Electrolytics? Standard or high reliability?

[trans4funks1]

Hi,
I was about to order a BOM for a linear power supply for a tube mic and saw that some of the electrolytics I was going to order were out of stock. I searched for an alternative part and before I knew it I had embarked on a journey of selecting what may, or may not, be upgrades for all of the electrolytic caps in the PSU.

I found caps with the same form factor, capacitance, and voltage ratings that had much longer life spans and much higher ripple current capacity.

I am writing to ask: Is there anything I am missing that I should know about? I saw that brands like Nichicon and Panasonic had many choices of caps with similar form factor and the prices were very similar. I went to both companies websites and found they had short descriptions of their different series of caps. Sometimes the prices reflected the longer life expectancy and higher ripple current capacity but sometimes I found no price difference.

I also noticed there were some described as extremely low ESR which as far as I can gather are primarily for switching power supply designs.

Is there anything I may be missing, or I should be aware of, or is it ok to "upgrade", if you will, to the "high reliability - low ESR" grade of caps if you don't mind spending just a bit more money?

Thanks.

 

[trans4funks1]

I Just want to point out that I am not asking a question like "what electrolytic sounds best in a power supply". I am asking in case I overlooked the consideration of some factor that makes the "high reliability" caps less suitable than a "standard cap" as a filter in a power supply.

It seems like there is a spectrum or range of characteristics and some trade offs so the manufacturers provide different options. Both Nichion and Panasonic had series they described as mentioned above, but their electronic specs weren't listed so as to make it easy to compare like to like so I got a bit confused.

Thanks.

 

[audiomixer]

It might be about marketing quite a bit....

If you get better life expectancy out of a other range, well great. in some case lower ESR could be a problem, but surely not in a regular PS. Price might also be affected by production quantities and so on.

by the way, for the best possible sound you must add small 100nF bypass caps made of very expensive material...  kryptonite is known to sound sweet..

- Michael

 

[soapfoot]

High-temperature is usually worth it even if you're not using it in a warm environment.

As I understand, there's sort of a three-way relationship between voltage rating, temperature rating, and service life.  If you over-rate for both temperature and voltage somewhat, I understand you can get a very long service life from an electrolytic.

For audio applications, I've always had pretty decent luck with the Nichicon KZ series. They're not prohibitively expensive, they're marketed as "audio grade" (whatever that may mean) and they seem to work very well.

 

[JohnRoberts]

I am not aware of reliability being a specific capacitor characteristic that we can select or not.

There are inherent reliability differences between general types of capacitors based on their construction. For example film caps are generally more robust that electrolytic. The main failure mode for electrolytic caps is loss of electrolyte so it is a fair assumption that higher temperature caps (105' vs 85') will probably be less affected in normal temperature operation. 

JR

 

[trans4funks1]

The Aluminum electrolytic caps I was looking at in the "high reliability - high ripple" Panasonic EE series had service life predictions of 5,000 and 10,000 hours while the Panasonic FC "standard" series had 1,000 or sometimes 2,000 hour predictions. Nichion described their PM and PW series as high reliability and the VZ as standard.

One thing that confused me is that it seemed to me like high ripple current capability would go hand in hand with very low ESR but it seems like those characteristics are offered as balanced, one against the other.

I've been trying to read through Cornell Dublier's Aluminum Electrolytic Capacitor Application Guide

http://www.cde.com/catalogs/AEappGUIDE.pdf

I found it very interesting but it is also sort of over whelming to read in one sitting. I found I ended up with more questions after I read it.

Thanks.

 

[JohnRoberts]

The Aluminum electrolytic caps I was looking at in the "high reliability - high ripple" Panasonic EE series had service life predictions of 5,000 and 10,000 hours while the Panasonic FC "standard" series had 1,000 or sometimes 2,000 hour predictions. Nichion described their PM and PW series as high reliability and the VZ as standard.

One thing that confused me is that it seemed to me like high ripple current capability would go hand in hand with very low ESR but it seems like those characteristics are offered as balanced, one against the other.

