HP 6034A Repair

[anewmanx]

Greetings,

My electronics background is somewhat limited. I have mostly worked with computer programming and micro-controllers such as PIC. I recently found myself in need of better equipment for collaboration with my father who works on vintage tube amplifiers, pedals, etc.

Long story short I needed a power supply. I found a good deal on this HP 6034A. Only problem is there is a hole blown in the power regulator board.

Here is the unit:




The damage to the regulator board can be seen here:


These are the two transistors from the board:


If could verify that this section of the manual is correct for what I"m looking at, I would appreciate it before I go ordering parts:


Any idea what could have caused that resistor that BLOW UP so dramatically on the previous owner? I hate to replace it just to blow again. I'm not sure what to test to find out the cause.

Is there anything else I should be checking before I replace the blown part and try to power this unit up? I have included below more pictures of the inside of the unit. The only photo that didn't turn out was of the transformer, but it looks physically pristine still.









Thank you for any and all advice you may have to offer. The unit has been powered down some time and all the capacitors are fully discharged, and I am observing safety precautions, ESD safety, etc.

[Ordinaryman1971]

I have one of those not working as well. In my case it powers up but it blinks all the lights and doesn't go through the post. If you need any pictures posted of the board you have burned up let me know I will have it for you.

[anewmanx]

Thank you. I will keep that in mind.

You say the LED's are flashing? Those are fault codes. In the service manual (which you can find online free) there is a table for fault codes on pages 5-10 through 5-12.

If you decide you want to sell yours cheap for parts let me know. ^____^

[Tomorokoshi]

Well...

Starting with page 159 of the 06034-90006 manual on the Keysight web site, which you found, look at the circuits around R3 and Q7.

Input to R3 is the bulk DC supply from the transformer, rectified by CR8 and CR9, and then made somewhat DC by C50 at 6300 uF. It appears the capacitor is rated for 20V, so the rail when running might be 15V or something. There should be an explanation somewhere in that document.

From the bulk supply Q7 is a series-pass regulator into C54 at 1000 uF. It looks like J2 is a jumper, so the load can be disconnected. This may be advisable for testing purposes.

Between Q7 and C54 are "Power On Circuit" and "1ms Timer" circuits. They don't have anything for the control of the regulator.

It's helpful to know what R3 was damaged by. At 0.33 ohms, rated at 3W, it is at full power when conducting 3A, or connected across a 1V supply. However, for it to burn like that would have taken somewhat more. If the bulk supply is actually 15V, then it would take a 5 ohm load looking into Q7 to achieve its maximum power rating. So we might be looking for something less than maybe 3 ohms.

Now let's look for current paths:

From Q7, there are several over-current options that would damage R3:
1. Through a short in C55, 1 uF ceramic.
2. Through a short in U73, 7805 5V regulator.
3. Through a short in CR12, diode.
4. Through the Power On Circuit.
5. Through the 1ms Timer circuit.
6. Through any of C2...C30.
7. Through the load on +5V.

So one approach is to keep Q7 out of the circuit for now, and disconnect the jumper at J2. This isolates some of the circuits.

Do you have an LCR meter? Does your DMM have a capacitance function?

To start with, measure:
1. Capacitance across C50.
2. Capacitance across C55.
3. Capacitance across C54.
4. Capacitance across C2.

[anewmanx]

Well...

Starting with page 159 of the 06034-90006 manual on the Keysight web site, which you found, look at the circuits around R3 and Q7.

Input to R3 is the bulk DC supply from the transformer, rectified by CR8 and CR9, and then made somewhat DC by C50 at 6300 uF. It appears the capacitor is rated for 20V, so the rail when running might be 15V or something. There should be an explanation somewhere in that document.

From the bulk supply Q7 is a series-pass regulator into C54 at 1000 uF. It looks like J2 is a jumper, so the load can be disconnected. This may be advisable for testing purposes.

Between Q7 and C54 are "Power On Circuit" and "1ms Timer" circuits. They don't have anything for the control of the regulator.

It's helpful to know what R3 was damaged by. At 0.33 ohms, rated at 3W, it is at full power when conducting 3A, or connected across a 1V supply. However, for it to burn like that would have taken somewhat more. If the bulk supply is actually 15V, then it would take a 5 ohm load looking into Q7 to achieve its maximum power rating. So we might be looking for something less than maybe 3 ohms.

Now let's look for current paths:

From Q7, there are several over-current options that would damage R3:
1. Through a short in C55, 1 uF ceramic.
2. Through a short in U73, 7805 5V regulator.
3. Through a short in CR12, diode.
4. Through the Power On Circuit.
5. Through the 1ms Timer circuit.
6. Through any of C2...C30.
7. Through the load on +5V.

So one approach is to keep Q7 out of the circuit for now, and disconnect the jumper at J2. This isolates some of the circuits.

Do you have an LCR meter? Does your DMM have a capacitance function?

To start with, measure:
1. Capacitance across C50.
2. Capacitance across C55.
3. Capacitance across C54.
4. Capacitance across C2.

Bless you, this is exactly what I needed. I'm going to strip down the case and get the board out so I can test all of the capacitors and make sure there is no obvious damage on the other side. I can't reach the leads as it is since they are slam down on the board.

Now to buy an ESR meter!

[anewmanx]

Well, I have run through the capacitors you mentioned with an ESR meter and so far all of them show appropriate capacitance and ESR measurement. Quite good measurement considering how old they are! I went ahead and checked the rest of the capacitors I could find on the board and failed to find any problems. Same with the diodes, they all checked good. I can't find any obvious problems. Two of the wires going to the voltage regulator board were a little weak and snapped off so I resoldered those.

I went back to the 2N3791 T03 transistor that was on the board with the blown resistor and ran it through with my DMM again. Everything appears to check out, with one oddity. When I trigger it it goes form infinite resistance to 1 megaohm of resistance instead of it reading as an open. I put it in the transistor checker and it didn't show any leakage. I'm at a loss as to what caused that resistor to blow up.

Any other tests I can run in the meantime? I have a function generator arriving tomorrow and I have a Tektronics 5111A and a cheap digital scope on hand if there are any tests I could run with those that would provide insight before putting parts back in and turning it back on, (and praying it doesn't blow up again.)

[anewmanx]

I have one of those not working as well. In my case it powers up but it blinks all the lights and doesn't go through the post. If you need any pictures posted of the board you have burned up let me know I will have it for you.

Actually, yes, a photo of how the board would be very appreciated. At least one of it mounted up properly. There is enough extra wire that one would flip it around and mount it, so I would hate to put it in wrong.

Any luck on repairing yours?

[anewmanx]

From Q7, there are several over-current options that would damage R3:
1. Through a short in C55, 1 uF ceramic.
2. Through a short in U73, 7805 5V regulator.
3. Through a short in CR12, diode.
4. Through the Power On Circuit.
5. Through the 1ms Timer circuit.
6. Through any of C2...C30.
7. Through the load on +5V.

CR12 is reading a dead short both directions, but reads fine out of circuit.

Should I replace r3 with a flameproof wirewound resistor?

Can the 7805 be tested in circuit with all power off? If so, output and ground pins read 0 with the dmm set to 2000k ohms.

[Tomorokoshi]

I get between 4k and 6k with a couple different date codes for a 7805 between pins 2 and 3. Same for both polarities. Out of circuit.

With CR12 okay then U73 may be shorted. Likely internally, although it's possible for external shorts to happen.

A short in U74 or C54 is also a possibility. It looks like U73 is in one of those spidery heat sinks. Make sure those aren't shorting together. At this point removing U73 and C54 are probably the next test.

[anewmanx]

I get between 4k and 6k with a couple different date codes for a 7805 between pins 2 and 3. Same for both polarities. Out of circuit.

With CR12 okay then U73 may be shorted. Likely internally, although it's possible for external shorts to happen.

