Repairing 6V on HP 6236B

[thone]

Hi all,

I am super new to repair thing and have joined/read eevblog for a couple weeks. I bought HP 6236B and turned out to be bad. 20V output is maxed at 24V and -20V output is maxed at 20V. 6V output stops increasing voltage after 1/4 way and voltage stuck at ~0.8V. Checking voltage reference, they are a bit off closed to 5% tolerance (e.g. 7.5V is measured 7.9V), otherwise looked ok. I tried to chase down the 6V circuit first. Checking TP6 as and it was about -0.187V, as per manual, I checked for defected Q7, Q15, Q8, R54. Don't know what I am doing, measuring Q8 & Q7 in circuit looks weird, so I removed them and check for ~0.7V between B-E / B-C and high impedance on E-C, so I guess they are ok.

So, I looked at the op-amps U3 of the voltage comparator. The inverting input is -0.016V and non-inverting input is -0.112. The output is -0.187V which seems to be problematic (I assumed that the circuit is trying to make inverting input and non-inverting input closed together). Since I think that Q8 & Q7 are good, I thought, then it must be U3 which is bad, so I replaced it with LM1558 (as in older post mentioned it could be the replacement). Now, interesting thing is, I got the range from ~1V to ~4.6V, but the pot nob is very sensitive and only work as small turn can saturate the output. The voltage between inputs are still -0.0163 vs -0.603 (at 4.6V).

Can anyone give me a good suggestion on how should I proceed troubling this power supply? It's kind of fun troubleshooting, but look like I need help unblocking this . (Was thinking to replace Q8/Q7 next to see if they are defected)

[bitseeker]

Welcome to the forum, thone. Troubleshooting starts at the +20V supply. Since your +20V output is high, check TP1 - TP3 as per Table 5-4 in the manual. See if that gets you on the right track.

[thone]

All references outputs are interesting. I am not sure it's ok or not.

TP1 7.5V +-5%. I measured 7.9 on my DM. My scope showed min/max/avg 7.71/8.36/8.07. -- Look a bit high
TP2 -6.2V +-5%. I measured -6.23 on my DM and -6.56/-6.07/-6.25
TP3 -12.4V +-5%. I measured -12.9V on DM and -12.5/-12.9/-12.7.

Will you consider these are not normal?

[bitseeker]

Will you consider these are not normal?

The math will answer that.

7.5V ±5% = 7.5V ±0.375V = 7.125 to 7.875V
-6.2V ±5% = -6.2V ±0.31V = -6.51 to -5.89V
-12.4V ±5% = -12.4 ±0.62V = -13.02 to -11.78V

Based on these ranges, TP1 looks to be too high, TP2 and TP3 are OK. So, for TP1 high (table 5-4, step 1, response b)  the recommendation is to check VR3 for open. If VR3 is OK, start going through Table 5-5, troubleshooting for +20V channel. The first symptom there is voltage too high. Follow the actions and probable causes there to see what to check next.

[Tac Eht Xilef]

Bought one a few weeks ago "for parts, untested" - all it needed was a new X cap - but it's still sitting on the bench with the lid off, so I just measured the TP voltages

(All measured with 6v set to 5V, +20v set to 11V, & tracking on)
TP1, 7.929
TP2, -6.514
TP3, -13.024
TP4, 0.045
TP5, -0.149
TP6, -0.098
TP7, 0.719
TP8, -1.229
TP9, 0.630
TP10, -0.800
TP11, -2.169
TP12, -2.169
TP13, -0.018
TP14, 0.000
TP15, -0.014

As you can see, some of the voltages are similarly out of spec - but it works perfectly & passes all the performance tests in the manual well within specs.

Definitely follow bitseeker's advice & work through the troubleshooting procedure in Ch 5 of the manual. It's very comprehensive.

(While waiting for mine to turn up I did look a bit at suitable substitute parts. The pass transistors are supposedly a bit problematic - they're basically special RCA-made 2N3055 / 2955 parts with an extended SOAR, but MJ15003/MJ15004 look to be suitable replacements. I also saw the LM1558 was supposedly a suitable replacement for the CA3458T, but haven't been able to find a datasheet for the RCA part to confirm. I suspect, however, that it's not terribly critical & any modern bipolar dual op-amp that fits the voltage ratings would suffice.)

edit: Looking a bit deeper, I'm wrong on that last bit. CA3458 -> Harris HA-2655 (datasheet), which is actually a pretty cool part for the era. I'm not convinced the LM1458/1558 is a great substitute - input impedance is orders of magnitude lower, input bias & offset currents are orders of magnitude higher, & I can find nothing about the GBW or slew rate. Studying the 6236B circuit I wouldn't think it'd make much difference, but ... I suspect there's better options, albeit in more modern packages. Thoughts & ideas?

[bitseeker]

Yeah, some voltages can be a bit off and still operate properly. You'll just compensate the meters to make up for it, so to speak. If it's too far off, then the trim pots might not be enough and the aged part will have to be replaced. This supply does sound like it has bad pass transistor(s), comparator(s) or divider(s).

Even when you think you know where the issue is, it's beneficial to go through the troubleshooting steps as it can minimize getting misdirected by red herrings.

