开云体育

ctrl + shift + ? for shortcuts
© 2025 开云体育
  1. HP-Agilent-Keysight-Equipment
  2. Topics

HP 3403C True RMS Voltmeter display fault

 

On 3/3/25 12:41, ed breya wrote:
That's an interesting picture of the failed cap. I'd guess the molten blob on the can end is just some electrical solder that ended up there by coincidence sometime in the past, or maybe from inside the part. Try hitting it with a soldering iron - it should melt. If it doesn't, then I guess it could be Ta.
Metallic tantalum does not normally look like that, all silvery and shiny, because oxides form on the surface with a range of iridescent colors. It's more of a refractory metal that would not exist molten at a temperature that wouldn't have turned the rest of the part to ash. Besides that, the "solid" Ta caps are mostly Ta oxides and very little metal, sintered into solid slugs, and I can't picture there being that much actual Ta metal present (in proportion) inside a cap of that size. There could have been that much electrical solder in there though. It's used to make internal contacts, and especially to solder the electrical leads and glass/metal hermetic sealing disc into the can.
Ordinarily I would agree, but I've seen this "blob of metal on the side of a dipped Ta capacitor" failure many times, and it's always shiny and looks a lot like solder.

-Dave

--
Dave McGuire, AK4HZ
New Kensington, PA
 

On 3/2/25 19:05, AArnaud via groups.io wrote:
The other one ooks just as weird but is healthy according to the chinese component tester.
Are there guidelines on HP capacitor replacements? Such as caps that never/very rarely fail and those who absolutely always do?
I want to recap this unit to keep it. It's not calibrated anymore of course but it's so great. Quality long life capacitor are expensive, I don't want to replace what does not need to be.
I usually "replace like with like", or at least a modern equivalent.

You should see all the problems that come up with people with no electronics knowledge on the vintage computing mailing lists going off the deep end with the "recapping" craze, thinking it will cure all ills...frequently they replace electrolytics with high-K ceramic capacitors and wonder why their power supply rails don't work. ;)

-Dave

--
Dave McGuire, AK4HZ
New Kensington, PA
 

I know what you mean Dave. I've seen soldery-looking stuff on some blown dipped Ta caps - the ones that haven't been burnt to a crisp. There is some solder in those too I think, for connection to the outside surface of the slug. As I recall, the Ta oxide slug is sintered around the steel anode lead and directly bonded, while the outside of the slug is coated or plated with solder or some equivalent alloy that can connect to the steel cathode lead, then all dipped in an epoxy.
?
I think SMT ones have some solder on each end of the slug for connection to the lead/pad structures, and have the cap element structure and polarizing features worked out somehow internally.
?
I have some rather odd "leadless" SMT style solid Ta caps with interesting structure. These are orange colored with a thin coating, and no plastic encapsulation, and go directly in-circuit via the solder-tinned ends. One end has a short wire stub poking out, which is the internal anode connection and polarity marker. There's no marking at all on the bodies, so takes some guesswork to figure out. I found datasheets years ago so I could figure out the ratings by part size and measured C. I keep these for ultra-low R and L uses.
?
Ed
 

I recently had one destroy itself on an HP 16717a logic blade. Burned all of the
way through the pcb, great fun. I guess that was early 90s vintage.

Paul

On Mon, Mar 03, 2025 at 09:55:33AM -0800, ed breya wrote:
I forgot to comment on hermetic solid Ta caps. I have seen only a few failures of these over the years, and I consider them among the best of parts for reliability. I have seen lots and lots of failures in all kinds of non-hermetic Ta caps, especially certain brands of dipped style. My experience here is with the older parts in the same vintage as most of my gear and such, so not so much about more modern parts which should be better. For instance, I have seen very few failures in surface mount style solid Ta types that are so common now. These are non-hermetic but seem to survive OK. Or, maybe after these rack up many years they may degrade like the others. Only time will tell.

