Looking for Tandon TM-100 service manual (w/schematics & waveforms)

Tony Duell ard at p850ug1.demon.co.uk
Fri Jun 12 15:40:35 CDT 2009

> I'm looking for Tandon TM100 service manual w/schematics; one of my 
> TM-100-2 drives in an IBM 5150 here started acting up (stepper motor is 
> stuck at track 40 and won't decrement) and I'm trying to find a useful 
> service manual to debug it. The two tm100 documentation/service pdfs 
> I've seen are a scan of a staple-removed 21 page or so packet (this scan 
> is missing pages 8, 9, 12, and 13) which has no schematics and is 
> missing some pages; The other is 7 pages scanned from part of a larger 
> document (100 pages or so), which looks like it would be useful if I had 
> the whole thing. Neither of them are particularly helpful.
> Does a scan of the larger tm100 service manual exist?
> Does anyone have a paper copy I could borrow/copy/etc?

At last! After far too many OS-advocacy messages, we have something 
related to classic computers. Can we have more of these please?

I don't have a full service manual, but I do have scheamtics (in front of 
me). The stepper control circuit is pretty simple, and I would have 
thought you could sort it out from schematics alone.

A disk drive exerciser [1] is very useful for doing tests on this 
circuit, but you can probably use your host system with suitable 
programming. What uou need to be able to do is send step pulses with the 
direction line in both states. 

[1] Those who know me will realise this is just about the only sort of 
exerciser I am ever likely to use :-)

There are 2 inputs to the stepper corciot from the disk controller, the 
step pulse and the direction signal. These are buffered by U2Ee ('14) and 
U3Dc ('04) respectively). The inverted step pulse (output of U2Ee) is on 

The stepper motor control counter is built from the 2 sections of U4C 
('74) along with hte XOR gates U5Da and U5Db ('86), which control the 
count sequence depending on the state of the direction line.

Now look at the clock input to this counter. It comes from U4Ba ('20), 
which gates the step pulse (TP12, as above). The counter can only count 
if the drive is selected and it's not writing. And if one other signal 
(output of U4Fc, '38) is high. This gate prevents the drive stepping 
outwards if the track 0 signal is asserted.

And I wonder if that's where your problem is. If track 0 is stuck 
asserted, then you won't be able to step outwards. The track 0 signal is 
based round U4Bb ('20), whioch combines the output of the track 0 sensor 
circuit (in a minute...) with the outptus of the counter flip-flops (thus 
making the exact trigger postiion of the sensor less critical). The 
output of U4Bb is inverted by U3Da, and fed both to the interface 
connector (via U1Fd, '38) and into the step clock gating circuit.

Start by looking at the output of U4Bb (also on TP8). If that's low, then 
the drive thinks it's at track 0, and won't allow outward stepping. 

The track 0 zensor it a microswitch (!) connected to P11. It's debounced 
by U4Fd and U4Fa ('38), cross-coupled to make an SR flip-flop. 

Now, I've had problems with the microswitches in these drives (not 
suprising given their age). I had to replace one in my HP9836A. It worked 
_some of the time_. Other times the track 0 signal would latch on (i.e. 
the switch made on one contact, flipped the SR one way, but didn't make on 
the other contact when the heads moved off cylinder 0, so the SR was 
never flopped back). 

You can still get a suitable microswitch. It's a normal V3 package but 
with a low (15g?) operating force. I found a suitable one from RS 
components (IIRC, may have been Farnell). Fitting it is not too hard, but 
you should do a track 0 alignment afterwards. I found that if you mark 
the position of the bracket on the chassis before removing it, and if you 
mark the position of the swtich on the bracket, and put everything back 
in the same place, it'll be near enough to work. I have a 'Microtest' 
alignment system so doing it properly was easy.

Let me know how you get on, I'l suggest further things to check/try.


More information about the cctech mailing list