High ripple current capability is the product of low ESR which defines how much heat that current generates (P=I^2 x R), and the heat dissipation out of the cap that will define internal temperature rise from dealing with that heat output.

I've been trying to read through Cornell Dublier's Aluminum Electrolytic Capacitor Application Guide

http://www.cde.com/catalogs/AEappGUIDE.pdf

I found it very interesting but it is also sort of over whelming to read in one sitting. I found I ended up with more questions after I read it.

Thanks.

Another general rule of thumb about reliability is that every 10' temp rise cuts MTBF in half. So heat from ripple current is a factor.

JR

 

[trans4funks1]

Hi John,
Thanks very much for the explanation.

 

[Rochey]

Does anyone know if the commercial and high-rel versions have a different manufacturing flow?

 

[JohnRoberts]

Does anyone know if the commercial and high-rel versions have a different manufacturing flow?

I have only been inside one capacitor factory and they only made (assembled) film caps and didn't even make the guts. This was back before SMD caps were widely used and this cap factory in Mexico mainly took small SMD like chip capacitors and finished them by attaching leads and dipping them in epoxy. I think the guts were made in a different factory inside the US and shipped to Mexico for the value added.

At Peavey I worked with a number of difference capacitor makers and some I had problems with I wouldn't trust as far as I can throw them, others were serious about process control and even when we had problems with their parts they bent over backwards to make it right.

I would buy caps from a highly respected well know brand with comprehensive data specs. I have seen sleazy parts distributers that rebrand cheap caps they bought who knows where and just re-wrapped them with their brand plastic sleeve. One of these sleaze-ball distributers tried to hire me to come work for him because I knew more about his capacitors than he did. :-( I dissected several to troubleshoot a problem with his parts.

arghh... the best cap is no cap...

JR

 

[PRR]

> not aware of reliability being a specific capacitor characteristic that we can select or not.

Maybe not from Mouser.

I just read the report on the 1956 Trans-Atlantic telephone cable.

This is vacuum tube amps and caps and such, 2 miles down under the sea. The goal was NO failures for 20 years. (Repairs were possible but brutally costly and slow.)

WE built a clean-room where workers built about 900 capacitors in 18 months. Less than 2 per day. Every worker was pre-qualified on test material. Every operation was directly over-seen by an inspector. Every operation was signed-off by both worker and inspector. Then the finished cap got extensive testing, like 4 months at 1.5X operating voltage (prototypes had been run at 2X and then discarded). As much as 95% of the top-grade mica they bought was discarded for flaws or damage.

In fact the 1956 cable served 22 years, and was retired mostly because so many newer cables and channels had been laid that it wasn't worth keeping on the books.
_________________________

In a tube-mike, ESR and Ripple are utter non-issues. I'd take the highest commercial temperature rating (usually 105C). I'd also note that, unlike an underwater cable designed for 175,000 hours continuous revenue work, a tube-mike runs a few hours a week (say 200 hours a year), you can always switch to your 2nd-pick mike to capture the music of the moment, and you don't have to hire a ship and drag the bottom to bring it up for re-cap.

 

[JamesW]

PRR I love that you know that shit

 

[dfuruta]

This is vacuum tube amps and caps and such, 2 miles down under the sea. The goal was NO failures for 20 years. (Repairs were possible but brutally costly and slow.)

Pardon the digression, but do you know how these were powered?

 

[PRR]

> how these were powered?

2,500VDC redundant at both ends. 52 repeaters powered series-string. 28% of power wasted in cable resistance. Signal/power split networks each end of each repeater. About 3,200dB total signal loss (at 160kcps) to be made-up by a repeater every 37 miles. "Unmeasurable" error of dB loss per mile adds-up in 2,000 miles so the line was 'live' while laying and EQ and repeater spacing was adjusted while the cable was dropping over the stern of the ship.

http://www3.alcatel-lucent.com/bstj/vol36-1957/bstj-vol36-issue01.html