Ed
--
Paul Amaranth, GCIH | Manchester MI, USA
Aurora Group of Michigan, LLC | Security, Systems & Software
paul@... | Linux/Unix - We don't do windows
 

Hi AArnaud,
?
Congratulations for the successful repair, and for having saved another valuable HP instrument from the scrap heap !
My conclusion is somewhat identical to my initial thoughts : start by checking the "usual suspects", supply voltages and clocks.
The obvious next step for your 3403C is a complete performance check and recalibration, if you have the appropriate equipment. If you're lucky, first adjusting the calibration in DC-only voltage mode may bring the instrument back into full compliance, which of course would be great news.
After that it'll be time for another instrument repair project, and for more fun !
Cheers,
Joe;
 

开云体育

Yes, Ed, I know exactly what you mean regarding the orange leadless tantalum caps. ?Sprague used to make them, and they told us that the orange material prevented them from printing the value, voltage, etc on them. ?A real bummer because many times the wrong voltage rating parts were installed and blew up! ?I stopped designing boards with them for that reason. ?Life’s too short to have exploding caps in your products. ?Too short not to intentionally blow them up on the bench, though! ?LOL!??

Jim Ford?
Laguna Hills, California, USA?


On Mar 3, 2025, at 10:59?AM, Joel Setton via groups.io <setton@...> wrote:

?
Hi AArnaud,
?
Congratulations for the successful repair, and for having saved another valuable HP instrument from the scrap heap !
My conclusion is somewhat identical to my initial thoughts : start by checking the "usual suspects", supply voltages and clocks.
The obvious next step for your 3403C is a complete performance check and recalibration, if you have the appropriate equipment. If you're lucky, first adjusting the calibration in DC-only voltage mode may bring the instrument back into full compliance, which of course would be great news.
After that it'll be time for another instrument repair project, and for more fun !
Cheers,
Joe;
 

开云体育

When the orange “tombstone” plastic encapsulated surface mount tantalum caps fail, they fail shorted.

Last year, while powering up a 9000/715, I saw one actually flame because I had the good fortune to have the case open. Fortunately, damage was minimal and i happened in have a partial real of an exact replacement.

I’ve had other tantalum capacitors fail, orange drop epoxy dipped, actually incandesce and shoot an glowing slug about 5 feet away onto the carpet.
 

Hey everyone,
?
I'll check whether or not the blob will melt. Both capacitors have this blob, although it is more pronunced on the dead one. Considering that we can now see inside the capacitor I think it's indeed solder. It was there when I got it, and the caps date code match the rest of the instrument.?
Is it wise to reuse the old one which is good, but looks justs as weird? I'll post a picture later when on lunch pause.
I have already mapped the original part numbers for the capacitors in the unit based on the manual, I need to check the actual reality of it. I'll post the sheet when I'm done. They're basically all still available new, albeit at a premium. Once again a costly order.
?
After that I'll adjust the voltage again, and try to calibrate it. I do not own a trustable RF source yet as my Marconi 2018A is in repair aswell (and part of the order), but I do own an Adret 104A DC voltage standard which is good enough for decent DC calibration.
It's probably been one or two decades since this thing had someone try to calibrate it. I'm worried about the converter internal state. Testing it, it's of course not accurate but it's not widdly far from it. I've tested it up to about 20MHz. DC is particularly close.
?
Cheers,
Arnaud.
 

Arnaud, you can replace Ta caps with alternatives in most applications, so there's no need to spend too much for an original type. Usually ones that burn out are in supply bypassing spots, which are not critical. The main issue then is space available for new parts, since the common fix is to go with a much larger capacitance (and physical size) aluminum electrolytic type to get ESR comparable to the Ta ones. I'd say put in an Al cap of at least 2-10 times as much C as the original Ta, with enough V rating for the supply voltage. Or, use the biggest C that can physically fit into the space available. You can look at various brands and families of caps to choose the lowest ESR you can find, but they won't be dramatically different.
?
Another option is to go with an OSCON organic dielectric type Al cap. These are built for low ESR and can beat Ta in this respect, so fitting is easier. The problem here is that (I think) OSCONs are only available in radial mount style (can with leads out the bottom), so trickier to fit in axial-lead spots. If axials are available, then no problem.
?
Remember this all only applies to supply bypassing spots. Sometimes Ta is needed for particularly low DC leakage and better stability like in timing circuits. Ones in these kinds of spots hardly ever fail because they're not stressed. Supply bypassing ones fail a lot, but value stability and leakage hardly matter here, so bigger Al types or OSCONs (which have the worst leakage) work just fine to get decent ESR.
?
Ed
?
?
 

