SA-800/801 saga continues / schematic request
cube1 at charter.net
Tue Aug 25 21:34:42 CDT 2015
On 8/24/2015 8:21 PM, Jay Jaeger wrote:
> On 8/23/2015 9:39 PM, Jay Jaeger wrote:
>> When I connect up drive "X", the +24V series pass transistor gets very
>> warm and the +24V drops to around +8.5V (regardless of whether or not
>> the stepper motor leads are connected). On the circuit board, a
>> Shugart ASIC 16270-1 gets quite hot, and is clearly dead. This chip is
>> labeled WR/CHNL in an SA-850 schematic, and is listed in the Shugart
>> spares catalog as "IC, Write Channel". It is indeed fed from the +24V
>> line on pin 15, and connects to the erase and center tap terminals on
>> the head. [I have not measured the head to see if it was damaged - nope
>> not]. I expect that if I pull that chip, the board will not overload
>> the +24V supply.
>> When I connect up drive "Y", the +24V series pass transistor on the
>> power supply gets very hot, but nothing on the circuit board gets hot at
>> all, and the +24V drops to 0. The board measures as a dead short with
>> an ohmmeter (even "X" shows *some* resistance). The stepper driver IC,
>> a standard part UNL2074 (quad darlington power transistors in a 14-pin
>> DIP no less) has a suspicious bump on its head, so I will probably start
>> by pulling it and see what changes (I hope that is it), and if that is
>> the issue, I will also check the voltage clamping diodes in the output
>> circuit and the stepper windings and the 7404 that drives it to make
>> sure it wasn't collateral damage. I expect regardless of the cause, the
>> dead short will be not *too* difficult to find and correct.
>> So, now the saga of the Altos power supply is fully understood. It was
>> the second drive from the DSD-440 that actually killed it. Fortunately,
>> the DSD-440 power supply was apparently robust enough to put up with the
>> With a little luck, I can make one whole SA-801 out of the pair
>> (focusing on "Y" since I have no source for that 16270 ASIC), and then
>> use the mechanism of "X" with the board from the Altos SA-800 to make a
>> working SA-800 for the Altos.
> Some GOOD NEWS for a change. I pulled the UNL2074 from drive "Y", and
> the short remained. The next most obvious component for a dead short was
> actually a little 4.7uf filter capacitor - and indeed that was the
> culprit. The board from drive "X" donated its corresponding capacitor,
> as the odds of me fixing that board are not very good, owing to the
> 16270-1 Shugart ASIC. Returned the UNL2074 back into the board on "Y".
> Erased (with a tape eraser for open reel tapes back in the 60's),
> formatted a floppy and tested, including interchange with the original
> second drive in the Altos, and all is well.
> Tomorrow I get to test the 2nd 801 (drive "X") mechanism with the board
> from drive "Y". I expect it will be fine as well. Then I can swap out
> the electronics from the original first Altos drive onto drive "Y" (they
> are plug compatible that way) to get another working drive.
> BTW, the replacement FD-1791B-02 (I actually tested to) did help with
> the hung floppy controller after errors syndrome - it happened
> significantly less often that with a FD-1791B-01 .
Both SA-801 mechanisms tested fine (with the board from "Y"). So now
the mechanism from "X" and the board from the original are in the Altos,
as the second drive, and doing fine. Drive "Y" was restored to its
original condition. The original SA-800 had its fancy new bearings
swapped back for the originals, and has the bad board from "X". - so
three drives to make two good ones. Drive "Y" and the bad drive will
take up residence in the Data Systems chassis for now.
All four of my good drives (the original Altos SA-800, the SA-801 drive
"Y" and the "hybrid fix" drive and my Siemens FD 100-8 (which says it is
an FDD 120-8 on its serial number plate) also tested fine.
More information about the cctalk