Checking tantalum capacitors
 

General question for the team. Is an ESR meter any good for checking tantalum electrolytics? My experience of the dreaded tants is that they are either good or bad and fail either short or explode. I don't think they can fail in the same way that aluminium electrolytics do and develop a high ESR.

Ref the failure mode suggested above, I have only found shorted tants or tants with the top blown off....or just two wires where the bead was and a smoky residue. Never an O/C one or one that seems to have gone low capacitance/high ESR.

Re: Checking tantalum capacitors
 

I have an HP 8566B Spectrum analyser on the to do pile. There is a power driver board which pushes appreciable currents into the tuning coils of the main YIG oscillator. The board contains about a dozen tants and is in a hot part of the instrument. They all read high on ESR, half of them read very high to the point where they might as well not be there. The driver misbehaves with overshoots and rings which produce PLL unlock warnings.

These are premium grade metal cased wet tants. for one of the values, Farnell want £80 each. I may need to get inventive!

So yes, tants can go high ESR

David

Re: Checking tantalum capacitors
 

Yes, tantalum capacitors can go high ESR & as already commented on they can explode, catch fire etc. Like sheep they know a thousand ways to die!
Where I've worked in the past tantalum's have caused all sorts of problems when incorrectly spec'd.
An interesting paper into the failure modes & causes can be found here:

https://sh.kemet.com/Lists/Technical...aid%20Qazi.pdf

Re: Checking tantalum capacitors
 

I don't even bother to measure them anymore, just change them- they are designed by the Devil himself

Re: Checking tantalum capacitors
 

I have always either found them to be short circuit or open circuit with just a pair of leads and a patch of soot around where they used to be.

Re: Checking tantalum capacitors
 

As Beardyman pointed out in post #3 correct specification seems to be key.

I used to produce a small series of boards for emergency service control rooms and always used 1) components from reputable manufacturers and 2) a large safety margin regarding operating voltage vs component rating.

The only failures I had was when the apprentice decided polarity didn't matter, somehow one of these boards slipped through quality control and failed with spectacular results.

Re: Checking tantalum capacitors
 

I have also seen another spectacular Tantalum failure. There was a control module, that was being tested, in the early days of EMC testing. Before regulation, but the OEM was concerned about possible malfunction caused by external RF on the behaviour of bought in modules. They introduced a requirement for immunity testing. On the whole 'thing' it was tested with EM fields, but as a pretest individual modules were tested with bulk current injection (BCI). Essentially an RF rated transformer had the wiring to the module under test fed through it, with a relevant signal generator and power amplifier, which induced RF current in the wiring. The poor guy doing the testing came running out of the screened chamber followed by a large cloud of horrible acrid smoke when a Tant let go. That's when I first started to dislike them.

Then I bought a Solartron bench DMM that didn't work when it arrived. The seller assured me it was in GWO when stored away some years before, but multiple Tants were SC. Replacing them means it now functions, but most of the ranges are quite poor on accuracy, I suspect due to more dodgy caps. I do have others of the same model that work perfectly, but I'm half expecting that the next time I switch on, they will be dead.

But the final straw is working on some Marconi instruments. The counter timer wasn't so bad, only took a couple of days to replace on sight. But the signal generator has so many of them buried deep and in boxes under other components I've not got halfway, and keep pushing it to the back of the bench!

Re: Checking tantalum capacitors
 

Tantalums generally never had a particularly low-ESR to start with. I seem to recall about 10R for a 10uF - Although not too much different from small electrolytics around that value (and ESR meters were meant to be used with an appreciation that larger capacitors usually had much lower-ESR). Although I do recall there were some Surface-Mount lower-ESR ones, with a marking code on the side to indicate this.

So Tants were never really suitable for use on a high ripple-current power rail by themselves - Although might get away with many in parallel across a board, to share-out the ripple.

