Mixing console PSU recap - What about those tantalums?

[j-fall]

Recapping my TAC 750 PSU and wondered what to do with the tantalum caps. I have 3 options and would love to hear your opinion on them:

    Leaving them alone
    Replacing with new tantalums
    Replacing with electrolytic caps

I've seen a video where the guy (DoItYourselfMusician) opted for option 3 and replaced the tantalums with 10u/50v electrolytic caps. Would you also recommend doing that?

Thanks!

[floobydust]

Tantalum capacitors have a place with their moderate ESR, they are not to be demonized especially in audio gear. For the few dollars I would replace them if you're in there doing a recap. Not sure how old the console is.  There's two? C5, C10? Or also C4, C9? They look like they are properly (voltage) derated 1uF 35V on a 18V rail.
Electrolytics would be used for the 48V phantom power Vreg.

[j-fall]

Thanks floobydust!

That would also be my preferred solution after I did some more research yesterday.

The console is from 1987. C5 & C10 are the tantalums.

[IanJ]

They look like they are properly (voltage) derated 1uF 35V on a 18V rail.

The most important issue with Tants.

Ian.

[fzabkar]

Typical practice in those days was to use a combination of a high capacitance aluminium electrolytic in parallel with a 1uF tantalum cap. The former would do the bulk of the work at low frequencies, and the latter would handle high frequency noise. Personally, I avoid tantalums wherever possible -- I have seen too many catch fire.

[HalFoster]

While over voltage will kill a tantalum, the biggest danger is current.  Tantalum's just don't do high current surges well at all - all those exploding ones are typically across power rails.  An overvoltage event with the high current available (or just frequent current surges, such as bulk filters on a PS) and... fireworks.  For the supply you show there shouldn't be any heavy current demands so the existing tantalums should be fine as is.  If you really feel the need to replace them, another tantalum of the same ratings or a good quality electrolytic of 150-200% or so of the tantalums value would do fine as well.

Hal

[schmitt trigger]

Agree with the statements above. Can’t add anything else.

On a related note, what sort of opamps are used in the console’s signal chain?
Just curious, but they would be NE5534, most likely.

[magic]

C5 may have been chosen for a particular ESR necessary for stability of this custom "boosted LM338" voltage regulator, as tants are said to have better ESR stability.

[gnuarm]

The problem with tantalum caps, which can bite you if you replace them, is that they can develop microscopic, internal shorts when heated for soldering.  At the factory, they bring up the voltage with current limiting, allowing the shorts to build up insulation, without catastrophic damage, solving the problem. 

When you do rework on your boards, replacing the caps, you should do the same thing.  Put a 100 ohm resistor in series (by whatever means you can), then turn on power.  Once it has been on for a few seconds, the cap will have healed any shorts.  The trick now, is to remove the series resistor, without heating the cap again! 

I use a tantalum cap on a board I build in the thousands.  Initially the factory didn't tell me they were seeing lots of failures on this part (which was a bit pricey for a cap at $2 ea).   When I found out, in particular, when one failed nearly burning the board, I did my due diligence.  Better late than never, eh?

There are lots of app notes on this issue by the cap makers.  The universal wisdom is what I've given above.  Tantalum caps need to be "broken in" with current limiting.  Once that is done, they are very reliable with high MTBF numbers.   A 1 uF, 35V tantalum should not be expensive, or large.   Mine were 150 uF, 16V in a surface mount package.

I have a test fixture that already had a very small value series resistor to measure the current.  I swapped that out with 100 ohms.  It also has a jumper to bypass it, so this is a perfect setup.  I don't think that will work for you.  Maybe make a break in a trace you can solder a resistor over, then solder a bit of wire, when done? 

Oh, option 1 is not a bad choice either.  I'm from the school of, "If it ain't broke...".  Tantalum caps do not develop a high failure rate with time, like electrolytic caps do. 

[p.larner]

ive always found its dried main filter caps dried causing excess ripple that kills tants.

[David Hess]

Based on the schematic they were applied properly, so I would leave the tantalum capacitors alone.

If you must replace them, then an aluminum electrolytic capacitor with 4 to 10 times the capacitance will be needed to get an equivalent ESR.

[gnuarm]

ive always found its dried main filter caps dried causing excess ripple that kills tants.

I'm not following you.  Are the tantalum caps the "dried main filter caps"? 

[fenugrec]

modern "aluminium polymer " electrolytics also have very low ESR, nearly as good as comparable tants IIRC.

[David Hess]

modern "aluminium polymer " electrolytics also have very low ESR, nearly as good as comparable tants IIRC.

They are much better than solid tantalum capacitors.  There are also polymer tantalum capacitors with similarly low ESR.

Neither of the polymer type of electrolytic capacitors are really substitutes for standard aluminum electrolytic or solid tantalum capacitors.  Their extra low ESR can cause problems in some cases, and they are limited to lower voltages and capacitance.  Where they really shine is when high ripple currents are required, like with switching power supplies.

[HalFoster]

And, importantly, with the same small form factor.