HP34401A: Just bought one. Do I need to change the caps?

[sianturi]

Howdy,

I just bought an old HP34401A multimeter. The selftest passed, and all seems okay.
Do I have to change the electrolytic caps?

Thanks,
Edwin

[SeanB]

Simple answer is YES.

Would you buy a used car, just get in and drive away without inspecting the oil, fuel, tyres and such, expecting it to drive for 100 000km without any looking further?

[HighVoltage]

I have repaired many 34401A meters and have never seen a bad electrolytic cap in one of them.
But I have seen lots of failed Tantalum caps in broken 34401A meters.

On the other side, I have older and newer models in the lab and normally they don't fail. I would not install any new caps, if the meter does not show any problems. This meter does not have the typical electrolytic caps problem.

[sianturi]

Okay thx guys! I will probably open the multimeter to inspect visually the electrolytic caps.

[Kleinstein]

If this meter is not one known for failing because of leaking caps (e.g. like the Keithly 2000), I would not change them. With a good thermal design, godd quality and in a linear supply the caps don't age that much.
Visual inspection should be good enough for the next 10 years.

[bitseeker]

I've owned three 34401A, two were the first version, and caps were fine in all of them despite their age. The ones that I've seen with problems were due to Tantalum caps (as HighVoltage pointed out), VFD drivers (especially in the early version), issues with the current source, bad/corroded front/rear switch, etc.

The 34410A has been a good meter too, although I wish it was fanless like the 34401A.

[HighVoltage]

The 34410A has been a good meter too, although I wish it was fanless like the 34401A.

If you slow the fan of the 34410A down with a series resistor, it is almost as quiet as the 34401A

[bitseeker]

Ah, excellent. I shall give that a try. Thanks, HV!

[RMEEV]

Apologies if thread necropsy such as this is discouraged, but it's exactly my topic.

Just acquired a very, very nice 34401a. It's the nicest meter I have by far in both specs and build, so am planning to make it my main bench meter. Even considering sending it out for cal; the last (recorded in cal memory) cal was in 2002. Before I do, though, I'm debating whether to recap it. My gut says to replace the electrolytic caps, but leave the tantalum. I'm concerned about all the "tantalum caps are BAD!" posts/videos, as well as the short-circuit failure mode of them.

So, I thought I'd resurrect this thread and see if the collective wisdom has changed in the last 5 years. I'm definitely not looking for an excuse to replace those SMD tantalums, but will if it will ensure a long life for this unit.

Thanks!
Robert

[coppercone2]

I only see bad caps on the japanese made HP stuff

the 34401As are fine. I would not bother unless it reads bad ESR (you bother) or your trying to track to squeel or noise. too much work just to replace fine caps for no reason (i.e. I have a 1960's cap with ESR thats just fine)

tektronix however is suspect.

[TheSteve]

It wouldn't hurt to replace them due to age. The don't work very hard though as the power supply isn't of the switching variety so it's far from mandatory.

[coppercone2]

I have a 34401A that was on constantly for maybe a decade or two and there are no problems with it other then the display (was part of a physics setup for logging something). (they got it when it was state of the art and replaced it with a 7.5 digit meter that fits the same form factor IIRC).

Installed in lab ~1992-1998 (unknown, probobly early on given how expensive everything else there was, it had to be gotten when it was state of the art for the size) and acquired by me on 2013 and used till 2021 and.. its chugging just fine. did not even think about replacing caps, but maybe I should

[sahko123]

id say if it operates normally there's no need if its getting hot when operating then maybe change them preventatively, otherwise should be alright.

[coppercone2]

yeah I think some of the dead ones have to do with it being wedged in between a cube of linear supplies like a hot dog in a bun

[tkamiya]

My simple policy is "do no harm."  If there is a leakage or other obvious signs of degradation such as domed top or something, I'd leave it alone.  I went through quite a few dozens of test equipment, HP34401A included.  It's actually quite rare for me to find damaged caps.  I see BBQ'd registers and burned PCB far more than that.

PCB on most of recent test gear contains 6 layers or more PCB.  "Just" trying to remove and replace a cap can end up far more task intensive than you've ever planned on doing.

[IDontKnow]

While I totally agree not to touch running systems more than necessary: My one and only HP mixed signal oscilloscope (from the late 90s) just broke due to a single failing tantalum capacitor - the same variety that can be found in the HP34401A. Very annoying. Very frustrating when it happens.

So I will swap the Tantals in the HP34401A.

I had failing Tantals (burning or blowing up) in other test gear from the 80s and 90s occasionally. I agree that they fail much less than some think - but they do sometimes fail. In my W&G signal generator each and every tantalum was secured by a small series resistor (100R) that blew with the tantalum.
And three tantalums blew there up to know.

Not a bad idea to change them every 30 years or so.  Especially in the case of HP34401A they are super easy to change.

My first post here. Very nice content and company on this web site, I have to say.

Gerd.

[coromonadalix]

had 3x 34401a,  i measured the supply lines for any problems

only one who had an very dim vfd, i recapped this one,  removed chemicals and tants,  redone completly, new clone vfd,   sent to cal lab with flying colors, still in specs

others 2  where fine, EDIT  oh yeah  one had a defective x-former (wrong voltage put on it)  and in specs too,   no recapping

some say  don't change, some say change them,  i would say :  evaluate and test before ...  if the supply lines are noisy ....

if you are not well equipped, or have experience with smd   be careful, damage can be done pretty easily  ...

[iMo]

Mine is 1999, US made, I replaced the original 4 big elyts this spring, the old ones did not leak and afaik were still good.
IanJ (IanScottJohnston) advices in his great YT repair vids to replace the 4x or 5x 22u/25V yellow tantalum bricks with 22u/35V ones as the original ones are at 15V "too close" to 25V considering their age.. (..and showing in one of his 34401A repair videos one of the tantalum bricks smoking nicely)..

[Cyclotron]

I have repaired two of my acquired 34401a meters by replacing CAPs. They were both from the 1995-1999 range as I recall.  Neither had signs of leaking but I suspected each had a dry cap. Honestly, I'd never seen caps that appeared dry without leakage till I found these two.

[chilternview]

I'm a firm believer of 'if it ain't broke, don't fix it'.

34401A's hold their cal and last for years without any special attention, as long as they are not fiddled with. Replacing caps just because something is old is like saying everyone over 60 should have a heart transplant 'just in case'.

[iMo]

Replacing the caps in power supplies will not hurt the meter's cal.

What WILL do is a replacement of the zener diodes in the ADC power lines (they drop 18.5V->15V with serial zeners), or any messing/soldering in the ADC circuitry..
There are 3 pairs (6pcs.) of such zeners on various places in the ADC, afaik.

The zeners go short when you short the +/-15V lines to ground, and that happens also when the tantalum 22uF/25V caps go short.
After replacement of the tantalums doublecheck the zeners firts, otherwise you will get + and/or -18.5V to the ADC and it may smoke the opamps, etc.

[factory]

Several of the tants did fail on my early 90s 34401A, two of the Zeners were damaged as a result, thankfully I spotted the AC & frequency measurement wasn't working before the PCB got burnt, the self test will identify these failures too, if you mostly use it for DC measurements.

After this I did replace all the tantalum caps & the electrolytics on the mainboard, leaving 30 year old radial caps to potentially pee electrolyte over the board, is not a risk I'm prepared to take (having changed many in the non TE related day job & other TE I own that had caused corrosion). They hadn't leaked in mine, YMMV.

Parts I used to replace the tants were from the Kemet (now Yageo) T489 low leakage range.
6x 22uF 35V (was 20V across 15V rails), T489D226K035ATE300
1x 4.7uF 25V (was 10V across 5V), T489B475K025ATE1K0
2x 47uF 25V (was 10V across 5V), T489D476K025ATE250
3x 1uF 35V (same rating, 18V rails), T489B105K035ATE2K0
Zeners I used were Onsemi BZX84B3V3LT1G

I also bought 3x more of the 1uF 35V, for the display board but decided against replacing them, due to them being under the VFD, didn't want to risk damaging it, there are up to 3 used depending on the age/revision.

After this the board got cleaned with flux remover & IPA several times.

There is a later (1999) 34401A here, I haven't replaced anything in that one, but ordered enough tants to keep a spare set ready, in case that fails in the future.

David