A short in U74 or C54 is also a possibility. It looks like U73 is in one of those spidery heat sinks. Make sure those aren't shorting together. At this point removing U73 and C54 are probably the next test.

Thank you so much for taking the time to answer my questions Tomorokoshi, and for pointing me in the right direction. Tomorrow I will de-solder those parts and test them out of circuit. Are you an electrical engineer, or a very well versed hobbyist?

[anewmanx]

I get between 4k and 6k with a couple different date codes for a 7805 between pins 2 and 3. Same for both polarities. Out of circuit.

With CR12 okay then U73 may be shorted. Likely internally, although it's possible for external shorts to happen.

A short in U74 or C54 is also a possibility. It looks like U73 is in one of those spidery heat sinks. Make sure those aren't shorting together. At this point removing U73 and C54 are probably the next test.

You were so right!!

I removed and tested u73 and it was fine. Sadly I lifted a trace a little, but soldered back to it so I hope no permanent damage.

When I tested c54 it gave a very weird reading with the esr meter. I pulled it out of circuit and the short on the 5v rail disappeared. I tested it. Totally blown. Unreadable capacitance and resistance shows dead short from leg to leg. Gets very warm with just a meter on it!

I believe I will order parts now and hope nothing else’s is broken.

Any further advice while I wait on parts? I am so very grateful for your help.

[Tomorokoshi]

Glad it's working out. I apply my engineering experience into my hobby, and I use my hobby as continuing education for my engineering. To support that I get various equipment as needed, and because a lot is from ebay I'd say that perhaps 50% has needed repair of one kind or another. I've had to dig into HP linear power supply circuits a fair amount now, both as benchtop power supplies and as the supplies within units like signal generators.

By chance what is the manufacturer of the failed C54 capacitor?

If we analyze the failure, we can see if the observed and measured symptoms match with the behavior of the components in the circuit:

1. C54 fails short.
2. The regulator circuit attempts to maintain 5V.
3. Q7 gets turned on all the way.
4. Enough current flows through R3 to burn it out.

However, what if something happened before the C54 failure? Perhaps a fault in the U73 / Q7 circuits?

Alternatively, could Q7 have been damaged during the failure mode?

When you get it back together you will want to test it with J2 disconnected. Verify +5V output.

It's possible to test U73 and Q7, but as you already found out, removing parts from circuit boards carries a risk of damage to the board. Lifted tracks can be fixed with soldering a length of single-strand wire-wrap wire. Typically those "blue" wires you might find on boards to fix problems or errors.

Ironically, if you had another benchtop supply with good voltage control and current limiting, you could test the +5 circuit by connecting the other supply to C50. Turn the voltage up slowly, and verify that it regulates to 5V once C50 reaches maybe 7V.

[anewmanx]

Glad it's working out. I apply my engineering experience into my hobby, and I use my hobby as continuing education for my engineering. To support that I get various equipment as needed, and because a lot is from ebay I'd say that perhaps 50% has needed repair of one kind or another. I've had to dig into HP linear power supply circuits a fair amount now, both as benchtop power supplies and as the supplies within units like signal generators.

By chance what is the manufacturer of the failed C54 capacitor?

If we analyze the failure, we can see if the observed and measured symptoms match with the behavior of the components in the circuit:

1. C54 fails short.
2. The regulator circuit attempts to maintain 5V.
3. Q7 gets turned on all the way.
4. Enough current flows through R3 to burn it out.

However, what if something happened before the C54 failure? Perhaps a fault in the U73 / Q7 circuits?

Alternatively, could Q7 have been damaged during the failure mode?

When you get it back together you will want to test it with J2 disconnected. Verify +5V output.

It's possible to test U73 and Q7, but as you already found out, removing parts from circuit boards carries a risk of damage to the board. Lifted tracks can be fixed with soldering a length of single-strand wire-wrap wire. Typically those "blue" wires you might find on boards to fix problems or errors.

Ironically, if you had another benchtop supply with good voltage control and current limiting, you could test the +5 circuit by connecting the other supply to C50. Turn the voltage up slowly, and verify that it regulates to 5V once C50 reaches maybe 7V.

