Tantalum decoupling caps

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

diggy fresh

Well-known member
Joined
Mar 19, 2014
Messages
107
Location
Montreal, Canada
Hi y'all,
i'm refurbishing a Lexicon Primetime 93 and noticed alot of tantalum caps for decoupling the power rails, analog and digital side.

22uf and 4.7uf tantalums, They mesure all "good" with ESR similar or a tiny bit less ESR than my Panasonic FC electrolytics caps.

I'm a bit nervous about the tantalums because they could fail short and i want to preserve this piece, but i think that in theory, tantalums could be better in that application because of smaller size (less loop area) and better at high frequency, not sure.

Would you let the old tantalum there?
Replace with fresh tantalums ?
Or put a new low esr electrolytic?
Or maybe an electro with film bypass?

Curious about your opinions, i know this has been talked about a bit before but in this context of digital/analog gear and lower noise.
Not about tantalum in the signal path.


edit : I'm leaning toward replacing with low esr electrolytics for piece of mind but i don't know if it could degrade the noise. Not sure.
Thanks!
 
Would you let the old tantalum there?
Replace with fresh tantalums ?
Or put a new low esr electrolytic?
Or maybe an electro with film bypass?

My take - bearing in mind that I've never had a problem with tant caps except where they have been installed with incorrect polarity.
And yes - they do smoke and smell - when they 'go.

So - yes - I'd leave the existing tants in there if they look okay and the circuit is operating as expected.
afaik tants don't have the same "wear out" mechanism as wet electrolytics. I'm interested if anyone knows (rather than just thinks) that this is not true.
Replace ? - shouldn't do any harm. I'd recommend SMT TAJ series.
Film bypass - no (although no harm). Individual devices should, in any case, have local decoupling capacitors. I recommend SMT Ceramic X7R type (assuming typ 47n or 100n) since 'better' ceramic type dielectrics are either unavailable or too large to fit. But any film type should be okay.
 
There are photos of Apogee Ensemble Firewires online, with burnt tantalum caps in the power supply section. One i got my hands on a few years ago (used) had an internally shorted one next to the quad ompamp for the first two inputs. Fortunately in that case, the 10ohm resistor upstream of it saved the board from any physical damage.
 
Tants as decoupling on PS rails are a crap shoot as they get older. I have no explanation why, just observation.

Just a few weeks ago we had one short in one of our Ampex ATR-100s. Those machines were made in the mid 1970s to mid 80's IIRC.

Bri
 
Do you feel lucky? Tants across PS rails have a bad reputation.

Low ESR aluminum electrolytic caps could be a safe substitute.

JR

Maybe I've just been lucky. Or it hasn't caught up with me. Nevertheless I've experienced no problems with TAJ series Tants that I have often designed in around standard voltage regulators.
Modern wet electrolytics can be good. But they are basically going downhill over time. Not sure about tants in that regard tbh.
 
I bought the remains of a Trident fleximix in the mid 90's ,
it only had 4 channel modules and the master section left.
When I powered it up the epoxy dipped tants all over the boards started to fail short .
Its was a total mess , in the end I scrapped it , only taking the Zutt input transformers for use later .
 
If it were mine, and none had failed, I’d probably leave them

If one failed, I’d replace either with a tantalum of good voltage rating (some stock parts are borderline) or a low-ESR aluminum electrolytic

If two or more failed, I’d replace them all.
 
In my case with the ATR-100 there are two identical dipped tants (one per PS rail). While I have that I/O module (the one in the overbridge with the VU meter) taken apart I'm replacing both with a low ESR lytic. I've stocked a few extra lytics just in case another module spits fire.

Bri
 
Tantalum caps can’t handle overvoltage, which is often why they blow catastrophically - so in PS rails and in the at-chip PS filtering positions it’s worth chucking them out and replacing them with low ESR electrolytic caps. If used as coupling or audio filtering then just put in new ones (tants), if they’re a bit off value and you’re concerned that the circuit has been designed with those caps and want to preserve original performance, but my preference would probably be to replace them with low ESR, low self inductance (ESL) electrolytic caps. They can just as easily blow in a coupling scenario with too much signal level for their rated voltage.
 
