Author Topic: [SOLVED] Agilent E3631A problems.... I just keep getting lucky... NOT.  (Read 4094 times)

0 Members and 1 Guest are viewing this topic.

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
So I have an E3631A power supply on the bench.  (Given to me) It turns on, beeps twice and comes up with output off.  The FULL self test passes, NO ERRORS.  If I recall correctly, it should only beep ONCE.

I can set the limits on all 3 supplies but when I enable the output on ANY of them, I have NO control to set voltage or current.

On output enable the +6v supply reads -.245vdc and .037A CC.  Voltage reads same on external meter.

+25 reads .09v and .042A CV.  Also correct on external meter.
-25 reads .29v and .028A CC.  Also correct on external meter. 
None are adjustable, even though the digit blinks and the encoder IS working.

Preliminary measurements show all voltages on the power rails as well as bias supplies are in spec, no shorted outputs.  Seems the power stages are good.  All fuses are good.

Aaand... it feels like a logic issue.... at which point I am totally out of my field of expertise. 
Any thoughts?  Seems that digital stuff and I just don't get along.  No such problems with older HP analog stuff.  But NOOOO... I wanna be FANCY!!   |O

Where do I start on THIS one?
« Last Edit: August 04, 2021, 05:43:56 pm by Runco990 »
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #1 on: July 19, 2021, 05:37:07 pm »
Doing a little digging, I get the DAC inputs to U13, but no change out on any lines.

plus and minus 5 are ok, so is the inhibit pulse.

So my next guess is U13 is bad.  Well, it's pennies, so I'll replace it and see.

 

Offline antimatter503

  • Contributor
  • Posts: 15
  • Country: us
  • Ion Implantation Equipment Engineer
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #2 on: July 19, 2021, 06:33:15 pm »
Got any pictures of the board or schematics?
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #3 on: July 19, 2021, 06:45:12 pm »
Here's a manual:  https://engineering.case.edu/lab/circuitslab/sites/engineering.case.edu.lab.circuitslab/files/docs/Agilent_E3631_Power_Supply_Service_Guide.pdf

I ordered U13, as the scope shows the DAC pulses going in, and changing with the rotary encoder, but nothing going out.  Basically flatlined.  Will try that and see.  Easy enough to change.  To me, that makes sense since all 3 supplies are equally affected.
 

Offline AaronLee

  • Regular Contributor
  • *
  • Posts: 229
  • Country: kr
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #4 on: July 20, 2021, 01:34:21 pm »
So I have an E3631A power supply on the bench.  (Given to me) It turns on, beeps twice and comes up with output off.  The FULL self test passes, NO ERRORS.  If I recall correctly, it should only beep ONCE.

I can't really help you much other than I powered on my E3631A and indeed just a single beep, so your recall is correct. On mine, the sequence is first it lights all the segments on the display, then displays, "ADDR 5" while the single beep, then "OUTPUT OFF".
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #5 on: July 20, 2021, 03:05:30 pm »
Yes, I get the same sequence but no controllable output.  As I said, it looks like U13 is at fault, but I cannot be 100% certain.
So I'll pop in a new one when it arrives and see.  Go figure the "self test" is useless...  NO ERRORS.
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23017
  • Country: gb
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #6 on: July 20, 2021, 04:37:48 pm »
Probably is the 4051 at U13.

TTi power supplies tend to die with similar issue and it's always the bloody CMOS 4052 that is knackered, every time!
 

Offline perieanuo

  • Frequent Contributor
  • **
  • Posts: 838
  • Country: fr
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #7 on: July 21, 2021, 05:58:30 am »
Doing a little digging, I get the DAC inputs to U13, but no change out on any lines.

plus and minus 5 are ok, so is the inhibit pulse.

So my next guess is U13 is bad.  Well, it's pennies, so I'll replace it and see.
hi, before ordering, you could just put a potentiometer or a resistor divider and simulate the DAC output in order to get output from your PS, of course with the dac output separated from this test rig
it's not a problem of pennies or euros, it's the correct diagnose flow
i rarely saw a dead dac ( dead duck :) ), but why not, anything can happen if you live enough to see it
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #8 on: July 21, 2021, 02:54:49 pm »
Doing a little digging, I get the DAC inputs to U13, but no change out on any lines.

plus and minus 5 are ok, so is the inhibit pulse.

So my next guess is U13 is bad.  Well, it's pennies, so I'll replace it and see.
hi, before ordering, you could just put a potentiometer or a resistor divider and simulate the DAC output in order to get output from your PS, of course with the dac output separated from this test rig
it's not a problem of pennies or euros, it's the correct diagnose flow
i rarely saw a dead dac ( dead duck :) ), but why not, anything can happen if you live enough to see it

U13 is the multiplexer that splits the DAC output into 6 analog channels.  The inputs are correct, there is nothing at all coming out of the multiplexer.  So I tend to believe the Multiplexer is bad.  Dac looks fine.
Will know for sure once the new chip gets here.

The problem is that I am also getting ready to move, so I have to shotgun a bit.  Time is ticking.
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #9 on: July 24, 2021, 03:06:45 am »
Swapped the multiplexer IC (U13) aaand... same thing.  I really hate digital stuff... maybe my age is showing, but I didn't grow up with this stuff, so having a hard time grasping the trouble shooting flow of DAC's and CPU's... etc.  Been repairing analog gear my whole life without problems, but this stuff is just making a damn fool out of me.  |O

So tomorrow I'll do some more probing, I printed out some clear schematics rather than staring at a PDF.

All 3 supplies are affected so there HAS to be a common point there.  All Power rails and bias supplies check good.  Like Dave said, thou shall check all power supplies first, which I did.  I have also checked diodes and resistors, which also check good.  (Read a few past repairs on these)

Thanks Agilent for the built in self diagnostics which say everything is fine!

Hmmm...  But I am stubborn and eager to learn, So I'll keep at it.  Eventually I WILL understand this stuff...  or die trying!  :-DD
 

Offline perieanuo

  • Frequent Contributor
  • **
  • Posts: 838
  • Country: fr
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #10 on: July 26, 2021, 05:37:08 am »
told you before, simulate dac output first in order to see it works, than you attack from processor to dac output with some logic in repairing steps
isolate dac output and see what happens, but if your processor gets feedback and that feedback is not working (like processor says to dac: go 50%, the dac doesn't do those 50% correctly or he does but the analog part fails to impose 50% to power module, your feedback reports this problem and maybe the processor stops imposing 50% and throws out an error, have you considered this scenario?)
try breaking this repair in logical verification steps like i suggested before, at least you eliminate some stuff from the equation...
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #11 on: July 26, 2021, 06:32:24 pm »
Yes, you told me before.  I was hoping to shotgun it fast because I actually have no time to be doing test equipment repairs at the moment.  I am in the process of getting the house ready to sell, packing, moving, etc.  I may have to pack this up and worry about it when I land in the new digs.  If I find some time I will try your approach, as it makes sense.  I have been studying the schematics a bit and am beginning to understand the flow of things.
 

Offline pnv57

  • Contributor
  • Posts: 18
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #12 on: July 26, 2021, 08:04:45 pm »
What happens when you attempt a calibration?
 

Offline perieanuo

  • Frequent Contributor
  • **
  • Posts: 838
  • Country: fr
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #13 on: July 27, 2021, 06:53:13 am »
no problem, to be continued
i doubt recal will repair a hardware damage like your case seems to be
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #14 on: July 27, 2021, 02:08:53 pm »
Recal does not work.  Voltage never changes.  Tried that in the beginning.

Will see if I have a little time to continue later.  Many boxes to fill, paint to do....  please kill me...  Moving sucks.
Just moving my SHOP will take an 18 wheeler!  Never mind the HOUSE!
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #15 on: July 27, 2021, 05:42:39 pm »
So while the painter is here....  Tested the ADC by injecting voltage into the front and the supply reads back correctly on all 3.  So that's fine.

Back to the DAC.  Tested everything I could and I "THINK" the actual DAC (U12) may be bad.  I am not sure how to test this... nothing changes on the output as I run the settings up and down. 

I'm getting pulses on the output lines, but they are always the same.  They used an audio DAC????
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23017
  • Country: gb
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #16 on: July 27, 2021, 06:20:21 pm »
This is where things get interesting. Have posted the schematic here for reference:



Circuit is a sample-and hold amplifier gated with U13. DAC is U12. U11 is a serial-parallel register which appears to act as an IO expander for the bus. Yuck but saves pins on the MCU I suppose and costs only one DAC. 90s design. We have nice I2C DACs these days with bazillions of outputs which would save all this faff.

I'd check the following in order to eliminate the DAC:

1. Check supply voltages at U12 1 and 16 are present (-5.1, +5v). Check supply voltage at U11 pin 15 present (+5v).
2. Check CLK (5), DATA (7), LE (6) pins for any changes at the DAC pins with scope or logic probe as you change stuff. DATA is interesting as it's routed through the U11 register which means that could be broken as well and stopping data getting to the DAC.
3. Check U13 pin 6 changes state. This goes LOW to send the voltage from X->Xn. That should confirm most of the shift register works.

After this I'd probably bust out the aliexpress logic analyser myself and start reverse engineering it properly
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #17 on: July 27, 2021, 08:18:23 pm »
Well, this is where MY knowledge is strained....

All supply voltages are present and correct.
Clock, Data and LE have logic pulses on them.
U13 pin 6 has logic pulses on it.

What is pin 3 on U13 supposed to be?  It is flatlined.  Is it supposed to be a DC voltage or a logic signal?
I did replace U13 because NOTHING comes out of it, but same exact result.  Here's where my digital knowledge needs updating... should the outputs from U13 be DC voltages or logic pulse trains?  I get basically zero.

At least I know the readback from the front works, as changing the input voltage to the front jacks with an external power supply reads correct on the display.  So all the ADC stuff is ok.

I watched Ian's videos on this and followed along as well, checked all the diodes and resistors which all test in spec.

Obviously (or not) all main rails are also correct.
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23017
  • Country: gb
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #18 on: July 27, 2021, 08:33:08 pm »
Pin 3 of U13 is the DAC output. Think of U13 as a bunch of invisible switches of which one can be active at once which are controlled by inputs A,B,C. If you set a binary value on ABC, then drop INHIBIT low it'll close the relevant switch and send what is on pin X (from the DAC) to one of the outputs as selected by ABC which will set the voltage on one of the amplifiers. That'll remember it until it goes through and does it again.

4051 guts roughly explained... excuse the chicken scratch!



This circuit should basically set the voltage on capacitors C10-15 etc based on the digital value shifted in.

Last thing to check is if anything is happening on U11 pin 10. If nothing is happening there then it's U11. Else replace the DAC.

As for reverse engineering I did some reading of the datasheets and basically what it does is shift out 24 bits per instruction. And yes you're right it is an audio DAC. Probably the best thing for the job here.
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #19 on: July 27, 2021, 08:58:09 pm »
U11 pin 10 has data.

So with basically 0V DC on pin 3 of U13 we can confirm the DAC is in fact dead?
« Last Edit: July 27, 2021, 09:02:48 pm by Runco990 »
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23017
  • Country: gb
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #20 on: July 27, 2021, 09:08:01 pm »
Smells like it might be that.

I'd fire up a logic analyser on the DAC input here and check it first but probably cheaper and less annoying to replace it and see than futz with that.
 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #21 on: July 27, 2021, 09:09:49 pm »
Already ordered.  $9 shipped.  At least I rock at surface mount soldering!   :-+
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23017
  • Country: gb
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #22 on: July 27, 2021, 09:13:00 pm »
That's pretty cheap. Good luck. Fingers cross for you here  :-+

 

Offline Runco990Topic starter

  • Regular Contributor
  • *
  • Posts: 239
  • Country: us
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #23 on: July 27, 2021, 10:29:35 pm »
I shall update when the DAC arrives.   :-+
 
The following users thanked this post: bd139

Offline perieanuo

  • Frequent Contributor
  • **
  • Posts: 838
  • Country: fr
Re: Agilent E3631A problems.... I just keep getting lucky... NOT.
« Reply #24 on: July 28, 2021, 05:03:03 am »
whenever i got patients like yourself, i order the logic IC chips, should cost almost nothing compared with value of the device itself.
4094 may be sending incorrect data, meaning sync between LE and DATA/CLK is critical
let's hope then it's only the DAC affected, but i never changed a DAC in 25 years :( usually the registers and multiplexers used to 'die' on industrial control boards, along with opamps.
a 4-channel scope would point that, if LE/CLK/DATA is consistent for the AD1851 and then you could be sure
4051 couldn't be excluded at this point, input pin no 3 can be damaged, that's why my original proposal could eliminate one suspect out of three (if 4094 pin 10 outputs something, i'll assume the 'guys' before that partial schematic are sending the correct signals with the correct timings between them). that doesn't eliminate 4094 as a suspect, other output may be problematic, but i doubt him.
let's pray 4051 it's fine...
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf