Tuesday, November 5, 2019

Digging into load failure on Telex 8020 drive A


When a threading attempt failed with the latest problem, it left the end of the tape down in the upper vacuum column instead of threaded through to the tape path down past the heads.

I trimmed the bent end of the tape and now it fails getting down near the bottom of the tape path, perhaps 6" shy of reaching the take-up reel. It times out but with a different situation inside the hidden area of the drive.


I observed that the speed with which the take-up reel rotates is much faster than the very slow rotation of the supply reel. Since that reel lets out the tape to thread through the path, if it is too slow then there is insufficient time for the tape to make it all the way to the take-up reel.

The maintenance manual has a procedure to set the rotational speed of the reels during the autoload. Unfortunately it involves a special Switch board that replaces the Control Logic board; I don't have this board nor do I have the clear plastic vacuum chamber cover I mentioned earlier, nor the card extender which is also used for the procedures. . Still, I know the rotation speeds based on the desired settings. I can observe the speeds during the stages of the autoload and grossly adjust the speed for any stage where it appears wrong.

The supply reel speeds during the stages are:
  • While looking for the end of the tape to start threading, counterclockwise at 2/3 rps
  • While threading into the path, clockwise at 1/2 rps
  • When lowering tape into the upper vacuum column, counterclockwise at 1.3 rps
The take-up reel speeds during the stages are less varied:
  • While threading into the path, clockwise at 2.1 rps
  • when lowering tape into the lower vacuum column, counterclockwise at 2.8 rps
Each reel has its own power amplifier PCB in the card cage - slots 1 and 2 are assigned to reel amplifiers. The pre-amplifier for both reels is a single PCB in slot 4. The Control Logic board in slot 5 gives commands to the pre-amp board, where it generates the drive voltages that are amplified by the boards in slots 1 and 2.

I may have problems in any of the cards in slots 1, 2, 4 and 5 as they are all involved in loading operations. I began to study the theory of operations and the schematics to figure out which board(s) could result in overly slow supply reel motion. Further, I set up some observation points to narrow down the fault, if any, to specific circuitry.

I took video of the drive attempting an autoload, allowing me to go back and assess speed after I break the sequence into the first couple of stages. This gave me the ability to validate the speed against the targets above.

I can see that the supply reel is moving far too slowly both when locating the tape end and when trying to thread it through the path. This is the cause of the failure to autoload, but we need to dig further to see which board at fault.

I swapped the two reel power amplifier boards. If the take-up reel slows down and the supply reel goes faster, I know it is one of those. If the problem remains associated with the supply reel, the potential locations of the fault are:
  • Reel pre-amplifier board section controlling the supply reel
  • Supply reel motor itself or wiring to its field coil
  • Incorrect adjustment of both speeds for the supply reel
  • Error in control logic board signals to the reel pre-amplifier
The swap of reel amplifier boards made no difference to the behavior of the drive. I then swapped the reel preamplifier boards in drives A and B to see if this changed things. If anything it was a bit slower than before, so I have to rule out adjustments on the reel preamp board in addition to the previous exclusion of the two reel amplifier boards. This means it must be a fault in the motor or wiring itself as everything else has been swapped.

Digging further into the schematics and theory of the reel motors showed me two areas to check next. I have to verify that the field coil ramps to the proper value on the supply reel motor, something I can detect using my voltmeter on the sense resistor. The other potential source of the slow drive is the autoload analog switch and its reference voltage, as these supply the voltage that is used by the preamplifier and amplifier to set motor speed.

I can do comparative testing between the take-up reel voltages, since these seem approximately correct, and the supply reel voltage. I don't know the voltage to speed function; could be linear with voltage or some other curve that makes it harder to determine the desired supply motor value.

Similarly I can do comparative testing of the field current as a way of checking that the supply motor itself seems healthy. That is, compare the voltage on the take-up reel sense resistor with the voltage at the supply reel resistor.

There are practical considerations that may complicate getting the meter on the sense resistors or the command voltage to the reel pre-amplifier board. I had to study the schematics and then the boards and backplanes to figure out how to access what I need.

The analog switch and reference voltages are produced on the reel preamplifier board, which I had swapped. This tends to rule out the problem being on that board, but I found test points to measure what I wanted to see. The test points are deep back on the PCB, suitable for access when using the extender card I don't have. Instead I need to attach mini-grabbers with long leads to carry the measured signal out to where I can reach them.

First to test were the reference voltages used for speed control - +4.8V and -4.8V. These were good of course since the take-up reel wouldn't work properly if they were bad. Next up were the driving levels for the two reels, picked off from two other test points.

The test points showed me the output of the analog switch levels. The take-up reel signal was at 0.7v when it was rotating during threading. The supply reel signal was at 0.8v when seeking the end of tape and then about 0.7v when threading. Based on these I would expect both motors to turn at the same rate, but they are not.

I then moved to the signal that exits the reel pre-amplifier and drives the reel amplifier, which I could pick off from the backplane easily. I hooked up to the take-up reel drive first, measuring
2.1V on drive 1 and .7 volts on drive 2 lines. I then hooked to the supply reel drive signal and saw 2.7V on drive 1 and .7 volts on drive 2, blipping to over 1V for a brief period during threading.

This suggests that the supply reel should be rotating much faster than it is, based on the drive current from the reel pre-amplifier board. Next up I starting measuring the current going to the two motors, making use of the sense resistor built into the power amplifier boards. This develops a voltage across it based on the current going into the motor.

The goal was to compared the voltage on the sense resistors of the supply and take-up power boards. That will tell me what the electronics are driving through the main coil of the motor. The results were odd, with the faster rotating take-up reel motor producing 21 millivolts and the slow supply reel displaying over 600 mv.

Since the sense resistor is 0.1 ohms, we can calculate the current flowing by I = E/R as 210ma on the take-up reel and 6A on the supply reel. These motors are designed to peak at about 22A of current. The higher current in the supply reel seems like it is trying hard to get itself spinning further but having no luck.

The preamplifier board sums the target speed and compares to the speed it has projected based on the current, acting as an artificial tachometer. The output of the summing point is the drive voltage to the motor.

There are two test points, TP5 and TP11, which let me see the voltage presented the power amplifier boards for the two reels. These will range from -10V to +10V depending on direction and speed. They are set at about + or - 0.8 volt when the reels are turning.


I cross wired the motors (drive B reel motor hooked to drive A electronics) to see how fast the reel turned during load. It looked somewhat faster but nothing like the take-up reel rate. Not sure this was enough to explain the load problem.


I looked closely at the supply motor rear and found a bizarre repair was done on it, with scotch tape wound around an outlet that is otherwise closed with a screw placed in the vent opening. I compared it to the motor on drive B which appears intact and correct.

Outlet on drive B motor (example of a good fitting)

Sketchy scotch tape repair on the suspect supply motor
This rear section of the supply reel feeds pneumatic pressure to the hub lock, to keep the tape itself from coming off the hub. Unless I feel air escaping, this should be fine. I have no noticed any issues with the tape seeming loose on the reel but I will test it. For now, this is closed as a non-issue.


I decided to switch over to restoration of drive B for a while, moving boards from the logic cage if necessary until it began to attempt a load. I moved the vacuum and blower hoses over to the other drive and shifted the main power input too.

No comments:

Post a Comment