HP 6214A Power Supply Teardown and Repair (Help!)

[TimNJ]

Hello everyone,

I bought this power supply off eBay for $30. I think it's around 40 years old. From initial testing, it seems to work. The only problem I can see is that its maximum output voltage is 9.2V, where the specified is 10V. A minor concern is the brittleness of the enclosure and other plastic tidbits. I'm keeping my fingers crossed and hoping I don't snap anything. Pots are also pretty scratchy and not-fluid though they seem to work fine.

The HP manual is brilliant (except for the illegible schematic grrr) and it shows reasons for why an output voltage may be low. At first I thought perhaps the caps weren't doing their job and excess ripple was causing a drop. But I'm not so sure about that now. Haven't done any load testing on it yet.

But anyway on to the pictures.

[TimNJ]

Okay so here's a few of my issues. I'm not exactly sure what to make of them from first glance.

The first picture is the output with no load attached. As you can see, there is a roughly 120Hz spike even with no load.

The second picture is with the supply set at 2V, and I believe it is in CV mode. If you look closely, the spike from before appears on this new waveform.

The third picture is with the supply set at 3-ish something volts CV, greater ripple.

The fourth picture is a zoom in on the wave, just for kicks.


120Hz sounds like the output of a bridge rectifier. and the waveform looks vaguely like RC charging/discharging. Perhaps the caps are indeed bad?  But this doesn't explain why the open circuit output voltage is roughly 0.8V too low. And the waveform is kinda messy

I'm also confused with some of the diode readings I'm getting. I'm supposing that it's because they are in circuit that I am getting funky readings? I measured the Vf to be 0.65V and then I'd swap the leads and I'd get a reading of 0.51V or 1.03V or something along those lines. It's not like they are zener diodes. They're rectifiers. What's wrong with my testing methodology?

This is the manual for reference if you have a minute to spare: http://cp.literature.agilent.com/litweb/pdf/06214-90002.pdf

So any ideas?

Thanks alot.

[krish2487]

Err...

Just stating the obvious...

I am by no means familiar with this particular power supply, but at a glance from the pics posted and the manual linked to.

The manual shows the schematic using discrete components, the pics you posted show one IC.

While the model number in the pics and the manual are consistent.

And the picture in the manual is too dark to discern whether it is an op amp or discrete matched transistors.

[edpalmer42]

If your scrope is dual trace, look at the voltage across the main cap with one channel and the output voltage with the second channel.  I think you'll see the same waveform.  I once came across an old power supply where the main capacitor was bad enough that it was actually getting warm from internal dissipation.  This was in a linear supply.  Something like that would explain both the waveform and the lower voltage.  Regardless of what's causing it, it looks like the voltage on the main capacitor is being pulled down far enough that the regulator doesn't have enough voltage to operate and the ripple is appearing on the output.

According to the manual, you should see a DC voltage of ~22V across the main cap.  What do you read? 

With no load on the output, does the ammeter show any current flow?  If so, it could be the output capacitor that's bad.

Old capacitors can fail when you power them up if they've been sitting idle for a long time.  Google on 'capacitor reforming' for an explanation.

Don't worry about your diode measurements.  Your readings are being messed up by the low DC resistance of the transformer secondary that's connected to the bridge.

Ed

[amspire]

The only problem I can see is that its maximum output voltage is 9.2V, where the specified is 10V.

That should be a simple calibration adjustment.

Pots are also pretty scratchy and not-fluid though they seem to work fine.

I think they are wirewound pots. You can feel the wiper move over the wires as you turn the pot, but they should last a long time.

The HP manual is brilliant (except for the illegible schematic grrr) and it shows reasons for why an output voltage may be low.

Did you get the 6213A manual I mentioned in the previous thread? It is much better quality and fairly similar to the 6214A. The quality of the 6214A pdf is a shocker.

You cannot read the writing on the schematic.

[amspire]

The manual shows the schematic using discrete components, the pics you posted show one IC.

That thing that looks like an IC is a resistor network.

[amspire]

Concerning the ripple, it may be that C5 has gone dry putting ripple on the regulator supply.

[TimNJ]

The only problem I can see is that its maximum output voltage is 9.2V, where the specified is 10V.

That should be a simple calibration adjustment.

Pots are also pretty scratchy and not-fluid though they seem to work fine.

