CurtPalme.com Home Theater Forum

[Spanky Ham]

Mac,
I am following this. I have my mouth shut for fear of saying something I shouldn't.

[draganm]

[macgyver655]

[macgyver655]

[Nashou66]

This is a thesis a Guess a grad student did with some interesting results.

http://digital.library.unt.edu/permalink/meta-dc-3104:1

One thing I must say , is for the novice electronics hobbyist manufactures data sheets do mislead
us to some extent as have other writing on the subject. Most of us do believe that the shelf life is about 5000 hours. And they do all talk about reforming and healing etc capacitors that have been stored a long time. So this person who researched this did a big favor to lots of us who will take the time to read it.

here are some excerpts.


Aluminum electrolytic capacitors have a limited shelf life. The aluminum oxide
dielectric is not stable. When a capacitor is in use (voltage applied) the dielectric is
continuously being produced (healing). Because no healing takes place during storage,
the dielectric strength will deplete due to dielectric polarization caused by impurities
which exist in the material from manufacturing (American Society for Testing and
Materials, 1978). The predominant effect on capacitors of this decrease in dielectric
strength is a large increase in the DC leakage current (DCL) (Greason & Critchley,
1986), the theory of which I have explained in chapter two.
There is not a set shelf life beyond which an electrolytic capacitor is guaranteed to
fail (Slaughter, 1996). Many manufacturers specify a typical shelf life for a capacitor or
an assembly containing capacitors. The manufacture’s stated shelf life for an aluminum
electrolytic capacitor can range from 2 to 10 years, depending on the quality of the
component (Kogler, 1999). Such a variation in shelf life can not be attributed only to
design and application variations. Instead, it is believed that the variations are based on
very conservative estimates of product life while in storage. The manufacturer will
typically specify time intervals that should be conservative under all normal conditions
(Davis & Funk, 1996). TXU Electric Comanche Peak Steam Electric Station established
a shelf life of five years on all aluminum electrolytic capacitors unless periodically
reformed (TXU Electric Comanche Peak SES, 1998).
The effect of aging on capacitors while in storage may be reduced or reversed by
performing periodic reconditioning (more commonly referred to as reforming) to restore
a capacitor’s internal properties (Davis & Funk, 1996). Different methods have been
described for reforming electrolytic capacitors. In general the process consists of
applying voltage to the capacitor without exceeding the specified leakage current. The
voltage is increased up to the rated value. This application of voltage tends to produce
aluminum oxide thereby restoring the dielectric layer (Davis & Funk, 1996).

After a very thorough testing process which I enjoyed reading about they came to this conclusion:


CHAPTER 5
CONCLUSION
The purpose of this research was to determine if the shelf life of aluminum
electrolytic capacitors was five years. An additional purpose of the research was to
determine if reforming aged aluminum electrolytic capacitors restored the aluminum
oxide dielectric film and in so doing reduced the DC leakage current.
Analysis of the test data resulted in the rejection of the null hypothesis and the
acceptance of the alternate hypothesis in both research questions for all sample
populations. It was concluded from this analysis, that this research clearly supports the
claim that the shelf life of aluminum electrolytic capacitors is some period in excess of
the five years currently assigned by TXU’s Comanche Peak Steam Electric Station
procedures. The claim that reforming of aged capacitors improved the DC leakage
characteristics of the capacitors was also proven.
The observed mean DC leakage currents of the aged unreformed capacitors, while
still within specification, did exhibit an increase that could be shown to be directly
associated with the duration of storage. For the longest aged can style capacitors, some
individual samples did have measured leakage currents slightly in excess of the design
values. Measurement of capacitance and ESR while not directly associated with the
research questions and hypotheses also supported the conclusion that the shelf life of the
capacitors is longer than five years.
The reforming process was shown to reduce the DC leakage current for all
populations sampled. Although none of the sample populations had pre-reforming DC
leakage values which were in excess of the design values, all populations show
improvement in this criteria. The process of applying voltage at rated voltage or slightly
above on a periodic basis can be associated with improvement to aged capacitors and
could be a basis for extended storage.