I'll add that since, in general, modern caps are smaller than vintage caps, make sure the the voltage rating of the new cap provides sufficient margin over the voltage rail. For instance, Tektronix often used caps rated for +15v on +12v rails. In all the cases where I have recapped equipment, a modern +25v cap was smaller in physical size than the original +15v caps.

DaveD
KC0WJN


On Tue, Mar 4, 2025 at 12:14 ed breya via <eb=[email protected]> wrote:
Arnaud, you can replace Ta caps with alternatives in most applications, so there's no need to spend too much for an original type. Usually ones that burn out are in supply bypassing spots, which are not critical. The main issue then is space available for new parts, since the common fix is to go with a much larger capacitance (and physical size) aluminum electrolytic type to get ESR comparable to the Ta ones. I'd say put in an Al cap of at least 2-10 times as much C as the original Ta, with enough V rating for the supply voltage. Or, use the biggest C that can physically fit into the space available. You can look at various brands and families of caps to choose the lowest ESR you can find, but they won't be dramatically different.
?
Another option is to go with an OSCON organic dielectric type Al cap. These are built for low ESR and can beat Ta in this respect, so fitting is easier. The problem here is that (I think) OSCONs are only available in radial mount style (can with leads out the bottom), so trickier to fit in axial-lead spots. If axials are available, then no problem.
?
Remember this all only applies to supply bypassing spots. Sometimes Ta is needed for particularly low DC leakage and better stability like in timing circuits. Ones in these kinds of spots hardly ever fail because they're not stressed. Supply bypassing ones fail a lot, but value stability and leakage hardly matter here, so bigger Al types or OSCONs (which have the worst leakage) work just fine to get decent ESR.
?
Ed
?
?
 

开云体育

The OSCON type capacitors aren’t available in higher voltage ratings.

From: [email protected] <[email protected]> on behalf of Dave Daniel via groups.io <kc0wjn@...>
Sent: Tuesday, March 4, 2025 11:30:05 AM
To: [email protected] <[email protected]>
Subject: Re: [HP-Agilent-Keysight-equipment] HP 3403C True RMS Voltmeter display fault
?
I'll add that since, in general, modern caps are smaller than vintage caps, make sure the the voltage rating of the new cap provides sufficient margin over the voltage rail. For instance, Tektronix often used caps rated for +15v on +12v rails. In all the cases where I have recapped equipment, a modern +25v cap was smaller in physical size than the original +15v caps.

DaveD
KC0WJN


On Tue, Mar 4, 2025 at 12:14 ed breya via <eb=[email protected]> wrote:
Arnaud, you can replace Ta caps with alternatives in most applications, so there's no need to spend too much for an original type. Usually ones that burn out are in supply bypassing spots, which are not critical. The main issue then is space available for new parts, since the common fix is to go with a much larger capacitance (and physical size) aluminum electrolytic type to get ESR comparable to the Ta ones. I'd say put in an Al cap of at least 2-10 times as much C as the original Ta, with enough V rating for the supply voltage. Or, use the biggest C that can physically fit into the space available. You can look at various brands and families of caps to choose the lowest ESR you can find, but they won't be dramatically different.
?
Another option is to go with an OSCON organic dielectric type Al cap. These are built for low ESR and can beat Ta in this respect, so fitting is easier. The problem here is that (I think) OSCONs are only available in radial mount style (can with leads out the bottom), so trickier to fit in axial-lead spots. If axials are available, then no problem.
?
Remember this all only applies to supply bypassing spots. Sometimes Ta is needed for particularly low DC leakage and better stability like in timing circuits. Ones in these kinds of spots hardly ever fail because they're not stressed. Supply bypassing ones fail a lot, but value stability and leakage hardly matter here, so bigger Al types or OSCONs (which have the worst leakage) work just fine to get decent ESR.
?
Ed
?
?
 

Ok, tgat's good to know.

DaveD
KC0WJN


On Tue, Mar 4, 2025 at 13:12 Ed Marciniak via <ed=[email protected]> wrote:
The OSCON type capacitors aren’t available in higher voltage ratings.
From: [email protected] <[email protected]> on behalf of Dave Daniel via <kc0wjn=[email protected]>
Sent: Tuesday, March 4, 2025 11:30:05 AM
To: [email protected] <[email protected]>
Subject: Re: [HP-Agilent-Keysight-equipment] HP 3403C True RMS Voltmeter display fault
?
I'll add that since, in general, modern caps are smaller than vintage caps, make sure the the voltage rating of the new cap provides sufficient margin over the voltage rail. For instance, Tektronix often used caps rated for +15v on +12v rails. In all the cases where I have recapped equipment, a modern +25v cap was smaller in physical size than the original +15v caps.

