better), I could help you locate the IC.

Datasheet for DM7445 BCD to Decimal Decoders/Drivers:
https://www.farnell.com/datasheets/59371.pdf

_________________
A backup a day keeps DR away.

Thanks, I really appreciate your help with this! I'll take a photo and upload it then!

So, here are a couple of photos... since I have a removed PCB, I could include both sides of it. Pretty good quality with that iPhone camera, by the way.

I can find only one IC with 16 pins however, and that doesn't seem to be the BCD-to-decimal decoder...? c".)

Earlier MFM drives used discrete logic chips. This model uses a microcontroller, so the HS bits would be handled by one of the unobtanium ICs. I don't know where to go from here. :-(

_________________
A backup a day keeps DR away.

Yes, that's what I suspected...that it was one of these chips.

Anyway, thanks for your efforts to help, I really appreciate that! Perhaps I'll find a solution somehow some day.

On the bright side, I just received a cheap 44 Meg Toshiba from the U.S. which gave a tremendous amount of CRC/ECC and SYNC errors as well as the infamous track 0 error, initially... but I guess it just needed some massage; several attempts to format it finally paid off and it seems to be fully functional now.

I wish they would use a different material than unobtanium, though... :-/ haha

Oh, by the way... there ARE a few surface mount resistors near the 20 pin data connector, which doesn't read the resistance they should - one 221 ohm and one 331 ohm. I was wondering if that could cause such issues?

221 = 22 x 10^1 = 220

Carbon resistors can degrade over time. They may read higher than their rated value. If you are measuring 3.0V, or close enough, at the terminators, then they should be OK.

_________________
A backup a day keeps DR away.

"7951" appears to be a Seagate in-house part number prefix, so any chips with these numbers would probably be special runs for Seagate.

_________________
A backup a day keeps DR away.

Yes, of course; I forgot that part with the "1"... so 220 and 330 ohm. Perhaps in the winter I'll give it a closer inspection; there's little else to do here in Denmark at that time of the year.

Once again, thanks for your tremendous help - I also got wiser from this, and that's definitely worth something, too!