Peculiar behavior HP 6236A Power Supply [SOLVED]

[enut11]

I was testing my HP 6263A for noise and its suitability as a voltage reference power source when I discovered this peculiar behaviour.
The power supply had been on about 5 min before testing. Load current was 38mA at 15v. The output dropped by 20mV over the course of 100 minutes. 

More to worry was the sudden drop in voltage:
Dip #1: -1mV
Then a recovery shortly after...
Dip #2: -5mV
Dip #3: -1.5mV

Perhaps a faulty 6.2v zener?

[enut11]

During the next 20 minutes the output was steady then suddenly spiked up by about 28mV before slowly coming back down again.

[enut11]

A bit later, the supply seems to have settled down.
Still, it would be nice to know why it behaved strangely for the first hour or so...

[bob91343]

It's just noise and drift.  Perhaps a bad filter capacitor or other component.

[enut11]

Thanks @bob91343. I just ran a similar test on my HP 6201B power supply and it did not show any of these jumps.

[sean0118]

I wonder if it's caused by temperature variation in the trimpots? If they are wirewound could that cause jumps like that?

DeOxit might help, although that is my solution to most problems. 

[RaymondMack]

RFI can be rectified by the electronics and appear to be mV shifts. Otherwise it might be a noisy zener diode.

[enut11]

I tested VR1, the -6.2v zener in the power supply. From cold start, there was no abnormal behaviour in zener output over the first hour of operation.

So, next is to check out the old pots...

[TimFox]

Sometimes the single-turn wire wound trimpots, left in place for years, corrode a bit.  Rotation for a few turns is often good enough to restore operation, after re-adjustment.

[enut11]

Thanks for all the feedback.

I retested the VR1 zener and all is good.

I have decided to replace the Vpot on the 20v supply.

[coppercone2]

you can open these pots, put them in a ultrasonic with soap, soak with alcohol, stimulate it with a mini nylon brush and lube it up and put it back together to maintain your budget on old HP, the appeal of them is how restoration friendly they are

these look a little newer though, if they are sealed pots you need to replace them. when they are really clean and fresh they feel kind of zippery

[WattsThat]

That unit has been modified. The 6236A/B does not have constant current capability. Sure, it has an HP knob but the rub on lettering typeface gives it away.

Somebody has been mucking about in the voltage regulator to implement CC and I’ll bet the problem is related to the modification. Start with the obvious as randomly replacing parts is never a good idea.

[enut11]

Yes, CC has been transferred from the motherboard (fixed trimpots) to the front panel and made variable. In all other respects it behaves in the intended way.
In this test, CC has not been activated so not likely to cause the sudden jumps in output voltage.

I have just replaced the single turn 10K voltage pot for a Bourns 10-turn. Will test and report on outcome.

[enut11]

Tested for 1 Hr from cold start, the new voltage pot appears to have solved the jumpy output problem. The mV steady rise is within specification and eventually flattens out.

I acknowledge that cleaning the original WW pot would probably have achieved the same stability but the helipot gives me improved output setting resolution.

[KE5FX]

I like these old-school HP/Harrison supplies, and I use several of them regularly, but the wirewound pots scare me.  The way they're designed, intermittent wiper contact doesn't have the effect of steering the voltage towards zero, but towards max.  And you always get intermittent wiper contact with these things, eventually.

I've thought about emulating the pots with an optical encoder tied to an EEPROM-based micro.  That'd eliminate the problem completely, and also eliminate the need for separate fine/coarse adjustments since the control response could be made rate-sensitive.  It's always annoying when the fine adjustment runs out of range just before reaching the level you want.

[bd139]

Almost always the pots on them (or if it has them the bloody rotary or end switches which are cancer). I had a 6236b with a faulty reference zener as well - that was jumpy then failed entirely causing some rather worryingly disparate voltages to appear on the output terminals!

[coppercone2]

deoxit L260-N1 goes into potentiometers for me.

I don't know what their normally packed with, but when you redo it yourself you feel the 'zipper' ringing of the coil against the wiper, the old ones feel like stuffed, and inside there is usually nasty black stuff similar to boogers... their usually at least 50 years old so its really a necessary step. Good to learn how to do because all HP stuff has the same exact aging method.

