SY, I'm really surprised to read these words from you....
Even if it may be apt. (I don't agree, obviously) it's still an insult or at least disrespectful, so a forum's policy infringement according DIY Audio rules, IMHO:
- Insulting, intimidating, bullying, harassing or other disrespectful or antisocial behavior.
Last edited:
1.2 Self-healing
Breakdowns can occur under heavy electrical load as a result of weaknesses or pores in the dielectric. The integrity of self-healing capacitors is not affected by such breakdowns.
When a breakdown occurs, the dielectric in a breakdown channel is broken down into its atomic
components by the electric arc that forms between the electrodes. At the high temperatures of as
much as 6000 K, a plasma is created that explodes out of the channel region and pushes the dielectric layers apart. The actual self-healing process starts with the continuation of the electric arc
in the propagating plasma. Here the metal layers are removed from the metal edges by evaporation.
Insulation areas are formed. The rapid expansion of the plasma beyond the areas of insulation and
its cooling in the areas of less field strength allow the discharge to extinguish after a few microseconds.
The area of insulation that is created is highly resistive and voltage-proof for all operating requirements of the capacitor. The self-healing breakdown is limited in current and so it does not represent
a short circuit. The self-healing process is so brief and low in energy that the capacitor also remains
fully functional during the breakdown.
Taken from: http://www.epcos.com/web/generator/...__en.pdf;/PDF_GeneralTechnicalInformation.pdf
I believe you are correct. That description does not apply to coupling capacitors, although it is an interesting read. Unless maybe the capacitor in question is severely under-rated.
Last edited:
I agree,
So if there is a very slight leakage it will continue to leak dependant on voltage / frequency until it self heals and voltage applied would have a direct effect. Applying higher voltages than in working conditions and frequencies over extended periods would/could have an effect..perhaps a music transient...if this was correct then the cap would get better with age to a point..(Just a thought).
http://www.wima.com/EN/selfhealing.htm
Regards
M. Gregg
Last edited:
There's a lot of research from 50-60 years ago on paper caps and the need for keeping their environment warm and dry (some of which I summarized above), and electrolytic cap manufacturers have certainly published a lot on the kinetics of oxide layer formation. Plastic dielectrics? Nope, with the exception of a recent paper by Doug Self on (relatively hygroscopic) Mylar.
In the case of paper (and perhaps Mylar), impurities, probably comprising water.
Clave, Perhaps you should have quoted SY's full text when you posted, as it reads very differently in the full context to the excerpt that you quoted. As SY said the conversation was drifting off topic. Please drop it and get on with the topic at hand.
Similar to stress testing? Running new components through extremes to weed out early failures.
It is called burn-in and is performed on completed equipment not components.
It is exactly what you think it is. Hook it up and run it. Nothing complicated.
In the manufacturing process some surface mount components will have to be baked at a certain temperature for a period of time. This is to reduce moisture content that can cause component failure. (Not to get a certain sound out of them.)
I hope that helps a little.
It is exactly what you think it is. Hook it up and run it. Nothing complicated.
In the manufacturing process some surface mount components will have to be baked at a certain temperature for a period of time. This is to reduce moisture content that can cause component failure. (Not to get a certain sound out of them.)
I hope that helps a little.
having worked in a semicondctor assembly and testing house many years ago, i know for a fact that we subject built-up devices in burn-in sequence, several hours of burning hot and freezing cold temperatures inside burn-in ovens with power applied to devices in several cycles.....
then comes final testing to see which devices survived the burn-in cycles.....devices that passed are then laser marked, tubed, boxed and shipped to the costumer.....devices that failed are then collected and analyzed for failure modes using scanning electron microscopes....
this is burn-in....
now clever audio salesmen hijacked this process, claimed it as their own and propagated myths about burn-ins.....
as P.T Barnum would say, there are suckers born every minute.....P. T. Barnum - Wikipedia, the free encyclopedia
the word gullible comes to mind...
then comes final testing to see which devices survived the burn-in cycles.....devices that passed are then laser marked, tubed, boxed and shipped to the costumer.....devices that failed are then collected and analyzed for failure modes using scanning electron microscopes....
this is burn-in....
now clever audio salesmen hijacked this process, claimed it as their own and propagated myths about burn-ins.....
as P.T Barnum would say, there are suckers born every minute.....P. T. Barnum - Wikipedia, the free encyclopedia
the word gullible comes to mind...
devices are tested to specs, if that device were an op amp, then it is tested for open loop gain, slew rate, input offset voltages, etc......this is still performance, only those devices that meet the specs are shipped........reliability is paramount if the semiconductor house is to establish good reputation among its clients...
