osiweb.org

I wanted to start a new thread in the hopes that people here can help me fix my OSI (MPI 51) drive. I've never worked on a disk drive before so please excuse my ignorance.

First I should start off by explaining what I have done so far:

- Replaced main drive bearings. Originals were shot.
- fixed (hopefully) severed wires leading from head to drive circuit board. Very fine wires and fiddly fix.
- replaced dying motor with one from another more modern MPI drive. Original was dying.
- checked the continuity of the cable coming from the OSI computer to the drive. All seems ok.

Basically as the drive is still not working I can not be 100% sure that I have fixed all of the above perfectly, however I took my time and did my due diligence on each of the fixes.

I have received some OSI boot floppies from Dave to test with, but I can not get the drive to boot from them.

I then decided to use a disk dump utility loaded via serial to give me a better idea what was happening. Dumping the contents of any disk (OSI or non OSI disks) I get mostly 00's dumped. An example of a typical track dump may look like this:

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 70 00 00 00 00 00 01 00 00 00 00 00
EC BF F9 6F FF F3 FC 7E 3F 5E FE FF FF 7F FF FE
FD FF FB 6F FF 9F DF 9F BF 9F 7F FF FB F7 67 FE
77 FF D7 FF FF BF FD FD BF 2F F1 F7 E1 E7 FF FF
DF FF FF FF FD FF 7F FB FF 8C FF AF 7E FD F6 FF
FF FE FF BF 5F 7E FF AF FE FF BF 7F FF 6A B7 7E
BF EF BD FD FF F7 7F EE FE FF 3F FF 7F FF FF 7F
FF 86 FB EE FF FF FE BF EF F2 BD FE 9F FA 7F FF
FF FD 2F 2F FF BF FC FA FD FF 7F DE 56 BF FF BF
FB BF DF FD FF E3 FF CF AF FE F7 ED ED EB 6B AF
DF F3 FF BC E1 FF 79 7E CF FE FF FF FF 5F FF CD
FF 7E FF FA DB 7F F3 F6 FE FF 7F EF 5F FE FF 6D
FF 7E FF 7D DF FF 7B EF FF B6 67 FB C7 7F FE ED
FD 2F 5F 7B FF FF FE D3 FF F7 FF FF FF FF FD FE
FF ED FD FF B9 FB E7 FB FF 3F ED FE FC FB FF FB
7F FF E1 BC FF FD F1 2F F9 EF F7 AF EB FF EB 7F
7F FF FF FF 7E BF FE 4F FE E6 FF AF DF 1C FF 3F
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00…etc. until the end of track is reached.

Sometimes the data starts at the beginning of the track, and other times it starts a little later as in the example above. Sometimes (especially on later tracks) there is no data shown at all, just 00's.

I then broke out my logic probe and tested the main drive connector on the MPI drive as it ran through the disk dump program, and these are the readings I get:

2) Density Select 1=Low/0=High - No signal
4) Reserved - No signal
6) Reserved - No signal
8) 0=Index - Looks to be pulsing high/low
10) 0=Motor Enable Drive 0 - GND
12) Drive Select 1 - No signal
14) Drive Select 0 - No signal
16) 0=Motor Enable Drive 1 - GND
18) 0=Direction Select - GND
20) 0=Head Step - No signal but pulse on track change
22) Write Data - Seems to be pulsing high/low
24) Floppy Write Enable, 0=Write Gate - No signal
26) 0=Track 00 - Low when on track 0, changes to high once head moves.
28) 0=Write Protect - High
30) Read Data - High/low pulse
32) 0=Head Select - No signal
34) 1=Disk Change/0=Ready - High/low pulse

That's about where I stand, and I'm stuck as to how to move forward. If anyone has any ideas I'd love some advice.

Much thanks again

Philip

Assuming you used (Ed's?) rawdump utility, then that is not what you should be seeing. There's a great article in the journals section of the archive (link here), Compute #20, p 136, describing the format of track 0 and the other tracks, and this does not match the expected format for track data. It would be enlightening to see a scope trace of the data, but that will be difficult on an analog scope. However, in real time, you may be able to get an idea of what the signal coming off the scope looks like; i.e., does it really look like square wave patterns in little groups? Also, you should check the index pulse to make sure it's the proper frequency (I think 300 RPM).

Dave

Hi Dave,
Thanks for the reply. Yes, I was using the rawdump utility. I wanted to get a better idea as to what the computer we receiving from the drive and I thought this program may give be a clue.

Tonight I battled with flakey connections to the back plane but I was able to get some readings using my old O-scope.

Firstly I tested the Index timing. I'm not sure exactly if I was using the scope correctly for this as I've never looked for something that pulses this slow. I set the timing on the scope to 20ms (per unit), and after a bit of fiddling I seemed to get a pulse every 4 units (see attached image)
(you can see the two blips at the bottom of the image)

So that means that it's a single pulse about every 80ms or 0.08 seconds. If my math is correct (and often it isn't) then 300RPM means 5 pulses per second (300/60) which is one pulse every 0.2 seconds. So from these calculations it seems that the drive is spinning more than twice as fast as it should be. Can this be correct??? I did change out the motor so maybe I screwed something up here.