Very interesting! I am so glad you were kind enough to help me. I know my way around programming and microcontrollers, but my knowledge of analog electronics is more limited. I understand the basic concepts, but prefer to think in logic patterns.

With U73 removed I was able to test it for functionality with a 9v battery and it produces 5v with no apparent short conditions of any kind.

My father shipped me an old healthkit transistor tester and Q7 so far reads out as having no leakage of any kind. On the DMM when I use a 9v battery as the current source and trigger the transistor the reading goes from the appropriate voltage down to nearly 0 on the base; so I suppose its a normally open transistor. It appears to be functioning correctly, and to have survived the failure. It tests out correctly with a DMM in the various diode mode tests, the only oddity is the resistance I read when it opens. Then again, since it is actually closing when it opens I suppose 1 megaohm is appropriate.

[anewmanx]

...

Greetings Tomorokoshi, we have made progress!

I replaced the parts that we suspected were faulty and the unit now powers up, but controls are disabled and voltage and current are both bottomed out despite the overvoltage error indicator not lighting up.

I entered self test mode and received the following two errors:

Interrupt test failure #5

HP-IB interrupt occurred when it should not. Possible fault areas: Fault is probably located in the interrupt circuit or the HP-IB circuits.

HP-IB Interface test failure #1

M2 = 1 CV = 0, but M2 ANDed with CV=1

Possible fault areas: Although the signals could be lost in the microcomputer circuit or front panel gates, latches, or connectors, the most likely cause is damage to the power mesh.

I removed the three J5 jumpers and poked around with a multimeter. The previously non-functional PNP tranistor circuit (Q7) now appears to be working. It reads roughly 15v on two of the legs and 5v on the other. All of the regulators with the spidery heat sinks appear to be working as well. The NPN transistor on the regulator board (Q6) has no power going to it at all on any of the three legs, and I had verified the transistor is good.

Any ideas on where else I should check? You are more familiar with these devices than I am. So far, I am very pleased I was able to bring it back to life, even if there is still something damaged inside.

Update: I found a wire disconnected from the board. One of the wires going to Q6 the NPN transistor.

Self test no longer shows any error codes, and the device still refuses to allow me to adjust voltage. The NPN transistor will briefly show a reading on the different legs then it disappears down to 0 volts on each leg. Very odd.

Update 2: I'm at the step on the far right of the chart:

Check CV reference circuit and dac. Use S.A. (Fig 5-15) to check first and second rank storage.

[anewmanx]

New update: While poking around with a multi meter before testing the control board, I did notice an oddity.

If I disconnect only the -15V J2 jumper the device will power up and show voltage on the main screen, but still has no ability to adjust voltage. The second I put the jumper back in for -15V the device goes down to no voltage. With the jumper disconnected, there is -15v on the source side of the jumper and +0.86 volts on the other side.

What I don't understand is this device is acting like its in overvoltage protection mode, but the overprotection mode LED is not lighting up. It flashes with all the others when the system starts up so I know its not burnt out.

[bsdphk]

When they put the TMS9981 into the HP6034 they clearly struggled to keep the price down.

If you study Firgure 4-3 and 4-4, you will notice that there are no A/D  converters, only D/A converters, which means the CPU is not very well aware of what is actually going on.

There are some workaround solutions, for instance the CPU measures output current by ramping the CC DAC down from the set value until output stage switches to constant current, then quickly set it back to what you asked for.

That kind of tricks introduce some really strange failure modes, where defect parts can give some very strange symptoms you would not expect.

So for the 6034 it is very important to stick slavishly to the trouble-shooting diagram and not skip any steps "because it couldn't possibly be that".

Good Luck!

[anewmanx]

Thank you for the information.

My fear is I will have to proceed to the signal analysis phase. That is something I Have never done. I suppose there is a first time for every new skill.

[Tomorokoshi]

Let's look at -15V then.

