Agilent 34410A, Repair of 2 dead meters [success]

[Macbeth]

Oh you tease! You know we are all waiting with baited breath 

[plesa]

Strange that my meters have the debug connector and yours not. Looking forward for test after cleaning. Maybe put it into oven and set it to 50°C for half hour.

The alcohol evaporated really quick. So, I am thinking that the oven is not needed. Would have been a good idea but it is already assembled now.
The fan in the 34410A in notoriously loud and I installed another fan with an additional 100 Ohm series resistor. The fan is still moving the air pretty well and I am just under the assumption that this should be enough airflow.

Next thing is to turn it on and take some measurements.

Hot air gun can do the job as well. Surface tension hold the liquid below IC for long period.
Yes, the fans are horrible, maybe you can try  to switch it off and check with thermal camera and later on stability against voltage standard.
I make same experiment with L4411A where the fan is more noise and without cover it is OK, but with covers fan needs to be blowing air out of unit.
34410A/34411A have more room for convection, so maybe it is not needed ( like 34401A which is passive only].

[HighVoltage]

It works !
Well, at least it seems it works well so far.
Here are some pictures of the first measurements.

The 1 MOhm standard fluctuates a little too much, I think
The 100 Ohm measurement is almost rock solid at 10 NPLC
The 1 V, 10V and 100 V measurements are also rock solid at 10 NPLC

I think the display strength has lost a little brightness on some segments.
At least it appears to be not so bright on some spots.

Ok, I will close the lid now and let it sit for a few hours and then take some detailed
measurements from my known 10V DC Geller Labs SVRT standard.
 
What else should I measure or look out for, to make sure it is all in order.

BTW, the fan turns well but make almost no noticeable noise at all.

It seems the repair of this first meter was a success so far.
Thank you all for your suggestions.

 

[plesa]

Congratulation to successful repair!! What about second unit?

[HighVoltage]

Congratulation to successful repair!! What about second unit?

I have to wait for the switch and relay from Digikey for the second meter.
But I will start cleaning it already and as you said, keep it in an oven at 50 degree for a few hours.

[HighVoltage]

After about one hour, the 34410A has settled to 10.00012 V
My Geller Labs SVRT is actually verified at 10.000 028 82 V by a calibration lab this summer.
So, it seems the 34410A is off by about 100 uV in the 10V scale

[plesa]

is it too early to measure input noise? (like that in DMM noise thread?) im esp interested in 1NPLC p-p noise

In this thread I posted noise measurement from 34411A,34410A and L4411A. On xdevs.com FTP are source data

[HighVoltage]

I will measure the input noise in a few days.
Then we have good comparison to plesa' meter.

[Kleinstein]

After cleaning with alcohol it may take a few hours in the oven or a few days of operation to drive out residual solvent from some parts or the PCB itself. So I would not take the absolute voltage reading to serious yet - this might still drift a little. At least it looks good.

Ohne thing to check might be input bias/leakage current in the voltag ranges (especially 0.1 V and 1 V). E.g. measure the voltage when shortet with a 10 M or 100 M resistor. Alternatively measuring drift with a good quality 10 nF cap at the input should give a similar information.

[HighVoltage]

Great suggestions, Kleinstein. I will do that in a few days, for now I will just keep it running on the Geller Labs SVRT.

Here is a picture of the unit 14 hours later:
The 34410A still shows 10.00012 V
In comparison the Keithley 2450 is exactly at 10.00000 V.

[TiN]

And the voltnut was born...
Good job, keep it up.

I don't remember seeing much DMMs with triple-color VFDs! Green is rare in my book

[jitter]

10.00012 V is within tolerance, isn't it? I wouldn't be surprised if this is "as good as it gets".

At work we have several 34401As and I hooked a few of them up to a calibrator set at a lowish value (I believe it was 5 mV) and the meters set to 6 digits readout. They all displayed a different value, but all within spec.

Guess I'm not a voltnut, I often switch these meters to 4 digit mode to reduce the clutter on the display (those least significant digits are just noise to me) and speed up the update rate of the reading. This helps me adjust pots without overshooting the target and I find that the instruments I calibrate with these meters set at 4 digits are no less well calibrated than if I had done it with the meter set to 5 digits.

[HighVoltage]

I just hooked up the Keithley as a current source and it seems the 34410A is also full in specification in this current range.
Now back on the SVRT.

[plesa]

If you have such a decent SMU you can measure the input bias directly by SMU
When I measured on 34410A it was 4.6 pA at 10M input impedance and about 0.65pA at Hi-Z mode.

[Kleinstein]

I would not trust the SMU very much at pA currents. Just the DMM itself and whatching a capacitor charge should be more reliable. Using just a resistor might measure a slightly different current, measureing mainly the active phase, if there is not much capacitance to average.

