Tantalum decoupling caps

[RoadrunnerOZ]

Failure of these tantalum rail caps may be related to gradual failure of the primary electrolytic caps at the front end of a power supply leading to increased ripple which may cause the tantalums to reach critical internal temperature.
Also when any electronic device runs hot, after power off as it cools and the ambient temperature drops, because of air flow through the vents it will attract moisture from the air to the cooling boards and the components on them. The reliability of the solid tantalum capacitor is also heavily influenced by environmental conditions such as humidity and temperature. Moisture can penetrate the polymer encapsulating material, degrade the characteristics of the solid electrolyte, cathode attachment materials, and tantalum pentoxide dielectric, and cause the capacitor to fail - these devices are not hermetically sealed. Thus a capacitor which has absorbed moisture as it cools or is operating in an high humidity environment and is presented with a turn-on surge or heavy ripple it may blow. High relative humidity and high temperature both affect water diffusion, but apparently increased ripple voltage in 85°/85% testing causes tantalum capacitor characteristics to weaken and capacitors to fail.
Water molecules can be absorbed or adsorbed by hygroscopic materials, which have moieties that can form hydrogen bonds with water or metal cations that can form a coordinate covalent bond with the oxygen atoms in water. Several polymers used in electronics, such as epoxy resins, polycarbonate and poly (methyl methacrylate), are examples of hygroscopic materials. The saturated amount of absorbed moisture by such a material is proportional to the relative humidity.
Absorption of moisture over time leads to the gradual degradation of susceptible components such as tantalums.

 

[Steven Paris]

AND, in many devices that had a LOT of tantalum capacitors used as power supply bypass, locating the particular cap that was causing the power supply to fail (because of the tant short) was an absolute NIGHTMARE. That's why audio DIY guru Rod Elliott has described tantalum capacitors as fit only for the wastebasket.

 

[RoadrunnerOZ]

AND, in many devices that had a LOT of tantalum capacitors used as power supply bypass, locating the particular cap that was causing the power supply to fail (because of the tant short) was an absolute NIGHTMARE. That's why audio DIY guru Rod Elliott has described tantalum capacitors as fit only for the wastebasket.

I’ve had instances of just cutting sections off on PCB’s with a tant on just about every chip for short testing by cutting the print track to the + rail in stages and metering each section for short circuit. I made a trace cutter from a stainless steel dentists tool - each end has the opposite curve with left hook or right hook accessing horizontal or vertical traces - I ground a flat sharp chisel point to each. Also useful for lifting DIP chips after desoldering.

 

[RoadrunnerOZ]

Note re absorption of moisture - the old Roland Juno 106 has the voice filter chips AR80017A encased in black thick epoxy which over time becomes conductive and causes filter drift or HF oscillation - the only solution to this is to remove, soak in acetone and peel off the epoxy and pick remnants out from between and behind the legs of the smd IC’s, redipping frequently in acetone as you go - I made special tools for doing this picking from dentists tools for picks and a super thin piece of stainless steel wire on a handle to push through the backs of the legs to push out the epoxy. After that I spray the boards with PCB lacquer.
If you buy NOS AR80017A chips they will have the same problem - they become hygroscopic very soon. The problem is that bad that there is a company, Analogue Renaissance, that now makes new replacements (without epoxy ).
Black boxed potted epoxy circuits can also go the same way.

 

[Electrobumps]

old tantalum capacitors have a track record of failing short circuit that can cause serious problems across PS rails.

JR

That's interesting, in my 480-L I just replaced the noisy fan with a low noise fan, and decided to recap the electrolytics as one of the PSU caps plastic cap had come loose from the shrink wrap so did all of them.

That board is terrible to work on the solder has a high melting point and the boards traces lift pretty easily. I don't realty want to work on it again unless it really is wise preventative maintenance.

Should the tantalums on the PSU rails be replaced with
C0G ceramics?

 

[RoadrunnerOZ]

That's interesting, in my 480-L I just replaced the noisy fan with a low noise fan, and decided to recap the electrolytics as one of the PSU caps plastic cap has come loose from the shrink wrap so did all of them.

That board is terrible to work on the solder has a high melting point and the boards traces lift pretty easily. I don't realty want to work on it again unless it really is wise preventative maintenance.

Should the tantalums on the PSU rails be replaced with
C0G ceramics?

To desolder easily I always resolder first with lead/tin solder - in thin delicate areas I use lower iron heat once I’ve resoldered, suck out the solder with a manual sucker then use desolder braid - once most of the lead free solder has gone or mixed, the leaded solder replacing it flows easily to the braid - you may need to use leaded solder a second time for stubborn components. Never push the braid around, just push it down onto the joint or you’ll tear the print. Try and avoid pulling the legs through while applying heat - this can break the feedthrough off and lift the print on the component side - you need to straighten the legs before pulling through. I use a tubular desolder tool that can fit around a capacitor leg while the solder is melted and straighten it and if the hole is the right size the tube goes through the hole and ensures clean separation. Desoldering guns are useful but pushing hard and sliding the tip around the joint can break the print easily and lift the solder tab off the board with excess heat. Desolder tubes:

 

[k brown]

That's interesting, in my 480-L I just replaced the noisy fan with a low noise fan, and decided to recap the electrolytics as one of the PSU caps plastic cap had come loose from the shrink wrap so did all of them.

That board is terrible to work on the solder has a high melting point and the boards traces lift pretty easily. I don't realty want to work on it again unless it really is wise preventative maintenance.

Should the tantalums on the PSU rails be replaced with
C0G ceramics?

