Power Designs 6050A strange faults

[zike]

Barely survived the HP 62xxA power supply fix-a-thon, and walked right into a nest of broken Power Designs supplies.  Here we go.

First up is a 6050A "Uniply" (isn't that a tire?). Would not quite make rated current, and the meter was intermittent. Latter was susceptible to "percussive maintenance" but by the time I got to it, you needed a hammer.  For reference, this is 6050A serial #501097; component date codes suggest early 1975.

The meter occasionally registered less than actual voltage/current, but was mostly either accurate or dead. Drive circuit tested fine. Popping the meter case open (still factory-sealed, I think) revealed a wirewound series resistor with a carbon comp added in parallel, presumably as trim to calibrate the movement for 1 mA FS. The resistors appeared to have been soldered by a child; big globs of cold solder, no wetting. Crimp to the galvanometer lug was also loose. 

Reflowing the joints and reassembling the meter completely resolved the fault.*

Here PD used a Modutec "S" meter (still made by Jewell I think). Not cheap by any means, but there couldn't have been any visual QA as it was glaringly shoddy. I wonder if anyone else has seen trouble with these flush-mount meters?

The current range was more of a puzzle. I was able to reach spec with trimmer R33, but it was near its limit; and rechecking a few minutes later, the trip point had drifted again. Eventually I noticed that the 0.5A/5A current range switch (coaxial with the current limit knob) felt a little crunchy. There should be only two positions about 30 degrees apart, but with a scant fingerful of extra torque, I found it spun 360 degrees (!). Range flipped up and down chaotically.

This control comprises a Clarostat 1k wirewound pot with a piggyback two-position rotary switch on a coaxial through shaft. No longer available, of course. Nothing to lose, I pried open the tabs and pulled the switch apart. The shaft engages a spring loaded over-center rotor carrying the contacts, but its motion is supposed to be limited by a metal key captured between tabs on the housing. The key was intermittently missing the tips of this ‘fork;' it may have gotten bent by an impact to the knob. I gently bent it a scosh to stay in its lane. Reassembly was tricky, as the drive needs to engage the over-center mechanism, which is spring-loaded. Eventually I figured out to engage the drive tip with the parts at a compound angle, and then rotate axially (to preload the spring) while also tilting the halves back into parallel.

The switch appears to work properly now, not sure how much I can trust it. Certainly not checking how much torque it takes to bust it again. Anyway, the current range is now stable and the trim wants to be much closer to its factory setting (marked in red paint by PD, of course). I believe the switch was not making reliable contact in the 5A position; it needs to be << 0.1 ohms, as it appears in series with the current-sense shunt.

Interested to hear other repair stories on these units, they are fascinating. Also have a 2005, 2005A, TP343A, 5015A and 4050 out in the waiting room, stay tuned.

*Hint for young players: if you clean or disassemble an analog meter, as your last reassembly step wipe the lens in and out with an anti-static laundry dryer sheet. Scent optional, I use "Country Fresh."