[HighVoltage]

OK, On to the repair of the second 34410A
The switch and relay arrived in the mail from Digikey and are installed.
The unit turns ON and shows numbers on the display

But no matter, if I use the front or rear input terminals, nothing gets shown ion the display
As if the input is broken somewhere.
No matter, if OHM, VDC VAC or anything else, if I apply an input signal to the terminals, it shows nothing
The displays stays at 0.0 and fluctuates around the noise level.

Then I looked closer and found the damaged resistor next to K201
So, will have a look what is broken there and for what reason.

Stay tuned....
 

[HighVoltage]

The big blue relay in the last picture is dead.
This is K200 in the schematics and the input signal stops here.

Its made by MEDER Electronics
BT05-BV50324

3200-0121 (COTO number)
0490-1914 (Keysight number)

Rly-Reed 2A,5VDC Coil
Keysight price is US$54.12

Why is that one so expensive?
I can not find anything comparable at Digikey or Mouser

For now I will install a bridge to continue the repair.

[dom0]

I spoke to Coto today and got a data sheet for this 3200-0121 special part. It has a max contact resistance of 200 (10mA) to 250 (50mA) ohms. The problem with this is that the design seems to be very sensitive to this resistance remaining stable. It is a 1kV rated part and low EMF (2uV).

The alternative Meder part is custom, too...

@Blown-resistor.jpg
Q305-Q307 ... HP continued to use the emitter follower stack/cascade to protect the ohms source? Why am I surprised to see it, it is a very nice design after all...

[HighVoltage]

OK, it seems the input protection was blown...
- Relay K200 was dead and did not turn ON
- R210, 316 Ohm, burned up
- L200, 10 uH, burned up
- Relay K201, contacts burned

I replaced K201 with a new relay and just installed a 1 Ohm temporary resistor instead or R210/L200 to bridge it.
And it worked !

My 10V SVRT shows 10.00008V

OK, now I need to find the relay and the two little SMD components
Hopefully all other measurements will work as well.

[HighVoltage]

I spoke to Coto today and got a data sheet for this 3200-0121 special part. It has a max contact resistance of 200 (10mA) to 250 (50mA) ohms. The problem with this is that the design seems to be very sensitive to this resistance remaining stable. It is a 1kV rated part and low EMF (2uV).

The alternative Meder part is custom, too...

@Blown-resistor.jpg
Q305-Q307 ... HP continued to use the emitter follower stack/cascade to protect the ohms source? Why am I surprised to see it, it is a very nice design after all...

Thanks for the info on the Meder Relay
I probably order one from Keysight to get the original relay. But 50+$

Yes, the emitter follower is really nice indeed.
And guess what... 2 W resistance also works perfectly.
 

[PA0PBZ]

OK, it seems the input protection was blown...
- Relay K200 was dead and did not turn ON
- R210, 316 Ohm, burned up
- L200, 10 uH, burned up
- Relay K201, contacts burned

Looks like my 'parts donor' theory was not so bad after all...
Good luck with the repair!

[ManateeMafia]

I ordered a replacement relay for one of my K2001 from here
http://www.testco-inc.com/

Good prices but I don't know if they have your model.

[Dr. Frank]

Hi, HighVoltage,

might you please take a photo of R225, that is the 100:1 HighVoltage    divider, this white ceramic part, from both sides?
I'd like to know, if that component is also used in the new 34465A/34470A... I assume that their schematics are very similar.

May you please check also, if that component is heated (100mW), when you turn on the DC 1000V range without any input voltage.

The heating connections are in the schematics, pin5 , versus pin6 GND.
You should find up to 2.5...3.3V across this heating resistor.

Would you mind to measure the value of this heating resistor, please?
It should have less than 100 Ohm, probably 50 Ohm ,I guess.

Very interesting part & principle, anyhow.

Thanks a lot
Frank

[HighVoltage]

I ordered a replacement relay for one of my K2001 from here
http://www.testco-inc.com/

Good prices but I don't know if they have your model.

After looking around, I can not really find this exact relay anywhere.
I will just order the original Keysight part number as a replacement.

[HighVoltage]

Hi, HighVoltage,

might you please take a photo of R225, that is the 100:1 HighVoltage    divider, this white ceramic part, from both sides?
I'd like to know, if that component is also used in the new 34465A/34470A... I assume that their schematics are very similar.

May you please check also, if that component is heated (100mW), when you turn on the DC 1000V range without any input voltage.

The heating connections are in the schematics, pin5 , versus pin6 GND.
You should find up to 2.5...3.3V across this heating resistor.

Would you mind to measure the value of this heating resistor, please?
It should have less than 100 Ohm, probably 50 Ohm ,I guess.

Very interesting part & principle, anyhow.

Thanks a lot
Frank

Hello Frank,

Here are some pictures of R225, the 10M Divider