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

[HighVoltage]

I am starting this thread at the beginning of the repair of 2 dead meters.
Both are AGILENT 34410A
Both units show garbage on the screen, like total nonsense values in any mode.
But they were cheap as a pair and I would be happy to get one good one out of the two, but I am aiming to repair both.

The first unit seems to be complete after taking it apart.,
The second one has a missing Rear/Front switch and a missing relay.

I ordered a switch and relay
Switch part number: 51902 (F8UEE), 8PDT PUSH-PUSH UNSLD
Relay part number: G6AK-274P-STLT-US-DC5

This is my first serious repair of a 34410A. So far, I have only repaired plenty of 34401A meters but never a 34410A
The good thing is, that the Service Manual and the Schematics are available for download.
But since this meter is very different from the 34401A, the repair job might be different as well.

Does anyone here have experience in repairing the 34410A?

Full schematic:
http://www.keysight.com/owc_discussions/thread.jspa?threadID=36971&tstart=-1

[HighVoltage]

PLESA posted a picture of a 34410A main board in his upgrade thread.
https://www.eevblog.com/forum/testgear/possible-upgrade-agilent-34410a-to-34411a/msg814698/#msg814698

Interestingly, his board has the typical green PCB color.
The main boards in both of my units are much darker.

Also, why is R3121 not populated?
May be only in the 34411A?


 

[Jay_Diddy_B]

Hi,

The schematic shows two shunt resistors in parallel R213 and R2131. R2131 is marked 'never load'. It is probably allows them to use two different kinds of resistors that don't have the same footprint.

Regards,

Jay_Diddy_B

[jitter]

PLESA posted a picture of a 34410A main board in his upgrade thread.
https://www.eevblog.com/forum/testgear/possible-upgrade-agilent-34410a-to-34411a/msg814698/#msg814698

Interestingly, his board has the typical green PCB color.
The main boards in both of my units are much darker.

Colour variations between pcbs are very normal and are not indicative of anything other than usual variations in processing.

Usually pcbs are manufactured in batches not too long before they're needed. Bare pcbs are not as durable as one might think. Solderability may only be guaranteed by the manufacturer for a year or so when stored in good conditions (vacuum sealed packaging and with a bag of dessicant).
Older pcbs, or those exposed to air must be baked before stuffing and soldering. Even older pcbs might not be solderable anymore without them getting a special treatment which I believe isn't very cheap.

Every batch can turn out slightly different in colour. And switching between manufacturers (most of which are in China) goes on all the time, even though the people selling them to you remain the same. I don't know how many times I've seen pcbs change from matt green to glossy green and back again...

Usually there will be a datecode somewhere on the pcb. Often you can also find the logo of the actual pcb-manufacturer rather than the brand of the product (unless they're specifically instructed to leave those off).

[PA0PBZ]

The second one has a missing Rear/Front switch and a missing relay.

Hmm, I hope this is not what I think: More defects and used as a parts donor....

[HighVoltage]

Thanks for the feedback

I do have a 34410A in good working condition, so I can take some measurements if needed.
There was so much dirt in both of these used units, there is a possibility that the dirt was part of the problem, creating some lower resistance on the PCB. So I took both main boards and gave them a good cleaning in alcohol.

Hmm, I hope this is not what I think: More defects and used as a parts donor....

Yes, it is a possibility.
But why would someone take a relay out as a donor part?
I will only be able to tell, after installing a new switch and relay.
The lever for the rear/front switch is also missing, but it is available from Keysight as a replacement part.
(Hopefully it is still available for sale)

In general, I can not believe how badly some people treat their equipment.
But then, they might not even know the value of such precision instrument.

[TheSteve]

Are they actually functioning but just have a scrambled display, is the any com via the gpib port?

[HighVoltage]

Are they actually functioning but just have a scrambled display, is the any com via the gpib port?

The displays where updating with very unstable readings but even the one without the relay and missing switch had some numbers on the display.
They just did not make any sense.
For instance, in VDC mode it showed values around 1 VDC but constantly changing

I now cleaned the boards complete and will let them sit over night.
Tomorrow I will make an attempt to re-assemble the fist unit and take some measurements

BTW, it was nasty sticky stuff on the PCB, no idea in what environment they have been used.
But everything came off very well with cleaning alcohol.