In my situation with the ATR-100 "smoker" the failure MIGHT be because of an over voltage incident. The cap in question is downstream from a "local regulator" which drops an internal 24V rail down to 15V. IIRC that shorted tant was rated at 25VDC,

I checked the pass transistor and it seems OK. The zener on the base of that regulator seems OK...not open. Thank goodness I have an Unobtanium extender for that IO module so I can take a quick measurement when powering it back on.

The module uses multiple uA318 opamps which are rated at 40V (max) rails, so even if that regulator is rogue, I'll be at 39V rail-rail and not blow up those metal can uA318s. In a perfect world I would pull out the chips to verify the rail voltages, but Ampex soldered all their chips onto the boards....no sockets.

Bri
 
Tantalum MnO2 capacitors have fairly low tolerance to ripple compared to the high ripple current tolerance of Tantalum polymer caps - the MnO2 supplies the oxygen needed for internal combustion but this is not present in the polymer types. Polymer caps have typically very little chance of failure causing ignition, have a lower overvoltage percentage than electrolytic but much higher than Tantalum MnO2, however susceptibility to moisture can cause short circuit failure and they have a higher leakage rate than can electrolytic.
There are hybrid caps which provide the low ESR of solid polymer but also resistance to humidity, self healing like aluminium electrolytic and low leakage and high ripple current tolerance.

https://www.mouser.com/pdfDocs/nichicon-hybrids-vs-polymers-vs-aluminums.pdf
 
Well, in my situation, it's a choice of what will easily fit in a tightly packed PCB (designed in the 1970's).

In a perfect world, I would use a film cap as a replacement. But a 25 uF @ 25V film cap would be half the size of that PC board! <g>

Bri
 
Well, in my situation, it's a choice of what will easily fit in a tightly packed PCB (designed in the 1970's).

In a perfect world, I would use a film cap as a replacement. But a 25 uF @ 25V film cap would be half the size of that PC board! <g>

Bri
Safest best choice is can electrolytic low ESR - when filter caps start to go downhill, they go in a hurry with the Tantalum and degradation of multiple caps in a PS rail can cause the explosive failure of the weakest. Had a piece of gear in for service recently that had a Tantalum on virtually every IC across the power rails and a whole stack went at once - blowing out some print as well - all due to a failed voltage regulator.
 
Safest best choice is can electrolytic low ESR - when filter caps start to go downhill, they go in a hurry with the Tantalum and degradation of multiple caps in a PS rail can cause the explosive failure of the weakest. Had a piece of gear in for service recently that had a Tantalum on virtually every IC across the power rails and a whole stack went at once - blowing out some print as well - all due to a failed voltage regulator.
Amen!

Bri
 
thanks all for sharing.

I'm gonna go with replacing with Panasonic fm/fm for peace of mind, also Newmarket is right there is probably lots of additional ceramics decoupling at the chip.

One of the reason i was curious is i read that tantalum have lower inpedance at higher freqs, and also i just watched a video about the best decoupling for high freqs is affected by the loop area, so smaller caps are suposedly better.

The lexicon Primetime is full of switching logic and ram, so i don't want to be injecting more noise in the psu. But in this case im probably ok with low esr electros.

Very interesting people! thanks
 
With poorly ventilated units, especially in a stacked rack, the withstand voltage of a tantalum capacitor falls exponentially with temperature increase - this added to temperature increase due to ripple current can tip a too low rated capacitor over the edge of failure. High turn-on current, high ripple current, over or reverse voltage can cause breakdown and failure of the junctions inside the capacitor. At higher temperatures and short rise time/higher frequencies the recommendation would be to have the working voltage at 33% of rated.
Some designs simply do not take into account all the possibilities that can cause failure of these devices, in other words the quiet failure of one device in a circuit up or down the chain can lead to explosive failure of a tantalum. If they are rated wisely, don’t overheat and are not subjected to any possibility of reverse voltage or high voltage spikes, they can be reliable, as they don’t degrade over time like wet electrolyte electrolytics.
 
[history] Tantalum capacitors were desirable back in the day because of compact size for capacitance delivered and relatively low HF ESR.

Low ESR aluminum capacitors are a relatively recent development to support DC-DC switching power supplies.

The tantalums were premium priced so I didn't use many (I'm cheap). [ /history]

JR
 

Latest posts

Back
Top