And there was a recommendation to put 10R in series with them to limit surge currents (or maybe stop a s/c one pulling down a power rail, as Sprague made MilSpec? ones with built in fuses)

Unlike Electrolytics, they shouldn't dry-up with heat / time, and probably don't usually require reforming. But it's possible that too much voltage / ripple-current causes partial damage / the ESR to go higher, which then causes them to get hot and eventually fail s/c or go bang! - Although when you read how Tantalum bead ones are made, you wonder how they last that long with so little effective insulating layer.

I've only ever found ones that have gone s/c / have exploded (or even caught fire!), so it was obvious they'd failed, and never had the need to remove them to measure ESR or capacitance.
If they were just used for audio-coupling, on a battery supply, I tended to just use small ceramic capacitors of the same value (Although not common , with move to SM, you can now get at least 22uF in 0.2" spacing, but maybe fairly-low voltages like 6.3V or 10V etc. But may need to ensure that much-lower ESR on these doesn't cause any instability in the original circuit if it has voltage-regulators whose feedback systems are conditionally stable)


And some types are even less reliable than others - I've got a Racal 1792 Receiver that worked when I bought it but when I tried it again after a few years I found quite a few of the notorious blue Thompson ones it has had failed s/c. So I've been collecting quite a few of the main 4 different values, to just do a blanket replacement with some yellow AVX etc. ones which don't generally seem as bad.
I also had quite a few of the blue ones fail in a '007' CB microphone in the 80's, so there shouldn't have been any high ripple-current with its 7.5V? watch-cells battery (although maybe that had got fitted reversed at some point)

Re: Checking tantalum capacitors
 

Back in the early 80's when I started work designing digital electronics, the datasheet for bead tants used in power supply applications were 6V on 5V rails, and 16V on 15V rails.

And that is what Tektronix did in test equipment of that era too.

The result after a passage of decades is that they go short circuit, with different effects depending on the stored energy in the supply. They can just go short and the power supply overcurrents quickly and that is the end of it. Or flames and explosion.

If any of my designs are still used out there, I fully expect mine to suffer the same fate.

Interestingly, in space applications hermetically sealed wet tants are used, never ever bead. The requirement is that the voltage margin should be at least 50%. So >7.5V on a 5V rail and >22V on a 15V rail. Price for space qualified wet tants? Think of a number and multiply by 100 (or more). As an example a gull-wing package FPGA was £35k each.

Craig

Re: Checking tantalum capacitors
 

I remember I had about 3 or 4 module repairs of an intermittent nature of synth going unlocked. This was several years or so from entering service.and over a period of several years as well.. Once I had found the first cause, the next similar fault was soon found as I remembered what it was and then made a note in my service manual. Each one was the same. A tant of maybe 10uf that was decoupling a supply of a few milliamps fed through a resistor. The cause was a tant that was inserted wrong way round! Not enough current could flow to cause any physical damage, but sufficient to kill the synth lock. Switching off a short while and the fault would go away.
Rob

Re: Checking tantalum capacitors
 

The polarity marking on some beads seems to be counter-intuitive with the + side indicated by a vertical line, this has often led to confusion and occasionally to beads being inserted the wrong way round.

Re: Checking tantalum capacitors
 

The main question now I suppose is, do tants have to be replaced by tants? It must depend on where they are in a particular circuit. Normal decoupling it probably doesn't matter and if space allows, a suitable aluminium type will be OK but are there particular instances where only a tant is suitable? (talking normal leaded applications, not surface mount of course)

Re: Checking tantalum capacitors
 

Does any OEM still fit tants in 2021? Methinks not. Modern 'wet' caps have come a long way. I have seen tants used in timing circuits, but considering how good film caps are now, that's unnecessary.

Once a tant does become incandescent, it can take a lot to put it out. It's a serious fire risk. In fact, inside an enclosure is probably the safest place for it. If one blows up in an exposed place outside of an enclosure, it can turn into a floating fireball, flying across the room and potentially igniting the area it lands on.