[Kleinstein]

A first step would be checking for damaged electrolytic caps in the supply area. Only than power tham up and check the internal supply voltages.

If the display is showing numbers, this allready quite a good sign - so display part seems to work.
The internal selftest may help a lot.

[MadTux]

As already mentioned, check the voltages first. Then check if the relays at the input work. Apart from that, it's crazy how few and custom made parts are in there compared to older DMMs like 3456/3457 or Keithley 2015. One PLCC at the input, one FPGA probably as ADC controller and BGA microcontroller with custom Agilent part number. Any one of these is dead, not much you can do, except for the FPGA. Also, BGA soldering is not fun.

[Macbeth]

Apart from that, it's crazy how few and custom made parts are in there compared to older DMMs like 3456/3457 or Keithley 2015.

Thats a good thing, surely? Less custom parts the better.

One PLCC at the input, one FPGA probably as ADC controller and BGA microcontroller with custom Agilent part number. Any one of these is dead, not much you can do, except for the FPGA. Also, BGA soldering is not fun.

BGA is a doddle, just ask Louis Rossman.

What is a fair price for a used/may need some repair 34410A? They do seem a bit nicer than the 34401A's...

[MadTux]

I was maybe a bit unclear, very few and nearly all custom Agilent made parts in there 

In old 3456/3457, it was only the reference resistor/voltage divider arrays that could go bad, but I never had one go bad. Also not too difficult to replace with discrete components. All digital stuff is complete off the shelf.

Here, basically 3 custom HP parts do all the magic. Repair nightmare, without parts unit 

[HighVoltage]

What is a fair price for a used/may need some repair 34410A? They do seem a bit nicer than the 34401A's...

Originally it was Euro 450 for one unit that was supposedly working perfectly and a free broken unit on top.
When both arrived from the seller and both were kind of dead and one had missing parts, I got a good refund.
Now I have two broken units for a total of Euro 250
It will be a total waste if they will not come back to life
It will be a steal if only one will work at the end

Good units go for around Euro 600 these days, so about twice as much as a 34401A.

[TheSteve]

I'm looking forward to seeing how this progresses. With TiN no longer working daily on his 3458A we need a good DMM repair thread

[Macbeth]

Here, basically 3 custom HP parts do all the magic. Repair nightmare, without parts unit 

Only if unobtanium, much like TiNs 3458A ASICs. Or else deliberately priced way beyond reasonable.

This "throw away" thing is something I really hate. Ok, when a replacement consumer TV costs less than a service engineers hourly wages then fair enough, though in my own case when my cheap shit fails I still persist on fixing it. Professionally I got out of that game decades ago and switched to IT.

But expensive test equipment? Or as Louis Rossman fixes, expensive Apple consumer equipment. There is no reason it should be beyond economical repair. Yes, component level repair not board swapping.

Go back to the 80's and earlier and you would find a service manual for everything. Really nice ones too - fold out sheets, transparent overlays, fully indexed. We don't need that anymore - we have the web. There is literally no cost in publishing this stuff anymore. Just a grim culture of keeping hold of it all.

If the RoHS and global warming advocates want to do something then they should force manufacturers to release service data and provide parts or alternative part numbers. Much better than filling up landfill. Bah! 

[lukier]

nice thread, i have been trying to find 34401a schematic (n 3457a, but no luck)
i see it has 2 ADC, does it mean it share from 1 multislope generator? could be the multislope going bonkers?

I believe 34410A/11A use Multi-Slope IV ADC, the same as the latest generation of 3446xA/34470A. From what I understand the 12 bit ADC (called FINE ADC) is used in a similar manner to the previous generation (34401A) - to read the residual charge at the end of the conversion. The much faster 8 bit ADC (COARSE ADC) is used instead of a zero crossing comparator. Maybe this provides more accurate zero crossing detection.

