Agilent E3634A - fails with error 718

[vinicius.jlantunes]

Hi guys,

There were some other old threads about this PSU and defects, but I thought I would start a new one and link the old ones at the bottom.

I bought this power supply a while ago and it wasn't working to begin with. After cleaning it up though it started working just fine, not sure what was the deal back then.

I used if for a while, very lightly, just to power a couple LED's and stuff, and then it stopped working again - it turns on fine, even passes self test every time I try, but as soon as I turn on the output it says "error 718":
Gain out of range for Gain Error Correction
The slope of the DAC gain is out of range. Hardware fails.

That's the same error the guy in the first thread below had, and advice was to replace ADC op amps U25 (AD711JR), U28 and U35 (OP-27GS) and U41 (AD706002JR). If someone can lend some more help though I would like to ask for any more pointers or ways to test these guys before I try to replace them - parts in Brazil (even the most stupid opamps) are hard to obtain and expensive. Also the other thread seemed not to be conclusive if the fix was successful.

I did basic checks such as probe all labeled bias points and known voltages. The output measures constant -0.55 V. If I connect the sense terminals it goes to -0.526 V.
Sometimes I clear the errors over and over and then it stops happening - the output will still not change if I try though, and an indicator "Unreg" will light up in the display.

I will print the schematics tomorrow and try to study them a bit more as well. Again, any pointers are appreciated.
---
Old threads for reference:
https://www.eevblog.com/forum/beginners/did-i-kill-my-e3633a/
https://www.eevblog.com/forum/repair/video-teardown-and-repair-of-an-agilent-e3634a-power-supply/
https://www.eevblog.com/forum/repair/agilent-e3634a-power-supply-teardown-repair/

[idpromnut]

Not that it will be much help to you, but most of my experiences fixing these lines of PSUs is that the output pass transistor blows (or gets subtley damaged). If you have the schematics, I would start by trying to understand what the DAC's function is in the circuit and feedback loop.

If you need parts, throw a request up on the forums; I'm not sure what the shipping is to Brazil, but if a member can help out by buying and shipping the part(s) you need, there are those willing to help out

[vinicius.jlantunes]

I have been studying the schematics, not quite sure I get it yet, but I probed a couple more things and it seems that there is a -10 V reference used in the A/D converter section that is off - it reads -14.7 V instead of -10 V. Looks like the opamp output is glued to the negative bias voltage.
The positive supply also seems to be a bit low, but maybe still acceptable: 14 V, 1 V less than the specified 15 V. The negative supply is solid at -15.3 V.

The part I am talking about is U31, page 131 of the linked service manual. The schematic is attached for convenience.
http://gerrysweeney.com/?wpdmdl=1745

The IC is an AD706JR dual opamp, and its pin 7 is the output that is supposed to have -10 V.

Does my fault analysis make sense? I don't really understand the rest of the schematic, but this -10 V reference does feed via a MUX into the opamps that have been said to influence in the error I am getting (more specifically it seems to feed into U28).

[free_electron]

problem is that a fault in this section tends to damage the entire chain.

you can;t really see anything in voltage domain as this whole thing is a charge balancer.  they are throwin currents aroud....

[vinicius.jlantunes]

Hum, you mean I wouldn't necessarily see the -10 V at that reference output as the U32 MUX (74HCT4053) input isn't drawing any current (high Z state)?

I haven't been able to work on it yet this week, but would you have any other hints? Should I just try to replace all the opamps in the hope it fixes the problem, or that would be most likely a futile exercise?
I got a friend who is in the US at the moment so I could probably order the parts and have it delivered to him, then he brings it back to me in a couple weeks.

[vinicius.jlantunes]

I needed some "free air" from my work and decided to study a bit more about the way this power supply works and stumbled upon this thread which has an amazing lot of information by @free_electron and others... linking in here so I can re-read later.

https://www.eevblog.com/forum/reviews/teardown-agilent-e3631a-power-supply-(picture-heavy)/

Thanks to all those who contributed so far, be via this thread or others I came across. Unfortunately my free time to work on the fix is quite limited at the moment but I hope I will be able to fix this PSU with all I've learned from you guys.

[vinicius.jlantunes]

OK, I've had some time to play with this today. I am getting more and more convinced that the voltage references are out... more specifically, the section in the attachment.

Would someone be willing to probe a couple things for me in case you have one of these PSU's in working condition? It seems many of these HP/Agilent PSU's are similar, perhaps some other models in the same family could do the trick as well. The places to probe are listed down below.
That would be much appreciated! 

So here is my thinking: the +15V power supply is OK I guess, no ripple, a bit low perhaps at 14.1V, but probably OK. Then the zener is dropping 4.97V, leaving 9.13V for the LT1021 reference input. According to the datasheet, the reference output should be between 4.995 and 5.050V - I am reading 4.45V... I thought that maybe the source has to be drawing a minimum current or something for it to work, but reading the datasheet it seems not to be the case, to the reference is simply out it seems.
I ignored that for a moment though and read the output of the U31 opamp that is supposed to be at +10V - it reads 9.56V - that makes sense actually - that circuit basically doubles the reference voltage - it is slightly low but well, perhaps within specs still.
Then I read the -10V reference again, that reads -14.7V. I thought it could be because the opamp output was not loaded, so I pulled it down with a 15k resistor just in case, but still same reading.

Therefore, just to check if these readings are indeed abnormal, I would appreciate if someone could probe the following voltages in a similar PSU:

• +15V rail
• Voltage drop accross CR32 zener diode
• Voltage output of the LT1021 +5V reference U40 (pin 6)
• Voltage output of the +10V reference (pin 1 of the U31 opamp)
• Voltage output of the -10V reference (pin 7 of the U31 opamp

The attached picture shows the region where the components are on the board. They're not hard to probe.

[Hugoneus]

You are saying that the -10V reference reads -14.7. This seems like the Op-Amp is saturated to one of its rails. I suggest taking the Op-amp out of the board, then measure the reference again and the +10V reference. Measured all the voltages of the suspect opamp before you take it out. What are the voltages?

[vinicius.jlantunes]

Thanks Shahriar ! I will try that. I am still waiting on the hot air gun I purchased to arrive, so removing the SMD part at the moment might be a bit difficult. I will try removing it anyways and report back.

The -10V reference is actually on the same opamp, it's a dual AD706 opamp - the +5V reference is fed into both amplifiers to generate the +10V and -10V references.

I have already requested samples of this and some other parts from AD, so if I have trouble dessoldering the opamp I might just cut it off the board.

[Hugoneus]

Ok, good luck. Let us know how it goes.

[vinicius.jlantunes]

I finally got all parts and tools to attempt some sort of repair... no success still 

I replaced the reference (U40) and the opamp (U31) I thought were problematic. I was wrong about the reference, it was actually fine (tested after removed), but the opamp was indeed fried. I didn't have the AD part so I used the direct replacement TI part OPA2277.
In the process I found that the precision 20k 0.1% resistor (R71) was open! I don't have a replacement at hand but I used two 10k resistors in series; they're not precision resistors, but better than nothing (~19.9k, thus ~5% off).

I continue to get the same errors though, no change in behavior. I noticed one thing while doing some measurements though - the reason why I initially thought the reference was out is because I was using a ground point a bit far from the reference (the "minus" binding post on the output). I did so as I assumed the ground would be fairly good across the whole board as all traces are quite beefy, but it turns out it isn't, or there is some problem this could be hinting at.

I am stuck at this point... anyone has any ideas? Any theories as to what could've taken out the opamp + resistor, and what else it could have burned? And the ground thing, was I just being stupid and doing the measurements incorrectly, or could that hint at some problem?

[vinicius.jlantunes]

Well, I replaced another couple of opamps (U28 and U35), still no success. When I first turned on the PSU, same error. But then something strange happened...

After a while, it ceases to issue the error... it happens like a dozen times right after I turn it on, then I can clear it and it stops happening, almost as if it needed to "charge up" or "heat up" for a few seconds for it to fix itself... I still see the "unreg" indicator though and am not able to set the voltage / current output.
If I then turn it off and on again, error doesn't happen. I have to leave it off for a good while for the errors to maybe happen again.

Am I looking at the wrong place? Should I be looking at the DAC piece instead of the ADC?

[ytsejam]

Hi vinicius,

If it is an "unreg" issue, I'll say the DAC portion first.

Verify the U30 and U29 op amp output, especially CV_REF (0 ~ -12V), CC_REF (0 ~ +12V), all within the range.

If you have verified DAC portion and confirmed it is fine, then check voltage versus Pg (+Vout) between Q6 pin2 and pin3 while the output is on.
If Q6 is not turn on while you turn on the output, and R46 has approximately 6V, check the PCB trace from R46 to Q6 pin1, I had a broken one which caused "unreg" issue.

[vinicius.jlantunes]

I know this is and old thread, but just in case someone is interested - I have listed this power supply for sale on ebay - I have not tried ytsejam suggestion for a fix as I had not seen it, plus my electronics stuff did not come with me in my last move.
It does have the "unreg" issue - after it warms up a bit it stops showing error 718, but does not regulate the voltage and shows the "Unreg" warning on the display

https://www.ebay.com/itm/134982462229

I really just want it gone to a good home