First, examine the schematic to determine where -15V is used. This comes from A4 U60, which is an LM320T-15. It looks to be good for around 1.5A. I probably missed some:

Page 165, A4, U10, CC DAC
Page 165, A4, U9, CV DAC
Page 166, A3, R82, CC FULL SCALE
Page 166, A3, R81, CV FULL SCALE
Page 167, A3, R96, CC ZERO
Page 167, A3, U10, CC PROGRAMMING VOLTAGE
Page 167, A3, R155, Q7 follower
Page 167, A3, R90, REMOTE CURRENT METER ZERO
Page 167, A3, U11, CV DAC OUTPUT
Page 167, A3, R7?, CV ZERO
Page 167, A3, U7, CV PROGRAMMING VOLTAGE
Page 167, A3, U8, CV DAC OUTPUT
Page 167, A3, R38, U4 input bias
Page 167, A3, U4, OVP
Page 167, A3, U16, LOCAL OVP

So it looks like there are a bunch of op-amps with several control potentiometers and a few adjustment trim pots. Losing -15V makes sense that it would appear to be in OVP mode and would make the OVP indicator malfunction.

With the -15V J2 jumper disconnected, what do you get for resistance and capacitance measurements on the -15V line past J2?

[anewmanx]

Thank you again for your guidance. I will perform this test procedure tomorrow.

I also wanted to mention that when I enter self test mode, all tests continually pass, BUT -- when I put it in LED test mode, moving the RPG does NOT cycle through the LED's. So that circuit is malfunctioning, or the RPG is dead.

[anewmanx]

With the -15V J2 jumper disconnected, what do you get for resistance and capacitance measurements on the -15V line past J2?

Hmm... between what two points? So many points.

Between the -15v line pin on j2 and the -15V test strap on the front panel I get resistance of 770 ohms and the capacitance meter reads 1.2 on the 20uf scale. The cheap ESR meter I have gets a totally different reading though, oddly, it reads R 3.5 and 7.3uf. usually it agrees quite nicely with the capacitance meter I have, so I'm scratching my head on that one. We have gotten deeper into this than my knowledge set was equipped to handle. Time to learn!

- Update:

I made a mistake and need to replace four resistors on the A1 board now before proceeding any further. R14, R21, & R67, R68. They were off of Q3 & Q4. I suppose they were there to protect the device from the hazardous condition I inadvertently created. They are even raised up off the PCB with nice little isolators so they can explode without burning anything... nice design choice.

[anewmanx]

I've been kicking myself all night because I thought I did something wrong to cause those four resistors to explode, but I am starting to think that it was not my fault.

I found a dead diode on the circuit... I'm starting to suspect it was failing all along and causing irregularities in the voltages. Diode CR11 failed. It's wide open both directions now. Reads completely different from CR12 both in and out of circuit. Perhaps something further along caused that diode to pull too much current or there was a voltage spike. I don't know yet.

While I'm waiting on parts, does anyone know the pinout for the HP 5060-0329 Rotary Encoder? I've got it out and I'd like to test it to see if it is generating pulses. The pin header is five pin, but only uses four of the spots. The four wires are black, white/black, white/orange, empty pin, and white/red. in that order.

Update:

The rotary encoder uses a 3mm incandescent bulb (from what I read online)as a light source for two sensors inside. Mine is burnt out. I'm going to attempt to replace it with a 3mm white LED and a 200 ohm series resistor.

[anewmanx]

Well, more bad news.

Again the fusing (I'm assuming since they are 1/8th) resistors blew, but it appears the new diodes were fine.

They blew in an interesting way though, the 51 ohm blew first followed by the 62 ohm.

Both the little 1/8th resistors for Q3 & Q4 blew. Both transistors appeared to be okay when tested with a multimeter, same with the FETS Q1 & Q2. The leg of R21 that connects through to R20 is visibly burnt. The same condition is not present in R14 where it connects to R13. The resistor pair of R21 & R68 is visibly more damaged than the resistor pair of R14 & R67.