I think they are wirewound pots. You can feel the wiper move over the wires as you turn the pot, but they should last a long time.

The HP manual is brilliant (except for the illegible schematic grrr) and it shows reasons for why an output voltage may be low.

Did you get the 6213A manual I mentioned in the previous thread? It is much better quality and fairly similar to the 6214A. The quality of the 6214A pdf is a shocker.

You cannot read the writing on the schematic.

Yes I did see it. Just found a very nice scan of the 6214A schematic. It is attached.

So I did a bit of tracing and I was able to figure out where the test points they kept referring to were.

There were 6 basic voltage tests. It failed 3, passed 3. Some "probable causes" it lists for the ones it failed include two electrolytic caps, some rectifiers, two zener diodes, and one transistor (Q11).

---

With regard to a low output voltage, there is only one trimmer inside, and I believe it is for adjusting the meter.

[TimNJ]

If your scrope is dual trace, look at the voltage across the main cap with one channel and the output voltage with the second channel.  I think you'll see the same waveform.  I once came across an old power supply where the main capacitor was bad enough that it was actually getting warm from internal dissipation.  This was in a linear supply.  Something like that would explain both the waveform and the lower voltage.  Regardless of what's causing it, it looks like the voltage on the main capacitor is being pulled down far enough that the regulator doesn't have enough voltage to operate and the ripple is appearing on the output.

According to the manual, you should see a DC voltage of ~22V across the main cap.  What do you read? 

With no load on the output, does the ammeter show any current flow?  If so, it could be the output capacitor that's bad.

Old capacitors can fail when you power them up if they've been sitting idle for a long time.  Google on 'capacitor reforming' for an explanation.

Don't worry about your diode measurements.  Your readings are being messed up by the low DC resistance of the transformer secondary that's connected to the bridge.

Ed

Thanks,

By the main cap, I presume you mean C9. The schematic says there should be 19V across it. I read 14.33V.

Where am I measuring current? On the output?

Thanks for your help.

[TimNJ]

And one last thing, what's with the "changes" at the end of the manual. I don't understand..Am I supposed to physically change my supply, or is it just a changelog with different revisions that they made.

[amspire]

The changes in the appendix are changes and fixes done to the design after the manual was published. There can be a variety of reasons - to fix problems, to make manufacturing easier, etc. Often you will find you have many, if not all the updates already. You can implement the changes, but if you can get the supply working well, you may not need to.

I think your problem is probably capacitors, and as I said, I suspect the regulator supply.  First disconnect any current oscilloscope connections from the supply, and then connect the + output to the scope ground. Check the voltage and ripple on 7, 8, 9 and 11 in the schematic.  You can vary the output voltage to see if it changes anything.

I mentioned C5 may be bad. Add C2 as well.

If the main capacitor C9 was bad, there would not be more ripple at 3V out then at maximum out.

Faulty transistors, zeners, etc are more like to stop the supply regulating at all.

[SeanB]

Basically all you need to change are C9, C5, C14 and C2, using the next higher value and a higher voltage range than the existing. The new ones will fit, though you may have to add sleeving and bend the leads slightly to fit in the existing holes. Use 105C good quality ones and it will give another 40 years of trouble free operation.

[amspire]

Basically all you need to change are C9, C5, C14 and C2 ...

Do my eyes deceive me or are there two C2 capacitors? I think the C2 on the output is meant to be C12, which is not actually used according to the parts list..

So when we say C2, take it as the one from the Q11 collector to +S

[edpalmer42]

By the main cap, I presume you mean C9. The schematic says there should be 19V across it. I read 14.33V.

Where am I measuring current? On the output?

Thanks for your help.

Yes, I meant C9.  The block drawing on page 4-1 of the manual you linked to shows 22V across C9 for the 6214A.  Maybe they changed it in later versions.  Either voltage will be fine.  In any case, 14V3 is too low.

I'm not talking about using an external ammeter to measure current.  With no load on the output, the ammeter in the power supply shouldn't show any current.  However, if C14, C2, or CR14 were going bad, the ammeter might show a current flow.  Actually, if they were pulling the C9 voltage down to 14V3 the ammeter will definitely show it and the regulator transistor should be getting very toasty!

Basically you have to start working through the troubleshooting procedures in the manual.  They look pretty decent.  When they give you a list of possible components, you have to figure out how to test that component.  Hopefully, you can test it in-circuit.  Sometimes you have to unsolder a lead to follow their procedures or test a component.  That's a nuisance, but that's life.

