On 08/14/2015 09:27 AM, Jay Jaeger wrote: Is this a drive that you've had in service for some time, or did you just take possession of the drive? If the latter, I suspect that someone's already tinkered with the drive. The first place that I'd look would be the spindle assembly itself--when you pull the spindle drive pulley, things tend to fall out all over the place. There's a significant preload on the spindle and bearings. The hub clamp itself more or less "floats", and centers itself, so no real issues there unless the clamp itself is damaged. My own worthless .02, Chuck

On Fri, 14 Aug 2015, Jay Jaeger wrote: If all thge disks were formatted in other drives, it could even be one of the other drives that actually has the problem. One of the first tests should be to take a virgin/bulk erased disk, format it, and WITHOUT unclamping, write and read it. If it reads/writes fine, then the clamping MIGHT be out of alignment, but is at least stable in its [wrong?] position. If you still get wobbles, then something is loose and unstable in its position. Then unclamp, remove, reinsert, reclamp, and try it. And, how well do disks formatted in this drive do on other drives?

The fact that it sometimes centres properly suggests to me that it is not excessive runout on the spindle, but none the less the first thing I would do is remove the clamp arm and check the inside surface of the spindle with a dial guage while turning it by hand. Then probably take out the spindle and its bearings and check all the parts are there (compare with the service manual or the good drive). If something there is coming out of place it will cause big problems. Take the centring cone apart (there are lots of bits and at least 2 springs) and again check for anything missing. Make sure the spring are in the right places, if you swap them over the cone will not expand properly and will thus not centre the disk properly. Now look for wear on the 2 main plastic bits of the centring cone. If the inner part sticks, again it will not centre correctly. If there's wear on the outside of the larger part, the disk may not run true. -tony

Not necessarily play. Anyway, half a track width is just over 10 thou and I am not sure I could notice that amount of runout by eye. Also I have learnt by bitter experience to check everything, otherwise you discover (far too late) that the thing that was 'obviously right and didn't need checking' was the problem Are you sure. The cone is in 2 parts. Normally there is a spring between them inside the cone and a (stronger) spring between the cone and the clamp arm. What exactly did you swap? The cone and internal spring? The obvious spring on top of it and all the washers? The clamp arm itself (check for burrs round the hole where the cone spindle goes)? If the cone is alright then it pretty much has to be the spindle itself. Are you sure the spindle bearings are correctly fitted with all the parts shown in the service manual? -tony

In every 8" (and 5.25") drive that I've seen the spindle (the bit driven, maybe indirectly, by the motor) has a female cone on the end. There is then the clamping cone, often plastic which fits into it through the hole in the disk. That plastic part may be shown as one piece in the service manual as individual parts are not available as spares, but most of the time it's actually 2 bits with a spring inside. If it's not the clamping cone then it pretty much has to be the spindle. A comment about pot metal got me thinking. If the cone has swollen due to oxidation of one component of the pot metal, it may be that the plastic cone is compressed too far too early and the media slips out of alignment. I would inspect the metal part of the spindle very carefully. I would check for excessive runout of the conical section. I would look for damage (burrs, etc). Do you have mechanical measuring tools (callipers, etc)? The ball races are most likely standard parts and can be obtained. Other parts may have to be made. My feeling is similar, that the clamping cone allows the media to slip about before it is finally clamped so that sometimes it works, sometimes it doesn't Firstly I think it's a long shot that somebody else will have had this and secondly even if they have had the same symptoms it might be due to a different cause. Better to actually work out what is wrong with _your_ drive (sorry, but I have never gone along with the idea of stock faults and cures [1]). [1] May be OK if you have a lot of similar units to repair, say you are a repair shop and you keep seeing the same model of TV come in. Stock faults may then let you get 90% off the bench very quickly. But they are not much help with one-off machines like we tend to get on this list -tony JRJ

How does that work? The disk media must fit around that male cone. Is the complete disk carried by the clamp arm or something? I need to find one of these drives to examine. Seems like an odd way to do it. The metal cone is not going to change in diameter. So the hole in the media has to just fit it, if too tight the media will be distorted, if too loose it won't centre. -tony