I don't know about discrete devices, but the ICs I design in my day job go through a qualification process. A sizable number of devices are cooked in an oven at 125~150 deg C with 100 % humidity at the worst case operating point (high supply voltage, max operating frequency, etc.) for weeks. At regular intervals the devices are removed from the torture chamber and tested. Any failures are analyzed and the root cause of the failure identified and fixed (if applicable). Qual is a big deal.
Once a number of products have been through this qualification, the amount of time new products on that process have to spend in the torture chamber is generally reduced.
This has nothing to do with burn-in, however. It's an accelerated lifetime test. Basically, we want to ensure that our customers get good service life out of the parts if operated within the datasheet limits.
Parts that reach customers have been through several tests to ensure functionality and performance, but have not been burnt-in. They may have been powered for a few seconds. That's about it.
~Tom
Last edited:
I wouldn't say "nothing." The accelerated lifetime test torture chamber has some similarities with burn-in. Parts or assemblies are subjected to extremes of their operating range and then tested for failures. Or exercised and tested in alternating sequences via computer control. Burn-in is usually less than weeks at a time, though.
Now you might say this has nothing to do with break-in...
Regards testing,
There are military grade components and there are industrial grade and there are consumer grade..well whats the difference if they are all the same..
Well take a look at the ceramic package 555 timer..industrial grade or the 741 for an example...As was said to me once, "the infancy of electronics is small"..ie the failure rate of a component..if it lasts a week it will last a month if it lasts a month it will probably last the year etc... Its interesting that the old analogue electronics tended to have a higher fault rate than the modern complex computer grade components..modern manufacturing?
Then what about component failure? If components don't change with time then why do they fail? And OK some faults are instantaneous others are slower. (So the aging process is it from day one or starts after 5 years..is it faster or slower at the start or the end of the period of use?) My experience says fast aging at the start then a slow period of use followed by failure..
On a batch of equipment I built all were soak tested for 48hours for each one of 50 units, critical units were run for two weeks before shipping.... two failed with logic IC cmos or PSU failure. (Cap blow out / regulator failure). Most units on production were tested like this..
Set production equipment to spec..run 10 test pieces and test every one adjust process is necessary...run 100 test pieces and test a batch of 5 out of every 20... Run production line test every piece with auto test and spot check after every 100 produced with much higher rated tests. Then test and inspect 5 out of every 100 units. These included vibration tests, oven, over voltage. over current...and exposure to vapour etc.
Here is a thought..how do you test a space probe..IE voyager...and guarantee it won't fail in the next 50years in the 70's? (Or 1 week after launch..)
Regards
M. Gregg
There are military grade components and there are industrial grade and there are consumer grade..well whats the difference if they are all the same..
Well take a look at the ceramic package 555 timer..industrial grade or the 741 for an example...As was said to me once, "the infancy of electronics is small"..ie the failure rate of a component..if it lasts a week it will last a month if it lasts a month it will probably last the year etc... Its interesting that the old analogue electronics tended to have a higher fault rate than the modern complex computer grade components..modern manufacturing?
Then what about component failure? If components don't change with time then why do they fail? And OK some faults are instantaneous others are slower. (So the aging process is it from day one or starts after 5 years..is it faster or slower at the start or the end of the period of use?) My experience says fast aging at the start then a slow period of use followed by failure..
On a batch of equipment I built all were soak tested for 48hours for each one of 50 units, critical units were run for two weeks before shipping.... two failed with logic IC cmos or PSU failure. (Cap blow out / regulator failure). Most units on production were tested like this..
Set production equipment to spec..run 10 test pieces and test every one adjust process is necessary...run 100 test pieces and test a batch of 5 out of every 20... Run production line test every piece with auto test and spot check after every 100 produced with much higher rated tests. Then test and inspect 5 out of every 100 units. These included vibration tests, oven, over voltage. over current...and exposure to vapour etc.
Here is a thought..how do you test a space probe..IE voyager...and guarantee it won't fail in the next 50years in the 70's? (Or 1 week after launch..)
Regards
M. Gregg
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.