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.
If replacing capacitors in any unit that has had an overvoltage failure it is wise to replace with a higher voltage rated unit. Bulging can electros or blown tantalums I usually replace with higher voltage if possible. 10V with 16V, 16V with 25V, 25V with 35V, 35V with 50V, 50V with 63V - these transitions are pretty well straightforward as for can electros there is little or no can size or body size change - if there is then it’s either a radial size change or a vertical height change and one or the other will usually fit. Next and equally important is to replace with highest possible temperature rating. Manufacturers of electrolytic capacitors specify the design lifetime at the maximum rated ambient temperature, usually 105°C. This design lifetime can vary from as little as 1,000 hours to 10,000 hours or more. Each 10°C drop from maximum doubles the life of the capacitor so if a capacitor in a circuit is an 85°C cap then replacing it with a 105°C one has obvious merit. A 10,000hr 105°C cap at 55°C has an expected life of 320,000hrs.
With older gear the newer caps tend to be smaller anyway. With tantalum they tend to be the same size.
The newer aluminium electrolytics from manufacturers such as Nichicon, Panasonic, Kemet, Rubycon,Murata and Vishay all have capacitors with lower ESR than standard through hole tantalums. The hybrid electrolytic caps even lower still. The lower ESR produces less internal heat generated from ripple current or signal passage and thus also prevents heat related electrolyte losses causing gradual failure that the old can electros exhibited and changes to electrolyte types means greater stability over long periods of time - this equates to tens of thousands of hours at max temperature and rated voltage. Best to look at the specs of capacitors before choosing.
Years of reported common failures in audio gear (and other) lead to sensible replacement of this type of component.
Unlike cars they don’t crash test recording consoles and FX gear!
 
bit of discussion and specifics from merlin (Blencowe) in this thread on tants in vintage Neves:

https://groupdiy.com/threads/select...ion-in-audio-applications.78333/#post-1001897
There’s no doubt that there is a sonic difference between the original tantalums and replacing with electrolytic which is why in specific circuits that have been designed with tantalums in the audio path I’ll replace them with tantalums but higher tolerance ones. For power rails I get rid of them.
 
Sonic signature of caps is an old story, I used tants in the 70's for my projects, though they sounded OK. These were replaced with film caps in the 80's, much bigger and awkward looking, but sounded better.
Not always practical.
The capacitor debate has raged for decades. Metalized vs. metal foil, aluminum vs. copper foil, teflon, polystyrene, polypropylene, polyester, paper-in-oil, etc have been contenders.
Nobody argues for ceramics!
Availability and size is an issue, besides cost.
Plan on replacing electrolytics in old equipment. That magic liquid inside them has a tendency of escaping.
 
Tantalium in powerrails need to be replaced like in Lexicon's.

In the signalpad like Neve I would keep them. My conclusion after listenîng tests.
 
Lyrics actually go up in value as they age. (Spell checker can be funny sometimes so I leave this one)

This is because the pressure of the wound assy forces liquid out over time resulting in the plates getting closer together which increases C.

Unfortunately the voltage resist level will drop as the path from one plate to the other is decreased.
 
The capacitor debate has raged for decades. Metalized vs. metal foil, aluminum vs. copper foil, teflon, polystyrene, polypropylene, polyester, paper-in-oil, etc have been contenders.
Nobody argues for ceramics!
Availability and size is an issue, besides cost.
Plan on replacing electrolytics in old equipment. That magic liquid inside them has a tendency of escaping.

wrt ceramic dielectrics. COG/NPO is limited in value Vs size. The debate seems generally to be for values where they are not available or practicable.
 
Lyrics actually go up in value as they age. (Spell checker can be funny sometimes so I leave this one)

This is because the pressure of the wound assy forces liquid out over time resulting in the plates getting closer together which increases C.

Unfortunately the voltage resist level will drop as the path from one plate to the other is decreased.
Yeah - first sign of failing can electrolytic is the rise in value - contrary to popular belief.
 
wrt ceramic dielectrics. COG/NPO is limited in value Vs size. The debate seems generally to be for values where they are not available or practicable.
Pick ceramics for their self resonant frequncy to get lowest impedance decoupling, if you have a specific frequency in mind. Could be an SMPS. Not much needed for audio, can be very microphonic.
Many dielectrics have voltage dependent capacitance, keep a big margin on voltages. High value caps with low voltage rating very risky. Seen shorted new ones recently.
 
Speaking of cap failures , these temperature compensated hi freq dogbone caps that go in the discriminator coil of marantz and other stereos sometimes fail or drift which causes loss of stereo fm and will drive you nuts. They just kind of fall apart mechanically.
 

Attachments

  • Dogbones.jpg
    Dogbones.jpg
    13.5 KB
Tantalums have one advantage over standard electrolytic capacitors. That is very high frequency response. Replacing with conventional capacitors will possibly lead to instability. I remember one case of a power supply on a bit of UREI kit that had been replaced with an electrolytic and the unit screamed at about 1.5Mhz. Replacing the cap with a tant cured the problem. Replace all of them with best quality tantalums that you can find. Expensive but solves all the problems.
 
I had an experience with a Tant that had such low ESR the 10V voltage reg feeding it became unstable, started oscillating and the cap caught fire and became glowing lump. Fun times.
Tantalum is a "conflict material" so I have been avoiding it for the last 15 - 20 years. Nowadays there are more alternatives, ceramics, niobium, organic electrolytes, film caps if you have space, for low ESR.
 
I remember one case of a power supply on a bit of UREI kit that had been replaced with an electrolytic and the unit screamed at about 1.5Mhz

Was the power supply itself supposed to work that high in frequency? If anything, sounds more like the electrolytic might've had too high ESR, and messed up the feedback/compensation of said power supply.
 
Regulated power supplies are "amplifiers" ( with 0V as AC reference) and subject to the same laws as any signal amp. Gain/phase etc is in play.
Have fun!
I made the observation years ago that inexpensive 3-terminal regulators like 78xx and 79xx rely upon negative feedback to deliver low output impedance, while using 741 era process technology.

JR
 

Latest posts

Back
Top