DaveD
KC0WJN


On Tue, Mar 4, 2025 at 12:14 ed breya via <eb=[email protected]> wrote:
Arnaud, you can replace Ta caps with alternatives in most applications, so there's no need to spend too much for an original type. Usually ones that burn out are in supply bypassing spots, which are not critical. The main issue then is space available for new parts, since the common fix is to go with a much larger capacitance (and physical size) aluminum electrolytic type to get ESR comparable to the Ta ones. I'd say put in an Al cap of at least 2-10 times as much C as the original Ta, with enough V rating for the supply voltage. Or, use the biggest C that can physically fit into the space available. You can look at various brands and families of caps to choose the lowest ESR you can find, but they won't be dramatically different.
?
Another option is to go with an OSCON organic dielectric type Al cap. These are built for low ESR and can beat Ta in this respect, so fitting is easier. The problem here is that (I think) OSCONs are only available in radial mount style (can with leads out the bottom), so trickier to fit in axial-lead spots. If axials are available, then no problem.
?
Remember this all only applies to supply bypassing spots. Sometimes Ta is needed for particularly low DC leakage and better stability like in timing circuits. Ones in these kinds of spots hardly ever fail because they're not stressed. Supply bypassing ones fail a lot, but value stability and leakage hardly matter here, so bigger Al types or OSCONs (which have the worst leakage) work just fine to get decent ESR.
?
Ed
?
?
 

Hey again,
Here are the photos. The one with the big blob is the blown one. It does melt with the iron so it's indeed solder. Thanks for the info on possible replacements.
?
?
 

开云体育

Yep, I have seen the fireworks in the lab, too. ?Been a while, fortunately. ?More than 25 years ago we had a technician who installed a few tant caps backwards into a VMEbus chassis. ?Blam, molten metal shot across the lab! ?Fortunately nobody was hurt.

Jim Ford?
Laguna Hills, California, USA


On Mar 4, 2025, at 9:30?AM, Dave Daniel via groups.io <kc0wjn@...> wrote:

?
I'll add that since, in general, modern caps are smaller than vintage caps, make sure the the voltage rating of the new cap provides sufficient margin over the voltage rail. For instance, Tektronix often used caps rated for +15v on +12v rails. In all the cases where I have recapped equipment, a modern +25v cap was smaller in physical size than the original +15v caps.

DaveD
KC0WJN


On Tue, Mar 4, 2025 at 12:14 ed breya via <eb=[email protected]> wrote:
Arnaud, you can replace Ta caps with alternatives in most applications, so there's no need to spend too much for an original type. Usually ones that burn out are in supply bypassing spots, which are not critical. The main issue then is space available for new parts, since the common fix is to go with a much larger capacitance (and physical size) aluminum electrolytic type to get ESR comparable to the Ta ones. I'd say put in an Al cap of at least 2-10 times as much C as the original Ta, with enough V rating for the supply voltage. Or, use the biggest C that can physically fit into the space available. You can look at various brands and families of caps to choose the lowest ESR you can find, but they won't be dramatically different.
?
Another option is to go with an OSCON organic dielectric type Al cap. These are built for low ESR and can beat Ta in this respect, so fitting is easier. The problem here is that (I think) OSCONs are only available in radial mount style (can with leads out the bottom), so trickier to fit in axial-lead spots. If axials are available, then no problem.
?
Remember this all only applies to supply bypassing spots. Sometimes Ta is needed for particularly low DC leakage and better stability like in timing circuits. Ones in these kinds of spots hardly ever fail because they're not stressed. Supply bypassing ones fail a lot, but value stability and leakage hardly matter here, so bigger Al types or OSCONs (which have the worst leakage) work just fine to get decent ESR.
?
Ed
?
?
 

开云体育

50+ years ago, 400 Hz power supply plugged into 60 Hz source. Runs for a few minutes and then throws its guts all over the bench. Back before the days of power supplies that work from 48 - 440 Hz, 100 - 240 vac. But you only do it once ?

Steve


On Mar 5, 2025, at 9:13?AM, Jim Ford via groups.io <james.ford@...> wrote:

?Yep, I have seen the fireworks in the lab, too. ?Been a while, fortunately. ?More than 25 years ago we had a technician who installed a few tant caps backwards into a VMEbus chassis. ?Blam, molten metal shot across the lab! ?Fortunately nobody was hurt.

Jim Ford?
Laguna Hills, California, USA


On Mar 4, 2025, at 9:30?AM, Dave Daniel via groups.io <kc0wjn@...> wrote:

?
I'll add that since, in general, modern caps are smaller than vintage caps, make sure the the voltage rating of the new cap provides sufficient margin over the voltage rail. For instance, Tektronix often used caps rated for +15v on +12v rails. In all the cases where I have recapped equipment, a modern +25v cap was smaller in physical size than the original +15v caps.

DaveD
KC0WJN


On Tue, Mar 4, 2025 at 12:14 ed breya via <eb=[email protected]> wrote:
Arnaud, you can replace Ta caps with alternatives in most applications, so there's no need to spend too much for an original type. Usually ones that burn out are in supply bypassing spots, which are not critical. The main issue then is space available for new parts, since the common fix is to go with a much larger capacitance (and physical size) aluminum electrolytic type to get ESR comparable to the Ta ones. I'd say put in an Al cap of at least 2-10 times as much C as the original Ta, with enough V rating for the supply voltage. Or, use the biggest C that can physically fit into the space available. You can look at various brands and families of caps to choose the lowest ESR you can find, but they won't be dramatically different.
?
Another option is to go with an OSCON organic dielectric type Al cap. These are built for low ESR and can beat Ta in this respect, so fitting is easier. The problem here is that (I think) OSCONs are only available in radial mount style (can with leads out the bottom), so trickier to fit in axial-lead spots. If axials are available, then no problem.
?
Remember this all only applies to supply bypassing spots. Sometimes Ta is needed for particularly low DC leakage and better stability like in timing circuits. Ones in these kinds of spots hardly ever fail because they're not stressed. Supply bypassing ones fail a lot, but value stability and leakage hardly matter here, so bigger Al types or OSCONs (which have the worst leakage) work just fine to get decent ESR.
?
Ed
?
?
 

Polymer capacitors are readily available for voltages of 100 and under although modern low ESR electrolytics make perfectly fine replacements for tantalums and are, of course, available in a wider range of voltages..? As for the radial leaded packages very small diameter PTFE tubing is readily available for only a few dollars - dress the leads with it or just put down the appropriate sized pieces of Kapton tape over any traces that may be a shorting hazard.? It will not as rigid of a attachment as a axial package but unless the item will be in a fairly high vibration environment that doesn't matter - and if so, use a dab of the appropriate adhesive.
?
Hal
 

Nah...

3-phase fed 500V raw DC rectifier stage in an 8kW TWTA, PSU let go.

The flames reached the ceiling in the workshop!? While the 32A wall breaker/RCD trip failed to open.

We had a few such events like that.? All were US built high power kit, with VDR's that failed in cascade, running on UK standard 400V 3P&N supplies, that were designed & built for use in the UK.? ("On the edge" rated?)

The high power RCT converter, & the 15kV HV secondary side of the PSU was tame by comparison.

Not fun at the time.

Dave 'KBV.
 

Hey everyone,
?
Having recently picked up an Hakko 474 I've decided to desolder every electrolytics in the PSU to test them before ordering some.
To my surprise they're all well within specs, close to new. I now understand why Sprague is regarded as such a good brand, I'm leaving them in. I'm only replacing the blown tantalum on the A7 board.
?
I'll post another update when some DC calibration will happen.
?
Arnaud.
 

Hey everyone,
?
Just to inform everyone that the 3403C is back together and running happily.
I let it warm up for 3 hours, set up the power supply and used my Adret 104A to setup the DC part. I had to trim the capacitor for the DPM to get back into alignment and that's about it, it's well within spec on the DC aspect and at least up to 100V.
For the AC I do not have a trustable source so I do not know for now, we'll see later.

Thanks everyone!
Arnaud.
 

Nice. I hope mine works as well when I eventually get back to it.? Ed

21 - 40 of 41