Repair worklog : Old rusty HP 3458A

[Macbeth]

Turn it on its side - it will balance out! 

[TiN]

Perhaps in other life...

[Kleinstein]

Though the temperature does not change that much, there seems to be a correlation from the drift of the 3458 reading and temperature. It might be worth doing a test with intentional larger change in temperature if the drift is really a temperaure effect.

Modereate heating (e.g. 40-50 C) of the ASIC could be an option to see if this temperature is the critical one. It might even reduce leakage from surface water films on contaminated ceramics.

[VK5RC]

TiN, A really dumb thought, please tell me if you have checked this out before, the method of 'dual / multislope' ADC relies on a good frequency source, the variability/drift changing with position made me think about crystals, how their resonant frequency changes with position. It couldn't be  bad crystal could it???

[lukier]

I've mentioned that a month ago in post #423, so probably it was ruled out.

Although, it would be quite funny if the drift can be traced to the crystal. It should be ratiometric in principle, but imagine that the physical time for run-up and run-down (not the crystal oscillation count) was different between reference calibration and when taking measurements.

[ManateeMafia]

Don't give TiN any ideas. Next thing you know, he will have it Rubidium powered. Atomic 3458A.

[plesa]

Don't give TiN any ideas. Next thing you know, he will have it Rubidium powered. Atomic 3458A.

LOOL, but if I remember it correctly it was TiN original idea to make KX Calibrator.
So it will be also frequency standard and not only voltage/current/resistance
Fingers crossed for 3458A repair.
Another one  on ebay for low price http://www.ebay.com/itm/141862586249

[VK5RC]

@lukier, thanks, missed that one!

[TiN]

TiN, A really dumb thought, please tell me if you have checked this out before, the method of 'dual / multislope' ADC relies on a good frequency source, the variability/drift changing with position made me think about crystals, how their resonant frequency changes with position. It couldn't be  bad crystal could it???

Tried that long ago

[cncjerry]

By the way, there is a nice looking 3458a for sale in the San Francisco Bay Area on craigslist. Not mine, of course. Opening price was $2,900 and the person dropped it to $1,950 last night.  Says it passes self test, all the buttons work .  I was kicking it around but I have way, way too many meters already.

[TiN]

Get it, you will never stop to amaze how much time and effort this beast can steal from you .
If you get lucky, than you may end up with working reference meter, which you can calibrate all your gear to.
If not, we will see 30 pages of debug and repair, like above 

[cncjerry]

Get it, you will never stop to amaze how much time and effort this beast can steal from you .
If you get lucky, than you may end up with working reference meter, which you can calibrate all your gear to.
If not, we will see 30 pages of debug and repair, like above 

This is/was my concern.  I was seriously thinking about it but then I just spent about 20hrs debugging one of my 3457a meters.  When I read thru the theory of ops on the 3457a, I quickly realized that subtle problems (let alone the drift as illustrated in the 3458a above) could be a real bear to fix, if ever.   The 3457a has separate circuitry to eliminate the effect of 10pf of stray capacitance in the input amplifier, a separate offset DAC to control offset voltages throughout the meter, etc.  My meter calibrated perfectly, showed no errors until I hit autocal, hell, I didn’t even know it was broken! But didn't someone say for like $2300 Agilent would take your old3458a, fix, calibrate and send it back almost like new?  So for $4,300 you would have a 20yr old meter that could break again at any time requiring another $2,300.  I don't know, I guess these meters are out of my league.

[TheSteve]

Kind of seems like they are showing there age as well - after 20 years stuff drifts. There are an awful lot of broken ones for sale these days.

[TiN]

I'm sorta disagree at last point. 20 years stuff is already drifted most of it's stress out, so it is _usual_ to have older meters being more stable. Of course that does not apply to parts like capacitors (which degrade with age) or connectors/keys/displays, but real important stuff in 3458 and other hi-end meters usually properly guarded or designed in a way to minimize those drifts. Caps are easy to replace, so with bit of knowledge, it's possible to get good device, given all important stuff is good. One of latest service notes for 3458A actually does imply that newer meters had excessive drift issue, as newer parts may be not as tightly manufactured as older ones.

In any case you will learn a hell lot about these sensitive designs in case  your meter go wank or need repair. Even with all shenanigans described above, I learned more in these 200 hours spent than few years messing with Keithley 2001's (which are meter of itself already to start with). This also goes far beyond repair itself, but also how to produce stable enough signals and references to test it, how to maintain proper wiring and connections, how change in ambient temperature affects parts and so on, it goes and goes. I don't regret on this project, even though it makes me angry half of time I touch it. It's worth those $$ spent.

Also I don't think it's fair idea to send self-repaired and bodged up units to Agilent to repair. As they cannot guarantee by any means that parts you changed by yourself will meet specs/reliability requirements they have for it, so it may just be as simple as replace all affected PCBA's, which would cost way more than calibration fee. That will also prevent you from going into unit again, which sorta kills lot of learning right here. So, no, I'll not send mine to official calibration, it's not worth that. Would be more interesting (and way more useful!) to build and calibrate 10V/10k references and make generator suitable for SCAL.