And then their recomendation

CHAPTER 6
RECOMMENDATIONS
The conclusion drawn from this research is that the shelf life of aluminum
electrolytic capacitors stored in a level “B” controlled environment is greater than five
years and exceeds the established limits currently in place at TXU’s Comanche Peak
Steam Electric Station. Changes to site procedures should be initiated to change the
assigned shelf life to a longer period, possible to 16 years from date of manufacture, a
period previously used based upon accelerated aging test. Assignment of an even longer
shelf life to aluminum electrolytic capacitors of the radial or axial style may be evaluated
upon further research. Additionally, the use of reforming techniques or periodic
application at rated voltage to extend storage and restore capacitors is warranted and
recommended.


So it is still dependent on some variable, like the quality of the capacitor, the storage temp and if they have been reformed or not. so 16 years would be a safe bet for a cap to still be good condition, but I ask do you trust this?
I would still recap after 10 years old, but thats just me.

Athanasios

[macgyver655]

[macgyver655]

Well I think at this point all we can say is that a specific shelf live cannot be just stated as 2 or 3 years and they are 100% no good and must be thrown away.

So what about if there is occasional use? As in the device is never dormant more then 2 years. Maybe used for a few months then off for a year, then on for a few weeks then off for a year? Kinda like some of our projectors, no?

[macgyver655]

Well I guess we still have some interest here so if someone can answer my last question "So who knows the numbers for hours vs lower temps?" then I will continue.......

[Sparky015]

Thanks Athanasios. One question I would have after reading those excerpts is what vintage caps are they using for this testing? 70s technology, 80s? You wouldn't believe the advances made in cap technology.

I think I agree and not agree with your statement about recapping after a certain period of time depending on the situation. If I knew the PJ sat for a decade, then no, because I know the caps are still good. Today's technology caps allow for a long shelf-life and since these PJs are used in predictable environments, it comes down to the quality of design, quality of the setup, and run-time (in my humble opinion). So if the pj is used, as 99% of them are unless your one of the lucky few to snag a NOS, I'd probably do the same. There is no document out there that says X number of hours with a typical setup for Y projector you need to change Z components. At least not out in the public, anyway. Something worthwhile for someone to offer. You don't want to change them after they start popping off the board or leak to the point the FETS and OP AMPS they support are strained and fail. Right now the only way to tell is to change them and then measure them. The key is to know whether the PJ is working properly. If it is, why change them even if they do fall out of the manufacturer specs?

These are obviously just my opinions based on what I've read and my EE education. I don't have the day in and day out experience with PJs and some do here on the forum to be able to provide solid data to back this up.

[Nashou66]

look at the end, they have charts of different brands, some are caps I like to use like Xicon low esr caps etc.

Athanasios

[macgyver655]

[v_erich]

Some values from the real life:

everytime when I change a capacitor in my 1998 BG808s, I measure it's values with a professional high grade LCR Meter, also the ESR I measure.
The values (old caps) are all in the range from -20 to -60% of the value printed on the cap.
The ESR comes up to max. 5 ohm, but most are in the range of 0,4-1 ohm.

In 2 weeks I will change the big caps in the SMPS (1000 and 2200µF), this will be interesting values.

When I solder a new e.g. Panasonic FC cap with 330µF/25V, the capacity is more or less this value (+- some percent), the ESR is around 0,1 - 0,2 ohm.

I think the ESR is measured with 1kHz.

Greetings,
Erich

[Nashou66]

There were lots of bad caps or counterfeit caps that came out of china, My iMac 20" G5 had them in the LVPS, they all popped, then the replacement Power supply had them too soon the began to bulge, before failure I re caped with panasonic FC and FM series and the comp has been running a year longer, being on 24/7. So i say look for caps from Japan as they tend to be better.

Athanasios

[Sparky015]

Mac, I was wondering when someone would start to pick apart my opinions!

[macgyver655]

[macgyver655]

[Sparky015]

[Sparky015]

tell you what, this thread alone is going to get me to 100 posts!

[macgyver655]

Sparky.... Thats not the counterfeit caps we are talking about. Many on here know about this issue so I guess I assumed you did. I would have to post links but it would be easier for you to just google bad caps or counterfeit caps. It is a LARGE and WELL KNOWN problem and goes way beyond just a few instances. If I feel like writing up something later I'll go into this problem on flat panels and other electronic devices.

[Sparky015]

ah, my bad. I'll check it out. I guess this problem hasn't hit my pro life yet because we are naturally behind the curve in updating designs. It takes years, sometimes a decade to get a new design qualified to go into production in order to drill a well. When it costs millions per day due to a down rig because of one failed 5 cent part, people lose jobs! I guess I gave some bored soles some reading material none the less!