I have just pulled the case on one of my HP9885s, which uses an SA801 chassis and HP electronics. The spindle clamping mechanism is as you describe. I took the cover plate off the clamp arm and looked at the (female) cone, etc. The cone is in 2 parts, with a bearing between them, the outer housing doesn't rotate in normal operation. There is also a spring ring round the outside of the nylon fingers. BUT I don't see how any of this has anything to do with centering. If the disk is pushed down onto the male spindle end (and yes, the disk guides do move with clamp arm) it will centre. So all that really affects things is the metal spinde. I suppose gross errors in how the clamp comes down could do it, but unless the clamp arm is hanging on one pivot only, or something, I don't think that's it. It virtually has to be a problem with the male cone end. I think you need to measure that carefully and compare the measurements with a good drive, look for burrs and wear. Remember that 10 thou is not much. What if, due to some problem with the spindle cone, it clamps on one side before the other? That could cause the disk to skew round and loose centering I think. -tony The part of the spindle on the media of the side of the deck looks like this: x x x x x x y y xxxxxxx xxxxxx At one point I thought, MAYBE, there was a little grouve at position "y" that I could catch a blade of an Xacto or tiny screwdriver on on the bad drive, but not the good one - but I am not at all sure of that. If that is true - if there really is a little groove there, then that would well explain the symptoms. It could get off center, and it would NOT precess. I would think that a spindle with a bearing issue would precess, so it would never work during an entire floppy test - it would always exhibit some wobble. I can see the entire nylon finger piece finger - it is completely exposed except for a small bit under a screw and washer. It has a bunch of fingers attached to a disk - all one nylon piece during manufacture. The fingers can spread out. The fingers have a flat part on them which then presses the media against the flat part of the spindle. They have a spring around the outside of all of the fingers for tension. Besides, as I have explained, I have switched out the entire assembly with the fingers to no avail. The plastic fingers EXPAND on this design, AROUND the spindle, not compress. No, but I could get some calipers. They aren't expensive. But if that is the problem, it is a lost cause for me to fix anyway. And I'd have no way to make them. And no point - it isn't like the world is short of these drives. Me neither, usually, but this one has me mystified, so I was curious of someone else had ever seen similar symptoms.

It strikes me that there are 3 main subassemblies associated with clamping the disk : The cone; the clamping arm; and the spindle. If you have proved the first 2 are OK by using them in another drive (admittedly not a perfect test, but it suggests they are not the problem) then I would look very carefully at the spindle. Making a replacement cone is possible, but not easy, given the accuracy needed. It appears that the spindle in my drive is in 2 parts, the cone is pressed onto the end of a metal spindle. I think if I had to make it, I would make the spindle rod first, rough out the blank for the cone, leaving it well over size then press the blank for the cone onto the end, hold the spindle in a collet (a 3 jaw chuck is nowhere near accurate enough) and then turn the cone itself. -tony

On 8/15/2015 3:13 PM, Jay Jaeger wrote: Got the power supply series pass transistor out from the backside without all *that* much difficulty. Soldering irons with temperature sensors are "the bomb", though I may have been able to manage it with the same 45W de-soldering iron with built-in bulb I used to clean out the holes. The power supply was definitely a victim of over-current: the series pass transistor is blown open (2N3055, fortunately and oddly, available in series with the output - that one is actually cracked. I have a spare of the same resistance and probably greater wattage as that one that I can pull from a different dead/unused power supply. The other resistors are all ordinary 1/2 W. The transistor driving the series pass transistor seems fine based on a simple ohmmeter test. So sometime tomorrow or so, I ought to be able to fix it. (It fooled me for a little while to - even with an open series pass transistor, there is still 24V output - it only dropped out (to about 2V) when the head loaded on a drive.) I am still not 100% sure if I shorted something while I was testing - the test drive was placed on top of the system, or I did some other silly thing with the jumpers such that two steppers ended up active at the same time (I don't think so), or if the SA-801 is just bad. I'll have to rig up a test where I power the 24V to the SA-801 from a current limited bench supply I have before I try it again. JRJ

On Mon, Aug 17, 2015 at 5:22 PM, Jay Jaeger <cube1 at charter.net> wrote: Isn't the center lead also the tab? Yes. There would have been a Qbus interface card for that. -ethan

Nothing to do with this drive (in fact I have never seen this in 8" drives) but some direct drive 5.25" drives have a left hand thread screw holding the rotor to the end of the spindle. The HD one in my PC/AT does, for example. Do not ask how I found this out! -tony

On 8/17/2015 11:14 PM, tony duell wrote: Yeah, but the TIP31A part is cheap, and is free air mounted on its leads - it isn't screwed down. I'd rather do the repair nicely. Anyway, it gave me an excuse to order a 750 ohm resistor so that replacement matches, instead of the 2 x 1.5K parallel pair I had put in. JRU

On 8/18/2015 4:11 PM, tony duell wrote: Yes, I do have them. But it really isn't worth it - the parts are due to arrive tomorrow. ;) Thief River Falls, MN isn't very far from Madison, WI. ;) Already replaced all the resistors that showed any sign of stress. JRJ

On 8/15/2015 11:59 AM, Chuck Guzis wrote: Agreed. But the darn thing is cast, with lots of fillets for support. I wonder how it would ever get warped, short of something REALLY heavy being dropped on it, or it itself being dropped from a considerable height - and there is no evidence of that. However, I thought I would test an SA 801 (jumpered as an SA 800) on the thing, so I could test the controller at double density. When I did that, something died (smelled electronically burn-y) after a few seconds of OK operation. Not sure if it was one of the tantalums or the stepper or something else, though. Anyway, I now have a drive that I can pull the clamp arm out of (the top guide frame, as you say) and try in the failing drive, to see if I can nail it down to that part of the spindle, just out of curiosity. Don't have a wiggler gauge, or anything to use it with.

It could be an indication of pot metal failure. I had one drive that the spindle from the driven side was clearly swelling. I replaced it from a spare parts drive. What happens is that the metal has oxidation at grain boundaries. It is said that it is caused by small amounts of contaminating metals in the mostly zinc pot metal. Aluminum has been said to be one of the problem sources. Dwight

On 08/14/2015 02:58 PM, Jay Jaeger wrote: The way most of the old 8" floppies work is they have a cylinder with an ID that matches the ID of the floppy hole. The floppy sits against the face of the cylinder, and a plastic, springy cone is pressed into the cylinder by a bearing and spring on the loading arm. This arm usually also lifts the head off the media when in the open position. So, the first thing to check is if the plastic springy thing still has all its fingers. If fingers have broken off, the result is obvious. If all the fingers are still there, check for smooth sliding into the cylinder. If the fingers have gotten bent or rough, the clamp may not always seat reliably against the hub. Finally, the spring or bearing may have a failure, not driving the clamp into the spindle. Almost certainly, careful examination of these parts will reveal the problem. Jon

On 08/15/2015 08:55 PM, Jay Jaeger wrote: Yes, I should have - sorry for the fumble. Jon

On 08/15/2015 09:49 PM, dwight wrote: That's the conclusion I came to--for a way to see how the whole mechanism works, see Shugart's patent 3,898,814 illustrations. I believe that the earlier Micropolis 5.25" floppy drives worked the same way. I recall that when they first came out, they'd mangle the hub area of a floppy because the spindle motor often had been turned off by the host. Eventually, the design was modified so that the motor spun up as the drive door closed, which solved the problem. Before that, Dysan had experimented with putting a reinforcing ring on their floppies. By the time high-density drives came out, the problem had been solved so there was no need to reinforce the HD floppy hub area. --Chuck

On 08/16/2015 08:46 AM, dwight wrote: Where I saw the problem was with the Micropolis 77-track (single sided) 100 tpi drives. --Chuck

On 2015-Aug-14, at 11:04 AM, Al Kossow wrote: OK, was wondering along those lines, had found the big chip - the D7220 - mentioned as a controller in intel bubble doc but didn't know what sort of interface it presented.

On 8/14/15 11:28 AM, Brent Hilpert wrote: The docs should be on bitsavers.

Hi Holm, If you don't find anyone else I'd be willing to do so for you. Regards, Todd * Holm Tiffe <holm at freibergnet.de> [150814 14:15]: