Agilent E3631A issue - display shows limit values, no errors but no output (0,0)

[Spectrum]

Hi,
I'm here again to ask for help. This E3631A (22 yo -2000) has a rather complicated problem.
With self-tests there are no errors displayed.
The pushbuttons and knob are in working condition, thus I think CPU are good.
When I switch on the device, the display shows OUTPUT OFF and is in +6 V mode, I push the output ON and the display shows - 7,324 V and - 6,212 A, if I set in +25V mode the display shows -34,47V and 1,941A, if I set the -25V mode display shows 34,47V and 1,941A (always in CV and note the opposite sign with the mode). No possibility to change these values
The outputs are always 0,00V, measured with an external DMM.
The button "Display Limit" works.
The only way to have some voltage in output is recall the setting stored in Store, but the values shown in the display are incorrect compared to those measured with the DMM.

I did some research and I found that these power supplies have problems with the ADC and DAC circuits, as well as nearby resistance (R4, R5, R15, R16 etc ..). I measured all components close to U9,U10, U11, U12, U13 with the DMM and all seem in order, apart some not exactly precise value in resistance. Let's say that some precision resistances measure as normal resistances.
The power supplies of these above integrated circuits are good.

Where should I start?

Thanks to those who can help me and greetings to all.. Stefano

[SoundTech-LG]

When you turn the rotary knob...  any clicking sounds? Does the display ever change while rotating the knob?

[HighVoltage]

Try if the remote control works over GPIB?

[Spectrum]

Sorry I have no GPIB cable and sw

[Spectrum]

When you turn the rotary knob...  any clicking sounds? Does the display ever change while rotating the knob?

Hi thanks for reply... Ja, I hear clicking and display changes only if you are on "Display Limit" but not in normal mode or when you see the values I wrote

[SoundTech-LG]

Hi,

What is the history of the power supply?
 Is it New to You? Purchased used?

or, you have owned it, and it has worked fine until now?

[Spectrum]

Bought in this not working conditions, but in virgin conditions, never open so far. I do not know his history. The strange thing is also the value on the output sockets when OUTPUT is ON... it is strictly 0,00V... it appears to be short-circuited, but it is not

[SoundTech-LG]

You have the schematic I presume, with all the part locations you mentioned...

[Spectrum]

Yes I have the Service Guide with schematics but I don't know where to start and if the section I started investigating is the correct one

[SoundTech-LG]

Beeps twice on turn on? display reads 0.000V 0.000A  after pushing output on? ...and Terminals all have 0.00 Volts?

[Spectrum]

I do not remember about the beeps.
No, the display shows always - 7,324 V and - 7,212 A in +6V mode, -34,47V and 1,941A in +25V mode, 34,47V and 1,941A in -25V mode (in CV or CC in +6V mode). Note the opposite sign with the mode. No possibility to change these values.
The outputs are always strictly 0,00V, measured with an external DMM.

I am going to sleep  good night

[free_electron]

check the +10 and -10 volt reference. next up are the 4051 multiplexers. check that the DAC is running and the sample and hold amplifiers work correctly. hook a scope between the dac and the multiplexer. you should see a stepped voltages come out. if you rotate the know one of the "steps" should move up and down.

[Spectrum]

Thanks guys, where are the +10 and -10 volt reference ? ADC U3 (pin ?) or DAC circuit ? sorry I'm not so expert, I need your help...
I'm going to do other measurements...

some other info:
- Beeps twice on turn on
- In a complete (extended) self test, no errors occur
- R5 is a 30K 0,1% green resistor but it measures in circuit 31,45 Kohm (5%), could be that ? Where could I find these green resistor with these strange values ?.

[Spectrum]

Here I am again.
I have the correct voltages +10V and -10V on pins 1 and 7 of the U3 (OP-AMP AD706), but when I go to U5 (pins 14 and 15 through R5 and R4) I have negative pulses at 6-700mV at a frequency of 187Khz, is it correct ? Out of circuit R4 is 30K, R5 is 31,45 K as mentioned

[free_electron]

you can't measure at U5. That thing is a charge balancer and works i current domain.

What do you get at U4 inputs. pin 13 15 and 1 are of interest ( VMON signals)

next thing to check is U10 and U9 outputs. if you spin the encoder for the voltage you should see the CV signals go up/down.  I've seen those TL074 die.

i'm not too worried about that resistor being 31k. the converter should still work.

[strawberry]

display readings show ADC overflow. could be U4 pulling to supply rails , maybe U7 ...

[IanJ]

Hi,

Have a look at my YouTube channel, i have repaired a few Agilent E36xx psu's, and one of them had same symptoms as you are getting.
Worth a look.

Ian.

[Spectrum]

Hi Ian,
searching in your videos I found two (part 1 part 2) about this PSU E3631A. A few days ago I searched on YT but I haven't seen your videos.

Unfortunately my understanding of spoken English is not very good and therefore I have not understood much of your speech (maybe putting some subtitles, even in English, can be useful).
Anyway the guilty area is always the same. I have reference voltage +/- 10V...something get out from DAC (pin 3 of U13), probably U9 and U10 are suspected as Electron Wrangler suggested.

[Spectrum]

Dear guys, the investigations continue...
I did some measurements with a Tek TDS2012 oscilloscope. First I can say that I find U9 and U10 slightly warm and during the measurements I noticed that the heatsink where U24 U25 and Q7 (Bias Supplies) are mounted, gets quite hot, a sign that something is charging it.
Sorry, I did not measurements about U7, as suggested from Strawberry, because I don't know what kind of signals I could find.
Let's see what I found:
U4 pin 1 - only in -25v mode, I have a negative square wave of 1,4Vpp
     pin 5 - no signal in every mode
     pin12 - no signal in every mode (100mV noise)
     pin13 - only in +6V mode, negative square wave 0,7 Vpp
     pin 14- the same as pin 13
     pin15 - only in +25 mode, negative square wave 1.4 Vpp (the same as pin 1 for -25 mode)

U10 pin 4/11 +/- 15 VDC
     pin 1  - in +25 and -25 mode I have a - 2.71VDC fixed, not variable with the encoder
     pin 8  - +6CV flat EEG (flat line)
     pin 14 - +25CV flat EEG

U9 pin 4/11 +/- 15 VDC
     pin 1  -  flat EEG (flat line)
     pin 7  -  +2.36 VDC fixed, not variable with the encoder
     pin 8  -  flat EEG (flat line)
     pin 14  - + 2.62 VDC fixed, not variable with the encoder

Pin 9-10 U12 (DAC output) I have something like this:
https://www.eevblog.com/forum/repair/agilent-e3631a-strange-output-behavior/?action=dlattach;attach=120724;image

Really thank you for your attention

[free_electron]

pin 10/9 of the dac is correct.
the outputs of the opamps u9 and u10 is wrong.


replace U13 , U9 and U10

[strawberry]

U4 seems strange. MUX is not loaded so expected ~0VDC on all inputs as PSU output terminals
if IC on 15V rail then U7 could be loading it

[Spectrum]

pin 10/9 of the dac is correct.
the outputs of the opamps u9 and u10 is wrong.


replace U13 , U9 and U10

Thanks, I search these components and let you know

[Spectrum]

U4 seems strange. MUX is not loaded so expected ~0VDC on all inputs as PSU output terminals
if IC on 15V rail then U7 could be loading it

So what do you think ? How would you proceed ? I can replace also U4, is the same MUX of U13, so I will buy nr.2 74HC4051

Thank you to all for the help

[free_electron]

i typically replace all that stuff in one shot. i've seen various failures and you spend more time than it is worth troubleshooting this stuff.

4051 lm339 , the opamps and comparators in the A/D

U3 U4 U5 U6 U7 U8 U9 U10 U11 and U13  are cheap. just replace all.

i've never seen the DAC go bad. it's always the parts that get exposed to the 15 volt rails that die. (U4 can die if U7 goes bad and shunts 15 volt to pin 3...)
i have long suspected latch-up due to the rails starting up out of sequence... i've repaired at least 20 of these supplies in various incarnations( single/double/triple ). 99% it is in that circuitry exposed to +/-15 volts.
i had one where the analog optos failed, one where the 5 volt for the outguard processor did not work and one where the cherry regulator fried. the easiest fix was the rom that had fallen out of its socket...

oh, and don't trust those schematics ... where is the power and ground for U5 ? pin 8 and 16 are missing ...

there is also two 15 and -15 volt rails and they are NOT the same !

These supplies are horribly over-engineered. The base regulator circuit is the same as a E3611/E3611 ! one voltage error amp and a current error amp driving a power transistor. All the rest is lipstick, glitter and hats.

Convoluted analog optical isolators with feedback (those opto's die ...), a VFD frontpanel that needs level shifting to -18 volt domain because of the filament (there are other ways to do this where you don't have to do the level shifting)
Another optical bridge for gpib/serial. two 8051 processors , an 80188 and an asic ( basically recycled from their 34401 multimeters)
They could have made all logic earth referenced and simply had shift register over optical drive two dacs. . They spent more money in analog optocouplers than would have been the cost of an additional dac. that whole dac followed by sample/hold is ridiculous. that may have been cost efficient in the 70's, but these things were designed in the 90's. But hey, they had the asic and the whole platform of the 34401... strip the input range , tack on a 3410 regulator or two and off we go.

[strawberry]

by removing U4 ADC should pickup static noise or external source and display on front panel
ADC 0.3V = 25V
U4 is directly connected to current shunt 25V_I_MON 0.2V@1A

oh, and don't trust those schematics ... where is the power and ground for U5 ? pin 8 and 16 are missing ...

table on page 128

[free_electron]

every other symbol on that page has power pins.... just not that one...  that one is referred on the processor page... in a table. very weird...


i've found mistakes in other schematics too... E3610 schematics show mosfets as power stages. some versions show 4 ... there's 2 and they are darlingtons... Biassed darlingtons on top of that ! i could understand if they shown two fets and two darlingtons to permutate between the 3610/3611 and the 3612 (that one uses irf440 fets. also a rare beast.) but no.. 4 fets two each with same ref des.

[IanJ]

Hi Ian,
searching in your videos I found two (part 1 part 2) about this PSU E3631A. A few days ago I searched on YT but I haven't seen your videos.

Unfortunately my understanding of spoken English is not very good and therefore I have not understood much of your speech (maybe putting some subtitles, even in English, can be useful).
Anyway the guilty area is always the same. I have reference voltage +/- 10V...something get out from DAC (pin 3 of U13), probably U9 and U10 are suspected as Electron Wrangler suggested.

I just had a quick look at my own repair video(s) on the E3631A and note that most if not all of the problems on my unit were open circuit resistors, specifically just about all the GREEN coloured resistors. So, before changing out any IC's I would go over those resistors and make sure they are ok.

Ian.

[Spectrum]

Dear Wrangler  , I repaired another E36xx series 4 years ago (3632 if I'm not wrong, 15V7A or 30V4A 120W) and in the desk I think have some OP27 and AD706, so I can replace all. I just ordered 4 TL074C and 2 4051, probably next week will arrive the package. I think it is better to desolder the 220nF capacitors near U9 and U10, to have more space and work more accurately.

All true what you say, in fact these PSUs are famous for this, however they are very beautiful, expensive (even used) and although I would never spend the money they cost, when I find a used one for about 90/100 euros, I can not resist and buy it.

All the rest is lipstick, glitter and hats.

here we say A lot of smoke and little roast 

Ciao and good night

[Spectrum]

Hi Ian, yes I know. if you read at the beginning I have checked these green resistors, also out of circuit to be more certain (R4 and R5, R5 is 31,5 Kohm, datasheet 30K 0,1%  )

[free_electron]

most if not all of the problems on my unit were open circuit resistors, specifically just about all the GREEN coloured resistors. So, before changing out any IC's I would go over those resistors and make sure they are ok.

the electrodes on those things dissolve over time ... i've seen that happen too. but only on supplies that were in relatively damp areas.

[Spectrum]

Hi, finally I have the necessary ICs. Replaced U4, U7, U9, U10 U13 but nothing has changed....

[ThaiGuy]

HI Spectrum,
Would you mind sharing the schematics?
THanks

[ThaiGuy]

Hi Spectrum,
Would you mind sharing the schematics?
Thanks

[alm]

Just find an older version of the service manual before Agilent removed the schematics. For me, the second and third Google result for e3631a service manual were:
https://engineering.case.edu/lab/circuitslab/sites/engineering.case.edu.lab.circuitslab/files/docs/Agilent_E3631_Power_Supply_Service_Guide.pdf
https://dev.xdevs.com/attachments/146

Both contain the exact same file that as far as I can see includes complete schematics.

[ThaiGuy]

Hey Alm,
Thanks a lot pal.