Re: Checking tantalum capacitors
 

Ceramic multilayer caps (MLC) can also evaporate the PCB when they fail short circuit! I don't see manufactures stopping their use any time soon.

I generally find TANT much more reliable than AL electrolytic but with the potential to catch fire where there is enough energy to do so. I might consider replacing them with low ESR AL where there could be a lot of energy, if there is enough space, otherwise fit an increased voltage TANT.

I've seen a few situations where it might be difficult to match the low leakage and low ESR of a TANT.

dc

Re: Checking tantalum capacitors
 

Yes, but there have been many things done in design or process to prevent or at least reduce the risk. Putting two in series, at right angles. Doing stress measurements on the board throughout the process (not every board, but a trial board). Having special components with floating terminations to prevent the cracks forming.

Re: Checking tantalum capacitors
 

It was interesting, but only to a point. It described how to do failure analysis, get the failed component on the board to be sure it's been fitted right-way-round; sawing open, etc. But it didn't mention degredation with time, for instance (some of the posts on here describe tants found dead after storage).

I've specified 3.3μF 25V tants on 15V supplies to op-amps (the tants between supply pins and ground, thus seeing 15V, and then 33Ω to the actual 15V power rails) and never had a problem. And I have hundreds out there in the field... Dubilier tants, changing to AVX when Dubilier disappeared.

Re: Checking tantalum capacitors
 

Tantalums do go leaky - they're a problem familiar to owners of Clansman military-radios; in particular there are some small high-voltage Tantalums in the synthesizer power-supply [module 5 on the PRC320] which go electrically-leaky which causes the 'stabilized' 120V VCO-supply to wander up and down by a few hundred millivolts - which in turn puts a 'wobble' - typically of several hundred Hertz - on the VCO output and makes SSB operation gruesome-to-impossible.

Identical original-spec Tantalums - they're "wet" types in an aluminium can with a red epoxy-resin seal where the +ve lead comes out - are available - complete with NATO stock-number - but at a price [around £80+VAT!] which is just silly - I fit Panasonic 105C electrolytics which are available from RS etc at a far less wallet-frightening price.

The more-usual orange- or blue-blob Tants just tend to go BANG! and leave a nasty smell/stain as evidence of their demise. I've recently done a complete replacement of these capacitors in my AEL3030 transceiver after they started going pop.

Re: Checking tantalum capacitors
 

Going leaky, and also high-ESR, are failure modes I don't doubt - as reported in this thread - but what is going on inside?

Aluminium is straightforward: going leaky is slow degradation of the oxide film due to chemical action while unpowered; high ESR is drying-out (or occasionally, failure of a crimped or welded internal connection). But solid tantalum...?

(And yes, I've come across wet tants too, axial-leaded 125μF, 150V tantalum types, used as reservoir capacitors following a 400Hz power transformer and bridge rectifier, rather worryingly given above posts. Small, heavy, and £66 each...)

Re: Checking tantalum capacitors
 

Seen a few types of wet tantalum capacitors leak corrosive electrolyte (sulphuric acid based) including Sprague 109D (used by hp a lot in 60's & 70's TE) and the ones with red seal (forget the brand), often the components around show signs of corrosion too, wiki seem to suggest the casing is solid tantalum on later ones which may explain the expensive price.

These came with a warning notice;
Attachment 234130

Some solid tants are disguised in plastic cases, we had a occasional failures with short circuits as with the bead type and excessive leakage in one timing application resulting in the product not going into standby after a period of no activity.
Attachment 234131

And finally a Racal 836 counter I repaired some time ago failed again in the same place, should have worked out something was arcing over before the smoke release as the shed radio was picking up the interference on MW :dunce:, I've used higher voltage types this time (25V instead of 16V), no damage apart from the cap as Racal fused all the PSU rails.
Attachment 234132 Attachment 234133 Attachment 234134

David

Re: Checking tantalum capacitors
 

Apple Macbooks probably still use them, or at least a few years ago. Other laptops and mobile devices likely as well.