[TheSteve]

TiN - I understand what you are saying, many older parts do drift less as they age. But as you say not all parts age gracefully, especially capacitors and such. For a project the 3458A may offer a great learning opportunity but it comes at a pretty high $ cost. It may also be the ASIC in them that is failing. I am sure HP/Agilent sold quite a few 3458's but the number parts units on ebay still concerns me. If I had no budget I'd certainly have one or more of them

[TiN]

What you say equally applies for any 7.5 and 8.5 meter. And I agree, these are not for faint heart. But benefit of old meters is that lot of voltnuts have them (and can help on cross testing, learning, etc). It's old school gear, meaning no BGAs, no linux, no megabytes of firmwares. Yes, there are few ASICs, but it's still workable in a way. Also service data and schematics are big plus. Take modern 34470 or 7510, if one get faulty, repair wouldn't be cheaper or easier.

And lets face, this is hobby for fun and selfeducation. If I really needed ppm-level measurements, than even my drifty ASIC would be lesser evil, as it have constant rate and can be compensated by post processing. Some people spent funds on hiend DMMs, others collect scopes, thirds buy Caesium standards, and some just lay on the beach for weeks

P.s. why i sound like thats me who buying 3458A?

[nidlaX]

http://www.ebay.com/itm/HP-Agilent-3458A-Parts-8-1-2-Digital-Multimeter-Parts-Unit-AS-IS-/141869789909

Wow! They tried to get away with selling this shit using just a stock photo? That's unbelievable!

[Jwalling]

http://www.ebay.com/itm/HP-Agilent-3458A-Parts-8-1-2-Digital-Multimeter-Parts-Unit-AS-IS-/141869789909

Wow! They tried to get away with selling this shit using just a stock photo? That's unbelievable!

Gotta love that one! "The item may have some signs of cosmetic wear, but is fully operational and functions as intended." countered by "**Parts Unit/AS-IS, note units are not complete before ordering**" and then "Guaranteed Working" 

Make up your mind, you moron! I hate sellers like this...

Jay

[alanambrose]

Can't believe there are obviously boards and other components missing - in fact the cases could be pretty much empty! Would be much better if they opened up the cases to show what they are really selling. Maybe these guys are dead lazy or maybe they don't actually want you to see what's in the box.

I'm constantly amused by a regular eBay ad for a Mont Blanc pen BOX - which is on for the same amount that I paid for the new pen. Also the ads for stuff which I can get much cheaper from Farnell!

Alan

[TiN]

That's part of their marketing plan to keep your attention crisp, eh?
I think bottom one does not have front panel PCBA either...

[cncjerry]

I think the units on craigslist are sketchier.  People have a little recourse on ebay whereas on craigslist pretty much all sales are final. I've seen items pilfered from work (Advantest spec analyzer, Tek tds744a) and serial numbers scratched off (tektronix 3054b), etc on craigslist.  Also, it looks like someone bought the 3458a locally as it is off the list.

[alanambrose]

Hey TiN,

I've been doing some cals of 3245A and K2000 using a 3458A which I believe has a good cal and also a cal on a 3458A that seems to have a dodgy cal even though its SRAM was fine. That led me to observe the drift on those items in an informal way. My observation was that the 3245A / K2000 combination seemed rock solid with only 1 digit variation over a day and a fair few degrees of temperature change. The two 3458As seemed noticeably more drifty with temperature (maybe twice as much) even if you only observed 6.5 digits. So I'm wondering whether you're just seeing normal temperature drift on your unit?

I'm musing over some tests which several people could run in parallel where a few units are tracked over a 24h period together with internal temperature, ambient and maybe the temperature of the key parts ref, asic etc (using a couple of thermocouples and a K2000 for instance). With some deliberate temperature variations thrown in that would allow some measures of correlation between ambient/internal temperature and reading to compare against the specs and also a sense of how much temperature variation each component exhibits as a function of ambient. We would need a fairly similar testing set-up so the results could be compared, of course. Not sure what's going on below but this kind of thing with more traces and data allowing some correlation calcs?

Regards, Alan

[Jwalling]

http://www.ebay.com/itm/HP-Agilent-3458A-Parts-8-1-2-Digital-Multimeter-Parts-Unit-AS-IS-/141869789909

Wow! They tried to get away with selling this shit using just a stock photo? That's unbelievable!

Looks like it sold! Although Ebay says it sold for "best offer accepted" which would imply that it sold for less, Watchcount http://www.watchcount.com says it sold for $995 which was the asking price.
Is Watchcount broken? 

EDIT: Now Watchcount has a link to Ebay which is stating that it sold for $456 USD. Weird - it just changed in the last few minutes!

EDIT #2: Aha! the link from Watchcount points to the Ebay UK website! I wonder if you can get the best offer price by going through the UK site for everything?? 

Jay

[alanambrose]

Madness...

BTW the 3458A clearly knows the internal temperature to better than 0.1C (e.g. the numbers saved in the calram) - anyone know how to get a higher resolution value out?

[ManateeMafia]

@Alan

I have been monitoring four different 3458A's recently using a RPi and an Agilent 82357b adapter. The ambient temperature, RH, and pressure have been monitored with one of these ... https://www.adafruit.com/products/2652 Not the most accurate sensor but it is close enough for relative measurements. There has been a correlation with room temp and drift of the readings. TiN posted some results not that long ago but I am not sure if the sensor readings were included. Most of the time the readings did return to normal as room temperature changed. Unfortunately, the RPi seems to want to stop communicating through the USB ports regardless of the python script being restarted and the adapter getting hot plugged. I have some long periods of data, but the most recent meter has been an issue. I have ordered an externally powered usb hub and I hope this will fix it. It will be nice to show how a couple of these meters had excessive drift after being unpowered for a few months and then they seemed to settle down.