Experience tells us that in an old piece of equipment, the electrolytic capacitors are often the first things to go.  You've got C1, C2, C5, C9, C14 - not sure if I missed any.  Some people would just change them all before spending any time diagnosing the unit.  Even if they haven't failed, they're probably not meeting their original specs.  Or they might fail tomorrow with possibly spectacular results.  If the troubleshooting procedures point to one of them as possibly defective, you really want to remove it and test it.

Changes in the equipment could be made to:
- reduce cost
- make it easier to manufacture
- replace unavailable parts
- fix bugs
- add features or improve performance
- improve safety

Most of the time, if a change is necessary or recommended the manual will say so.  Otherwise, it's just to document the change so that the manual can be used with instruments of different ages.  Actually, I see that one of the changes listed is to change the voltage across C9 from 22 down to 19.  So you look at your serial number, read the change sheets, and figure out which value applies to the unit in front of you.

By the way, I noticed in one of your pictures that the 110 VAC leads for the fuse holder and the transformer primary are uninsulated.  Unplug the supply and cover those, and any other bare 110 VAC leads (power switch maybe?) with electrical tape for safety's sake.

Ed

[SeanB]

Change the 3 big capacitors and it will work well. You probably will go with 470uF 100V for C5, 100uF 35V for C2, 3300uF 63V or 4700uF 63V for C9, 10uF 63v for C1 and 100uF 63V for C14. All common values and readily available, plus will fit in with only a little bit of lead forming. If forming leads place a short piece of sleeving if the lead has to go past the case of the capacitor, so it cannot short out to the case.

[TimNJ]

Thanks everyone. I'll get my calipers out and measure all the electrolytics and find some replacements. Thanks.

[TimNJ]

Does anyone have one of these and know how to remove the PCB from the front panel?

[amspire]

The PCB slides into two slots from the front panel - just slide the board out. Be careful not to break the slots on the front panel.

You want to leave all the wires to the front panel still soldered - you do not need to unsolder them to replace the capacitors.

The replacement capacitors will be much smaller then the originals, so don't try too hard to get the same size.

[TimNJ]

Amspire,

Thanks, but sorry I think I should've worded my question differently. I was talking about the front panel PCB. Theres an electrolytic on there also. I think the binding posts are soldered directly to it and that's what's holding them in?

[amspire]

I wouldn't bother. First, that capacitor is not causing your current problems and you can probably just leave it. If it has become too low in value or too high in ESR, the supply will not regulate well with fast transients.

If it has gone dry, you can always just solder another cap to the rear of the PCB in parallel. You could unclip the meter and remove the capacitor in place, but there is a risk you will break the meter retaining clips.

When choosing caps, you can always use a higher voltage capacitor just to get a bigger case, but be careful about changing the value. A bigger capacitor may sound better, but it can mess up the functioning of the power supply. HP may have chosen the values so that one cap will charge or discharge quicker then another so that the supply behaves well on start up. If you go increasing values, your supply may start having nasty voltage spikes on startup or shutdown.

[TimNJ]

Ah yeah I guess you're right. No need to do that since that's not currently the problem.

[TimNJ]

Well going against what I previously posted, I decided to change out that cap on the front panel PCB. I have not yet received the other replacement caps, but I desoldered all of the ones I will be replacing and when I measured them out of circuit, only one of the four measured good.

The big 2000uF cap would not give me a reading at all on my AM-270 (it was going crazy trying to autorange) and my AM-220 gave me a reading of 0.8nF. It's one of those 4 legger caps, but I tried every combination just in case it wasn't the standard + center and - on the outside. But I'd say that's pretty bad...

Because I saw how bad these caps were, I decided to ease my mind and replace the 10uF cap on the front panel PCB. I think it's a decoupling cap of some sort. Value probably isn't critical anyway. When I pulled it, I measured 5.5uF. Still not good so I macgyvered a radial Panasonic 10uF onto the back of the PCB. Some peace of mind.

[TimNJ]

The new caps are in and it works a treat. 

Thank you everyone. I'll put a picture or two up so you can see how much smaller the replacement caps are.

[oilburner]

Hi Tim, did you every get a chance to take some pictures of your re-cap repair?
Thanks!

g