Bypasses don't need to be C0G - common types are fine.

 

[Script]

I made a trace cutter from a stainless steel dentists tool - each end has the opposite curve with left hook or right hook accessing horizontal or vertical traces

I cannot imagine what that looks like. Do you happen to have a picture?
(It must have been really many rants in that specific case).

 

[RoadrunnerOZ]

I cannot imagine what that looks like. Do you happen to have a picture?
(It must have been really many rants in that specific case).

Second from right, (first left of the tweezers) the third and fourth also modded to suit other jobs - a friend of mine who is a dentist gives me them when the tips break off and I modify them with a grinding wheel and jewellers files and 1200 then 2000 grit paper to finish. The kick/curvature on each end is opposite to the other. I made a fine chisel tip to pick epoxy from between the legs of the Juno smd IC’s and for removing bridging solder between legs of smd’s as well as the print cutter chisels.

 

[Script]

Now I see. Forgot what they are called exactly. Have that too over here -- needs sharpening though.

 

[CJ]

as said before, just energize shorted rail with external pwr supply and feel board for heat,

takes about 30 seconds to find the bad cap,

10 seconds if you have an IF camera

 

[RoadrunnerOZ]

as said before, just energize shorted rail with external pwr supply and feel board for heat,

takes about 30 seconds to find the bad cap,

10 seconds if you have an IF camera

All good if you have a supply that will drive a dead short

 

[JohnRoberts]

Should the tantalums on the PSU rails be replaced with
C0G ceramics?

cog/npo caps don't usually come in such large capacitance values.

Modern low Z aluminum electrolytic caps, designed for switching PS, are probably reasonable substitutes.

JR

 

[Brian Roth]

Tracking down WHICH (of perhaps dozens) power rail bypass cap has shorted is a PITA. I fiddled with a Toneohm unit but never quite got the knack.

One device that actually works for sniffing out which cap is shorted across a rail is an ESR meter. It takes a slight bit of experimentation to get the hang of it, but I've had good results "walking" along the power rail to determine the lowest resistance spot. The ESR meter is capable of detecting very low resistances. Here's the one I have:

https://anatekinstruments.com/products/fully-assembled-anatek-blue-esr-meter-besr
Bri

 

[Electrobumps]

To desolder easily I always resolder first with lead/tin solder - in thin delicate areas I use lower iron heat once I’ve resoldered, suck out the solder with a manual sucker then use desolder braid - once most of the lead free solder has gone or mixed, the leaded solder replacing it flows easily to the braid - you may need to use leaded solder a second time for stubborn components. Never push the braid around, just push it down onto the joint or you’ll tear the print. Try and avoid pulling the legs through while applying heat - this can break the feedthrough off and lift the print on the component side - you need to straighten the legs before pulling through. I use a tubular desolder tool that can fit around a capacitor leg while the solder is melted and straighten it and if the hole is the right size the tube goes through the hole and ensures clean separation. Desoldering guns are useful but pushing hard and sliding the tip around the joint can break the print easily and lift the solder tab off the board with excess heat. Desolder tubes:

I know, but none of this changes the quality of the board and a high tin content solder.

My question....is it worth doing the tantalums. Sometimes better to let sleeping dogs lie.

 

[RoadrunnerOZ]

I know, but none of this make up for the quality of the board and a high tin content solder.

You can use wherever solder you want afterwards but you’ll have difficulty getting components out that have been installed using lead free - it doesn’t flow well onto desolder braid. Cruddy board quality is not necessarily indicated by fine tracks that lift - it comes down to how much heat you can apply or if you’re applying lateral friction force under heat before the print lifts. It’ll happen on any board if the temp gets too high especially with age and a desolder gun on high. The soaking up of heat around big electrolytic caps may well be due to a large surface area of print they are soldered to - higher wattage iron for those puppies.

 

[k brown]

cog/npo caps don't usually come in such large capacitance values.

Modern low Z aluminum electrolytic caps, designed for switching PS, are probably reasonable substitutes.

JR

I thought the reference was to local bypasses at ICs.

 

[RoadrunnerOZ]

My question....is it worth doing the tantalums. Sometimes better to let sleeping dogs lie.

Yes it is worth it - the tantalums should be easier to remove than the power supply electrolytics. The power supply has large areas of print which soak up the heat from an iron. The tantalums are on finer print which will soak up very little heat. The unit is 30+ years old and the caps will be degrading by now. In a studio I designed, built and worked in, the local area humidity was really high - we ended up getting a dehumidifier to add to the aircon. Tantalum caps would blow in the older gear with regularity. They are somewhat hygroscopic and are not moistureproof. They may be ok now but as time goes on their withstand voltage starts to drop and eventually they won’t handle rail voltage. See my post above #101. There have been many tests and studies around explaining the failure of these devices. If a client asks me to recap to factory standards I’ll use them to replace existing but otherwise I’ll use low ESR electrolytic caps for PSU rail filters.
This sleeping dog will bite.

 

[CJ]

You can use a variac to drive a dead short, the voltage will never reach harmful levels, because of the short , have used this to burn off microscopic solder bridges due to bad pc process
Gotta use surgically light touch

 

[RoadrunnerOZ]

You can use a variac to drive a dead short, the voltage will never reach harmful levels, because of the short , have used this to burn off microscopic solder bridges due to bad pc process
Gotta use surgically light touch

Not real good for feeding AC to a supply with voltage regulators - they won’t turn on at a low rectified DC voltage then when they do they’ll cook I’d imagine.