[Tac Eht Xilef]

Yeah, I'm inclined to agree with your diagnosis (& suspect there may be multiple faults). I mainly posted the voltages on mine as a guide to what should probably be OK, since the manual doesn't give the voltages for a working unit & unless you follow the troubleshooting table you can easily be misled by a normal measurement.

e.g. if you jump around in the troubleshooting procedure, finding ~0.7V on TP7 may lead you to suspect CR6 - when in fact it seems to me to be perfectly reasonable value for a working unit, and is only really relevant as a fault indication if it suits the symptoms (i.e. if the o/p voltage is low).

In short: thone, follow The Procedure. Do not deviate from The Procedure. Do not try to second-guess The Procedure. The Procedure knows all, and will punish those who stray...

[bitseeker]

I mainly posted the voltages on mine as a guide to what should probably be OK, since the manual doesn't give the voltages for a working unit & unless you follow the troubleshooting table you can easily be misled by a normal measurement.

Yes, many thanks for posting that. It's always helpful to have a working unit for comparison. With that and the manual, hopefully thone will be able to locate the issue(s) in short order. It's a good first-repair project, especially since so much info is available and the parts are generally standard (except for the pass transistors, as you mentioned previously).

[thone]

Thank you both. I have been here and there, now I know that I need to strictly follow the procedure. I will start again and let you guys know.

[Tac Eht Xilef]

Well, I was half-joking about The Procedure. But the one in the manual is very good

Something else to check out that I don't think is in there: Since you said the +20V output is hard at +24V while the -20V is at -20V, what happens if you 'uncouple' the tracking i.e. adjust the Tracking Ratio knob on the top-right counter-clockwise from the switched position? If you can vary the -20V output with that, that will narrow things down a bit.

(As an exercise in practical circuit analysis for faultfinding, see where the voltage at TP1 - the one that was slightly out-of-spec - is used, and think about whether the extra 0.1V or so in those places will make much difference...)

And, I've got to ask: you didn't buy this on eBay a few weeks ago from a guy who had 2 for sale and who mistakenly sent you the one from the other auction, did you? Because, if so, I got "yours"

[bitseeker]

Thank you both. I have been here and there, now I know that I need to strictly follow the procedure. I will start again and let you guys know.

Sounds good. Looking forward to good news.

[thone]

Well, I was half-joking about The Procedure. But the one in the manual is very good

Something else to check out that I don't think is in there: Since you said the +20V output is hard at +24V while the -20V is at -20V, what happens if you 'uncouple' the tracking i.e. adjust the Tracking Ratio knob on the top-right counter-clockwise from the switched position? If you can vary the -20V output with that, that will narrow things down a bit.

(As an exercise in practical circuit analysis for faultfinding, see where the voltage at TP1 - the one that was slightly out-of-spec - is used, and think about whether the extra 0.1V or so in those places will make much difference...)

And, I've got to ask: you didn't buy this on eBay a few weeks ago from a guy who had 2 for sale and who mistakenly sent you the one from the other auction, did you? Because, if so, I got "yours"

Voltage tracking seems to be working fine. It was my bad saying that -20V max out at -20V. It's actually maxed out at -24V. I guess, it might not be in the switch position when I tested it first time.

Is that the guy from Arizona? The item is described as fulling working, so I ended up getting it for free and having fun fixing it .

BTW, just wondering, when I checked TP1, the voltage is a bit higher. If I want make sure that VR3 is not opened, is there way to test it without removing it from the circuit as it has 40 Ohm resister parallel to it? I couldn't find any problem following the procedure, but to be sure, I will probably start removing components out from circuit and test one by one.

I also found that my 6V is oscillating (I got 1V to 5V output). It's almost square wave. It seems to be coming from voltage op amp as checking output from current limit op amp, TP12, I got -2V constant. Doesn't seem to look good...

[PaulAm]

Have you put a scope on that?

I was fixing a 6236b that did not regulate and had a severe oscillation on the output.  Turned out to be the output filter capacitors.   Change those out before you change any op amps.

[thone]

Have you put a scope on that?

I was fixing a 6236b that did not regulate and had a severe oscillation on the output.  Turned out to be the output filter capacitors.   Change those out before you change any op amps.

Yep. It's the output cap. The oscillation happened after I replaced the op amps (will test old op amp to make sure).  Now, my scope has severe ripple. Manual pointed out that I may not have correct group loop setup as the power supply has GND cut out. I am picking new power plug and will update you guys.

[45Overload]

I disagree that troubleshooting begins with the +20 volt supply.  Everything in this supply depends upon the regulated voltage section at the top of the schematic.  This circuit is independent of the rest of the unit and has to be right for anything else to work.  When this thing was built, a 5% carbon composition resistor was still considered to be a precision part, so voltages will tend to have wider tolerances than we are used to today.  The regulator section depends upon temperature compensated zener diodes, so these will have to be checked in-circuit.  Getting this thing up and running should not be difficult.  Just have a good reason before tossing parts into it.

45O

[properharmony]

This is old, but for anyone else looking at fixing one of these, the opamps in there I was able to carefully get some tl072s jammed in the round footprint and revitalized the regulation just fine; beware, the square pad is not pin 1; the round pad with the line through it is pin 1. 

The reference IC well...pretty much have to use what's specified and pony up the 14 bucks for a new one of those but instead of dropping 45 bucks on the 3 goofy footprint opamps with LM1558s from Mouser or the like, just 3 bucks worth of tl072s worked fine.