There are also some differences in the current steering part, instead of traditional resistor network and switches for various slopes it seems to use some form of PWM DAC (AD_PWM to DAM_OUT signal) that generates current for the integrator.

That's what I thought at least after quick look at the schematics and Dave's 34461A teardown. It would be great if someone with much more ADC expertise could shed more light on the Multi-Slope IV architecture, what benefits the 8 bit ADC gives over comparator, is PWM DAC better in terms of injected noise than resistors + analog switches, how good this ADC scheme is and how far from 3458A INL/DNL?

[Macbeth]

yea i like em old schematics, some are so friendly with test voltage marks all over the circuit diagram

Indeed. But a modern young pup fixing old stuff and relying on those test point volts has to bear in mind they were taken with a 20Kohm/volt meter like an AVO 8, even well into the early '90s when we would all be DMMs with 10Mohm impedance. 

[bingo600]

I believe 34410A/11A use Multi-Slope IV ADC, the same as the latest generation of 3446xA/34470A. From what I understand the 12 bit ADC (called FINE ADC) is used in a similar manner to the previous generation (34401A) - to read the residual charge at the end of the conversion. The much faster 8 bit ADC (COARSE ADC) is used instead of a zero crossing comparator. Maybe this provides more accurate zero crossing detection.

There are also some differences in the current steering part, instead of traditional resistor network and switches for various slopes it seems to use some form of PWM DAC (AD_PWM to DAM_OUT signal) that generates current for the integrator.

That's what I thought at least after quick look at the schematics and Dave's 34461A teardown. It would be great if someone with much more ADC expertise could shed more light on the Multi-Slope IV architecture, what benefits the 8 bit ADC gives over comparator, is PWM DAC better in terms of injected noise than resistors + analog switches, how good this ADC scheme is and how far from 3458A INL/DNL?

There's some Multislope info in this HP Journal pg. 17 ,  (it's for the 3457A , so it might be from an older type ADC)
http://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1986-02.pdf

/Bingo

[Gyro]

The displays where updating with very unstable readings but even the one without the relay and missing switch had some numbers on the display.
They just did not make any sense.
For instance, in VDC mode it showed values around 1 VDC but constantly changing

Couldn't that just be it operating with a floating input? Particularly the one with the missing front-back switch and relay.

[plesa]

Also, why is R3121 not populated?
May be only in the 34411A?

R3121 is also not populated in 34411A. Your black PCB looks to me like dark green
And the picture are borrowed from internet, because I do not want to remove boards from chassis.
Pictures are from this site, where you can find internals of L4411A, which is different layout of 34411A.
http://letsdisassembleit.com/news.php?extend.4

Also the 34410A is there
http://letsdisassembleit.com/news.php?extend.2

The price of used 34410A should be between 400-600 USD. Not working 100-150 USD.
I posted few days ago big auction where was lot o K2001 and 34410A for such price.
https://www.eevblog.com/forum/buysellwanted/lot-of-keithley-2001-agilent-34410a-and-other-equipment/msg818608/#msg818608

Today I'm going to check the debug output in 4411A to check the differences between 34410A and 34411A.

[PA0PBZ]

Pictures are from this site, where you can ound the Internals o L4411A, which is dierent layout o 34411A.

You should really fix your 'F' key 

[HighVoltage]

The main board is clean now and all the sticky stuff has been removed by the alcohol washing.
This is the board with the switch and relay in place.

I measured all the gold switch elements and find measurements of 3 mOhm to 9 mOhm on all contacts
I would assume that this is OK.

Next I will take the display board apart, since it is also full of dirt and sticky stuff and some dirt on the rubber buttons are so sticky that they only come out slowly.

[HighVoltage]

OK, display PCB and assembly are all cleaned now.
Also a picture of the dirty fan and rust / corrosion where the fan was mounted.
This sticky stuff was everywhere inside the instrument in a thin layer.

Now its time to re-assemble the meter again.

[plesa]

Pictures are from this site, where you can ound the Internals o L4411A, which is dierent layout o 34411A.

You should really fix your 'F' key 

Done, it required bit of cleaning

OK, display PCB and assembly are all cleaned now.

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.

[HighVoltage]

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.