As for the data line. I tried to get a reading off the scope. I had to adjust the holdoff so that I could lock onto something. Please take a look at the attached image.
Phil

I just quickly wanted to follow up my last post with something about the motors. Unless my calculations are totally off, the new motor is spinning way too fast.

just a little back history. Basically the original motor is dead. It's made by a company called Buehler. I was able to buy two new motors made by the same company and the new motors even has MPI written on them so maybe that means that they were meant for an MPI drive. The new motors look a little different in design.

The original motor has a connector that has the pins ordered as follows (left to right) black, red, blue, green yellow. The red and blue lines go to the bottom of the motor mechanism, and the green and yellow go to the top. Black is the GND line.

The new motors has the same configuration as to where the lines go into the motor mechanism (red and blue to the bottom of the motor, and the green and yellow go to the top), but the order is different at the connector. Green, blue, red, yellow. Thinking logically, I re-orderd the connector on one of the new motors so that it was consistent with the original and installed it, however maybe this was not the thing to do.

See attached for a visual description.
I just quickly tried a test and plugged in my other new motor (the one without the pin re-ordering) and it also spins fine. Confused.

***Update*** I tried my other motor without the reordered lines. The drive spins a lot slower for sure, but I think, after looking at the schematics (http://www.osiweb.org/manuals/MPI_B51-B ... Manual.pdf), that this setup is not correct. I believe the yellow and green lines (J4 - 28 and 29) lines are for the tachometer levels that helps control the speed of the motor through a feedback circuit, and the blue and red lines (J4 - 30 and 31) are for powering the motor itself. So basically I'm still unsure why the motor spins so fast.

Phil

Tonight I had a little time to read some of the MPI 51 manual, mostly focusing on the speed control area. I have to admit to not to understand it very well, but in general, I think I get the basic concept.

I took a scope reading at J4-28 which is the tachometer output from the motor.
(This was taken at 1ms and 5v per division setting)
When comparing the output to the manual it is certainly wrong...
The manual says that the frequency should be 2.5ms, but mine is at about 1.2ms.
The voltage should range from -3V to +13V, but my reading say it's at -10V and +20V.
(just incase you are wondering, I am using a modern power source for testing purposes, and have tested that I get +5V and +12V at the inputs). Now I'm not really a whizz at using the scope, but I have check my settings and I think it's reading correctly. So, I'm just not understanding that I'm seeing here, or what could cause this.

I have checked a little further down the chain at 4F-5 and the frequency is still at about 1.2ms and the voltage is at about 15V.

So my readings are telling me that I'm getting about twice the frequency and twice the voltage!!!! Oh dear!!!

***UPDATE*** Just some measurements from the motor side. At J4-31 I am seeing about 10.5V. At 4F-1 I am seeing about 11.8V. Again this is not what it's supposed to read.

Phil

Just some more readings from the 4F ic.

4F-5 (5V/DIV 2ms) 4F-2 (2V/DIV 2ms) 4F-6 (1V.DIV 2ms)

4F-6 (1V.DIV 2ms) Some quick observations:

- 4F-5 is about 12V peak to peak (should be 8V according to the manual). Seems to dip slightly below 0V. Pulsing too fast
- 4F-2 is certainly not a square wave (low voltage is about 4V, high voltage is about 6V)
- 4F-6 is at about 0.7V but with a very slight ripple (visible when the Volts/DIV setting is cranked up) which seems to be in the millivolt region
- 4F-7 seems to be at about 0V, with a very slight ripple (visible when the Volts/DIV setting is cranked up) which seems to be in the millivolt region

If anyone can suggest any more areas to probe, or suspect parts, I'm all ears.

Thanks again

Phil

Just thinking about this a little more and looking into how a comparator (the LM339 @ F4) works.

"A comparator is a circuit that accepts two voltages, V1 (in my case it's 5V coming in at 4F-4) and V2 (in my case it's an oscillating input on 4F-5), and outputs zero volts if V1>V2, or outputs a positive voltage level if V2>V1"

If you look at the images from the o-scope that I posted previously, this is obviously not what I'm seeing on the output (4F-2). What I'm seeing is an oscillating sine wave like signal, and not a pulsing on/off digital like signal as I would expect. Can I assume from this that the LM339 at F4 maybe funky?

Phil

After some offline discussions with Dave, I decided to replace the LM339. The good news is that it now seems all the scope signals are now back on spec, and the drive now spins at the correct speed. I am able to get a dump of the contents of a OS65D floppy, but I am getting spurious errors. I think I next need to do some alignment calibration of the drive.

Phil