Need help with Tulin TL-226

[famicomaster2]

Hello, I am attempting to repair a hard disk drive, a Tulin TL-226. The brake releases, the spindle accelerates to speed, and the drive seeks to cylinder 0. All of this happens reliably. At the end of the seek, the drive flashes it's activity lamp 8 times. It never asserts it's READY signal and does not respond to commands. I assume that this is an error code.

When I first got the drive, it was doing the same thing. About 2 days later it decided to come to life. After a low-level format, it threw a bunch of seek errors and after a reboot it no longer completed it's self test with the same symptom. It has not yet recovered. Heating nor cooling the drive have any effect. Opening the lid revealed some rather beat up limit bumpers, but neither adjusting nor shimming these made any difference. The drive always spins up, seeks to 0 and flashes 8 times on a loop.

Unfortunately, information on the Tulin hard disks is very scarce, a thread I found from years ago was unhelpful and the only other real source of information so far has been an archived Microsoft KB article which says not to use it with the original Xebec controllers because it uses a wedge servo.

I figure a manual, service documentation, a technical document, a brochure, anything will be more helpful at this point than guessing. My concern is that there is a possibility my DTC 7287 controller overwrote the drive's servo data during the low level format and "bricked" the drive, but I won't know until I see a manual.

Any help is appreciated, even if it's to let me know you have a spare TL-213, TL-226, TL-240 etc that you would sell me. Thanks.

 

[Chuck(G)]

"Wedge servo" is another name for "embedded servo" see: here. If you really did an LLF, you wiped out the servo data and your drive is now an interesting paperweight. See here, middle colum. Read the sentence that begins with "Moreover".
Sorry about that.

 

[famicomaster2]

Well, crap. How would you even go about formatting this drive without overwriting that data? Sucks that the thing is junk now, it's an interesting drive. I wonder why it worked for a little while after the format, long enough to get a couple seeks in.
I get the impression that this might have been a common issue with these drives, I imagine there is some way to reproduce that data.

The same Microsoft KB article lists the CMI CM-3426 as also having wedge servo data, although after a low-level format with the same DTC controller and that drive still works flawlessly. I wonder why the Tulin drive would be affected?

In this case, what controller would be safe to format this drive with, and is there a manual anywhere detailing this? At this point, I'm just gonna have to go get another drive. Know anyone with one for sale besides the guy asking outrageous money for untested unit on eBay?

 

[Chuck(G)]

Best case for these things is to "high level" format; that is, write every sector.
There was a similar issue with Drivetec/Kodak high-density floppies. You purchased the media pre-formatted. If you put the disks in another drive and formatted them, they became ordinary 5.25" floppies (IIRC $15 down the rathole). If you have any Caleb UHD (144MB) 3.5" floppies and format them, the same thing happens.

 

[famicomaster2]

How would you high level format it? It's an MFM drive. The controller will not recognize it and all tracks are unreadable without an LLF!

 

[Chuck(G)]

That's the thing--it comes preformatted, which is also why you can't use just any controller with it.

 

[famicomaster2]

I see. I wonder what controller was meant for use with it. That's rather annoying.

 

[Chuck(G)]

See this article from 1985 under #7 "Sigma Designs HHK-02000". Note that the authors say that the nice thing is that the "It is a nice surprise that the manufacturer has done all the physical formatting for you". What they don't say is that there's no other way with this drive.

 

[famicomaster2]

See this article from 1985 under #7 "Sigma Designs HHK-02000". Note that the authors say that the nice thing is that the "It is a nice surprise that the manufacturer has done all the physical formatting for you". What they don't say is that there's no other way with this drive.

I see. I wonder why the drive continued to work temporarily after the format and was only ailed after a reboot. I'll have to try and find the drive and controller paired, then.

 

[andy]

I had a pair of these many years ago when I was first starting to play with old PCs. One drive behaved like yours, but the other worked fine. I know I did the usual low level format using whatever random controller I had at the time. I'm sure it wasn't the original controller. The drive was still working when I parted with it.

I did try swapping PCBs between my two drives, and the problem was with the disk drive, bot the board. I opened the bad drive, and didn't see any obvious problems. I remember that the working drive made a distinctive series of seeking sounds just before it became ready. The bad drive didn't do this. I assumed it was doing some kind of self test, or calibration, and if that failed, the drive would not become ready.

The fact that it didn't work at first, and still worked briefly after you formatted it would make me think there's an intermittent problem with your drive. I've never heard of an MFM drive that could be bricked by low level formatting it. It appears to use a stepper motor, so I don't see why it couldn't be low level formatted in the field.

I know some drives have special formatting on some hidden tracks. My Rodime RO-203E uses a hidden track for part of the initialization process. Somehow that track got damaged, and it requires a special device to rewrite it. Fortunately, in that case there was a jumper you could cut to disable that part of the self test. The info was all in the service manual.

I recently acquired another TL-226. This one doesn't work, and flashes 6 times. I would also be interested in finding any documentation on it.

 

[famicomaster2]

6 flashes is what I get if I deliberately do not allow the heads to retract to cylinder 0, I.E. by placing too wide of a shim behind it. Sounds like your motor just needs some exercise or oil. If you give up on it, let me know, I'd be very interested in it.

Unfortunately, without documentation, I will probably never be able to find out what 8 flashes means. I have had the lid off this drive as well to nothing glaring - The end stops were pretty beat up, but they are rigid plastic. Making them larger or physically adjusting their position made no difference at all other than when the heads were stopped too far forward, which caused the 6 flashes mentioned above.
I'd also never heard of a drive being killed by LLF before, but the cryptic mention in a Microsoft KB article to explicitly not use it with a few controllers and the mention of "wedge servo" does make me wonder.
During the time the drive did work, it definitely did make an interesting power on seek test. Right now it just buzzes back to cylinder 0 and stops with flashing.

I did not know the RO-200 series had engineering data! Is this specific to the 203E or the entire 200 / 200E series? I wonder if it applies to any of mine, or the Ampex Pyxis 27.

It would be wonderful to get the drive working again or to find another, it was such an interesting unit. I collect strange mechanisms, primarily. Junk like this Tulin, my Panasonic JU-128, PTI PT-357R, etc make me happy.
Either way, thanks very much for the input.

 

[andy]

The Rodime manual is available on several sites. It has a lot of good information in it:

Rodime 200 series 5.25 Winchester Disk Drives User Manual : Free Download, Borrow, and Streaming : Internet Archive

Rodime 200 series 5.25 Winchester Disk Drives User Manual

archive.org

Interesting that you can get 6 flashes by blocking the head from returning to track 0. Mine appears to have no trouble finding track 0. Occasionally, it will seek across the disk and back again after power on. My assumption has been that it fails to read something specific that is supposed to be on track 0, or a hidden tack below 0.

Did you take any pictures inside the mechanism?

 

[Chuck(G)]

Still using an RO-203 (no "e") in a CP/M system here. At last try about 6 months ago, it was still working. We started out using the RO-100 series, spec-ing them at 7 and 14 MB. Rodime discontinued those models and "upgraded" us to the 200-series for no additional charge. Problem was that the bean counters had a problem with that--we had already contracted to customers for 7 and 14 MB system drives. To solve that we had to leave the BIOS supporting 7 and 14 MB without telling the customers that they really had 10 and 20 MB drives. Strange minds prevailed.

I used a couple of half-height 50 MB Rodimes on PCs. They didn't survive.

 

[famicomaster2]

The Rodime manual is available on several sites. It has a lot of good information in it:

Rodime 200 series 5.25 Winchester Disk Drives User Manual : Free Download, Borrow, and Streaming : Internet Archive

Rodime 200 series 5.25 Winchester Disk Drives User Manual

archive.org

I'll have to check that out sometime.

Interesting that you can get 6 flashes by blocking the head from returning to track 0. Mine appears to have no trouble finding track 0. Occasionally, it will seek across the disk and back again after power on. My assumption has been that it fails to read something specific that is supposed to be on track 0, or a hidden tack below 0.

This is how it worked as well, physically blocking the stack from retracting to cylinder 0 would cause it to jump forward and retry two or three times before it would blink the code.

Did you take any pictures inside the mechanism?