I'm at a loss. I can't very well throw more resistors in just for them to blow up. What would my next step be?

This is the schematic for this section:



Update:

Q4 is blown. Q3 appears to function with a heathkit transistor checker.

Should I perform some kind of test on T4 before replacing Q4 and all blown resistors?

[coromonadalix]

I would do a total recheck, or find a way to isolate some component pin(s)

remove all q1 q2 q3 q4, res r11 r12 r13 14 15 16 18 19 20 21 22 23 67 68
 and c8 c9,   cr5 cr6 cr7 cr8 cr11 cr12

by chance this output stage is xformer coupled,  maybe  just maybe  try to check t1 and wth an lcr meter, t2 is the feedback signal to the regulator section.

With t3 t4  pins 3-4 5-6  open you could check if you have some drive signals ?

This psu seems to work like a buck boost converter,  at c16 c17 positive,  do you have a short to ground ??

[anewmanx]

Thank you for the advice. I will have to wait until the new multimeter arrives to proceed. In the meantime, Q11 has ended up being fine. Q3 & Q4 were both blown.

I'm worried now that T4 might be bad. If it doesn't read any shorts in the windings when the multimeter arrives, then I suppose I can test T3 with my function generator to determine what normal behavior is for T4 and see if it is malfunctioning. It's part number 5080-2018. I'm not sure where or how I would replace that, but its looking like that would be a retrofit situation. I can't find the part online.

I suppose I will check T1 as you suggested while I am at it. You might have the answer there. I don't have an LCR meter though (I suppose I could use a triangle wave form from the function generator and view it with the oscilloscope?). Both FETs tested out fine when I checked them so I don't believe them to be the problem. All of the capacitors appear to be functioning correctly based on the capacitor meter and ESR meter readings I've taken.

Update:

Transformers T1, T2, T3, & T4 all appear to be working correctly. I passed both sine and triangular waves through them and the devices properly stepped down their voltage and did so stably even when I adjusted the dc offset on the function generator. That was a fun new skill to learn! There was a bit of distortion in what passed through T3 from pin 2, but I believe it may have something to do with the presence VR3. Not that I would know, I have a lot to learn still.

CR11 & CR12, which were both new, both are now shorted open. So it appears each complete system (CR11, Q4, R68, R21) & (CR12, Q3, R67, R14) fried crispy last time I turned it on. Both Q3 & Q4 were already shorted when I turned it on and I didn't know. I suppose a shorted transistor could cause that kind of damage and draw enough current to punch through CR11 & CR12 and fry the resistors on the way, am I correct?

[anewmanx]

Update:

I have verified correct ON pulses at CR11 & CR12 from T3, and all transformers appear to be working correctly. The distortion I previously noticed on two of the transformers was due to the zener diodes in those circuits and is apparently normal.

I also found that the jumper for the OVP circuit had become disconnected while I was working on things which leads me to believe that was the cause of this circuit blowing. Well, another problem somewhere else that the OVP circuit would have shut down anyway is likely the problem.

Currently my plan is to replace Q3, Q4, CR11, CR12, and their associated resistors and turn it back on... praying nothing blows again. I'm not sure what else I could test as every other resistor, capacitor, diode, and transformer associated with the circuit in question is testing out perfect.

As one last precaution I may turn it on without replacing the transistors and probe around a bit for any errant voltage that shouldn't be there. Perhaps I can find a problem without sacrificing more parts. This is turning into a fun learning experience / repair.

Update:

With CR11, CR12, Q3, & Q4 out of circuit (along with R14, R21, R67, & R68 I do find something with my multimeter. Touching my probe to the leg of R21 that is connected to the MOSFETS and their associated circuitry (to include T1,) briefly shows a large negative voltage that quickly dissipates to nothing. The same thing happens at R14, but in that case a positive voltage.

Does that seem like normal behavior for the half of the circuit that still remains? Perhaps a capacitor somewhere in the circuit discharging?

-- further update:

Removed C17, C16, & C18 all fine so far. No dead short that I can find after further testing. Still hunting.