The only hard part is bending the fold, you need to bend it back a little with a knife but then don't keep using the knife because it can slip and the control is bad, then you switch to something like wide taper duck bill pliers that have a sharp point (or flat wide needle nose) and bend it really slowly and double check your alignment before you fold it over because it can break from bending it too much.

I usually do 1 day to disassemble and clean then the next day to reassemble so there are no expensive mistakes made

each time you save on a expensive potentiometer you can throw in a new set of cables with your new power supply for the same price

They are also great because you get nice screw terminals in the back and if you really are doing something complicated you can use them to interface with your electronics to act as a power stage.. this can save a shit load of time and money, I made a standard kit that has a whole bunch of fork terminals on different wires so I can quickly do whatever I need to do with them.. thats another good step too to polish and grease up that back screw panel so its ready to go. with modern supplies this easy to access and interface terminal block with high precision is
1) considered an extra feature
2) requires a custom DSUB cable

I feel like no one is gonna use that DSUB shit unless they are building a test bench for a company lol.. those HP power products can be used like gigantic 15 pound IC's with the most basic tools (home depot will do), high reliability hookup, etc.   

[enut11]

Just disassembled the 10K wire wound voltage control pot from my HP6236A power supply.

Photos show a dry scaly deposit so if it was originally greased there is no sign of it now.

[coppercone2]

ultrasonic it with simple green and rinse with alcohol and it will be good. I don't know how resistant the plastic is to alcohol, I would not advise a long alcohol dip. I think the simple green + ultrasonic has a slight deoxidizing effect too, more so then simple scrubbing, for whatever reason.

I usually get black snot, that one looks pretty decent actually, as far as bad pots I have seen

These work well on the coils without hurting them, after its been loosened up by ultrasonic I poke it with the 3rd brush in the picture from the left and then put it back there, run it under running water for a while, put it back in there, brush again, rinse good, squirt with alcohol so it pours out of the pot, shake it 'dry', let it dry the rest of the way and then deoxit in solvent then deoxit grease. I figure just putting thick grease in there will not entirely coat the thing. When you put grease wipe it around with another brush and smear it on the metal parts in there too.
https://www.amazon.com/Plastic-Handle-Keyboard-Cleaning-Black/dp/B076HP2TYB

I often wonder if a dip in HCl might help but I never tried. I really like putting as little money as possible into old equipment because you never know what the hell is gonna break and it can turn into a money pit if you are not careful (i.e. transformer).

You might also want to desolder the transistors, polish them up and reseat them with grease and new pads in those. I found the insulator sheet to be covered in like a oily liquid before, I think capacitor leakage got up in there. I rinse the whole PCB after it sits for simple green for 5 min then blow dry with an air gun because their usually disgusting and change color. That top side vent leaves the harrisons filthy. If you read my thread I put one with pots and everything into the ultrasonic after removing the panel meter and it was fine lol (thats not the one with the cracked trace, I was delicate with that one actually).. if you have those switches that are semi sealed, its easier to ultrasonic it panel down then take the switches apart with solder wires still attached and rinse with alcohol in a bottle then to desolder like 20 wires. One was so bad it was making smoke (tobbaco?) when I tried to desolder parts!

if you have a buzz from it don't blame the transformer immediately, it is part of the problem but the ceramic caps near the diodes are the biggest offenders, if your PSU has them.

[enut11]

Thanks @coppercone2. Sounds like you have done this a few times and you have all the right equipment. Sadly, no ultrasonic equipment here so all cleaning has to be done manually.

[sean0118]

deoxit L260-N1 goes into potentiometers for me.

That's interesting, I didn't know DeOxit made greases as well. Here in Australia there's only a limited range, Jaycar has the red and gold contact cleaners, but all the others have to be ordered from overseas with Digikey or similar.

Might be of interest to others here, The Signal Path did a video using DeOxit on a decade resistance box awhile ago (although this was on switch contacts, not trimpots):