I did, here are a few. You can see a mark on the surface of the disk in the shape of a slider, I purchased this drive through a computer reseller who has maintained the same warehouse for nearly 40 years. It's possible this drive sat untouched for a very, very long time. Note the wide gap between the two platters, where a third goes for the TL-240. Interesting that they stack them this way instead of closer together. I wonder if there's a reason. The final picture shows the capstan and rotary band, along with the end stops up close. They are hard plastic, but they look pretty chewed up, which is why I was trying to widen them with tape or paper in the first place.
Here is a video of the drive spinning up as it does now: Tulin TL-226 fails seek test

 

[andy]

Interesting pictures. It looks like the head must have been sitting in the middle of the platter for a long time to leave that mark. It doesn't bode well for the condition of the heads or platters.

I think the end stops were rubber which has hardened from age. I'm more worried about what the outgassing rubber might do to the drive internals than the actual condition of the end stops. Rubber can give off corrosive sulfur compounds that tarnish metal surfaces, and leave deposits on things.

Can you see how the track zero sensor works? Is there an optical sensor, or switch visible?

I suspect the drive is looking for something on the disk when it initializes. If it fails to find that, it gives an error. As with the Rodime, it's probably on a track that's hidden below track 0. Maybe the controller compatibility warning was because some controllers might accidentally overwrite these tracks?

 

[famicomaster2]

Interesting pictures. It looks like the head must have been sitting in the middle of the platter for a long time to leave that mark. It doesn't bode well for the condition of the heads or platters.

After the low level format, I was able to successfully high level format and partition the drive via SpeedStor. All the seek errors and failure happened after running SpinRite 5 until it got stuck in a loop seeking the drive back and forth. Being 2am, I had just shut the machine down as it was and when I came back to it later the drive was dead again.

I think the end stops were rubber which has hardened from age. I'm more worried about what the outgassing rubber might do to the drive internals than the actual condition of the end stops. Rubber can give off corrosive sulfur compounds that tarnish metal surfaces, and leave deposits on things.

They were absolutely not rubber. They are very, very solid plastic. If they're rubber they must have been something very dense. I'm used to decaying rubber, since I work on a lot of junk cars and I've repaired my fair share of vintage disk drives. In my experience, rubber failures either get really sticky and soft or get very hard, but generally even expired rubbers can still be deformed with a screwdriver or knife. I could not cut through this with an x-acto knife when attempting to remove them, which led me to believe it was molded plastic.

Can you see how the track zero sensor works? Is there an optical sensor, or switch visible?

I suspect the drive is looking for something on the disk when it initializes. If it fails to find that, it gives an error. As with the Rodime, it's probably on a track that's hidden below track 0. Maybe the controller compatibility warning was because some controllers might accidentally overwrite these tracks?

There was no visible switch or sensor anywhere. The stepper motor is sealed underneath, I didn't pull the PCB off, but I didn't see anywhere for there to be a sensor. Inside the drive shows (obviously) nothing as well. Even blocking the heads from retracting, they will always settle in the same final position during this spinup. Without mechanical adjustment or the addition of material, the head stack never touches the end stop, even though it can travel further. If you push the head back to the stop and power the drive up, it will push itself forward into what looks like the same position. I think there's something else it's missing.

I figured that maybe this was a temperature issue initially, either with the HDA changing shape or something cranky on the PCB. When I got the drive and it didn't work, I was bringing it inside on a hot day. When it did work, the drive was much colder, and after sitting overnight it was unhappy again. I threw it in the fridge for an hour or two to no effect, though. Was a good long shot, it's actually salvaged a couple drives of mine before.

The only real way to know for sure what that code means would be either to manually dump and disassemble the firmware ourselves or find a manual. With a working drive, you might be able to "simulate" faults as well, I suppose.

 

[famicomaster2]

Hello, I did wind up tracking down a second drive, this one with a different issue and of a different revision. Parts of the mechanism are slightly different and the entire control PCB is different. It also did not come to me with a controller.

This unit started off non-working as well, they must not have been very reliable mechanisms in their day. After much work, the drive was for about half an hour spinning up and working reliably, but out of nowhere (no power cycle even), it fails mechanical tests in DOS. The power on self test seems to pass without issue, but requesting a full sweep causes the drive to seek a few tracks, make a sound like plucking a guitar string, and then flash an error code. I've seen both 4 and 5 flashes. I am still unsure what the codes mean. If anyone has any information, please let me know.

The hope was that with the drive "functional" I would be able to duplicate the apparently un-hidden wedge servo data and clone it to the other mechanism, thus fixing the other drive as well. At this point I would just like ONE example that works.