No, not another futile "capacitor sound quality" thread, this is about quality tests for electrolytic capacitors in the power supply.
Plans to tri-amp my home theater speakers mean quite a few amplifier power supplies, and the capacitors are a major cost component.
So I would like some objective tests of electrolytic capacitors to see if the less established brands are actually worse or merely less expensive while they try to make a name for themselves.
Capacitance and parallel resistance are easy to test, my main concern is ESR and lifetime.
Capacitor life time is correlated with size so I will only consider capacitors that are similarly sized to established brands.
My current candidate from Jaycar (Australia) is 50 mm diameter by 80 mm for the 10 000 uF 100 V benchmark.
There were widespread problems with "bad" capacitor electrolyte formulations a while back but now solved I believe.
So, for a non-bad capacitor I suspect ESR over time is a reasonable proxy for lifetime.
For a 10 000 uF capacitor the ESR should be in the order of 0.01 ohm.
To measure this sensibly really asks for a resolution of 0.001 ohm.
There is a cheapish ESR meter around (on Ebay and Amazon) that claims to do this, the MESR-100 v2.
Anyone have any experience with it?
Obviously 0.001 ohm resolution is a lot to ask of a cheap meter so I am curious how well it does.
I suspect it will need care with lead layout and the like.
Anyone have any other recommendations for measurement equipment?
Alternatively, a source of excellent but cheap capacitors would solve my problem.😉
Best wishes
Plans to tri-amp my home theater speakers mean quite a few amplifier power supplies, and the capacitors are a major cost component.
So I would like some objective tests of electrolytic capacitors to see if the less established brands are actually worse or merely less expensive while they try to make a name for themselves.
Capacitance and parallel resistance are easy to test, my main concern is ESR and lifetime.
Capacitor life time is correlated with size so I will only consider capacitors that are similarly sized to established brands.
My current candidate from Jaycar (Australia) is 50 mm diameter by 80 mm for the 10 000 uF 100 V benchmark.
There were widespread problems with "bad" capacitor electrolyte formulations a while back but now solved I believe.
So, for a non-bad capacitor I suspect ESR over time is a reasonable proxy for lifetime.
For a 10 000 uF capacitor the ESR should be in the order of 0.01 ohm.
To measure this sensibly really asks for a resolution of 0.001 ohm.
There is a cheapish ESR meter around (on Ebay and Amazon) that claims to do this, the MESR-100 v2.
Anyone have any experience with it?
Obviously 0.001 ohm resolution is a lot to ask of a cheap meter so I am curious how well it does.
I suspect it will need care with lead layout and the like.
Anyone have any other recommendations for measurement equipment?
Alternatively, a source of excellent but cheap capacitors would solve my problem.😉
Best wishes
Last edited:
Dave, not criticizing Jaycar, but mate, if you want quality, buy Nichicon, Chemicon, Elna, Panasonic etc.
These Yow-Ping, Flying Pigeon brand Chinese caps will get you in trouble in the long term. I know, I see them leaking all the time.
Jaycar sell stuff for hobbyists, and do a good job of it, but their stuff is cheap Chinese crap marked up.
You can get quality component shipped via Element14, RS, Mouser or Digikey, for less than a trip to Jaycar. I regularly get Fed-Ex to my door in 4-5 days from Mouser, Texas in the US. Can't beat that.
These Yow-Ping, Flying Pigeon brand Chinese caps will get you in trouble in the long term. I know, I see them leaking all the time.
Jaycar sell stuff for hobbyists, and do a good job of it, but their stuff is cheap Chinese crap marked up.
You can get quality component shipped via Element14, RS, Mouser or Digikey, for less than a trip to Jaycar. I regularly get Fed-Ex to my door in 4-5 days from Mouser, Texas in the US. Can't beat that.
Network Analyzer....costly but an invaluable tool for many tasks.
Less costly : Pulse generator and a good scope. Measuring Capacitor ESR and Inductance
This.
By eleventy billion.
How much do those cheap parts cost when you have to replace them even once.
And what about the collateral damage to things connected.
I still see posts in all areas of technology where cheap yum cha caps have failed.
lcd monitor/tv panel power supplies, induction cook tops, hi fi, general electronics.
Spend now, save later.
Thanks, I had already looked at this link before I posted my query.
It looks useful as a check of the RF properties of typical small to medium capacitors but I doubt that it will yield useful results on a power supply cap. with an ESR of 10 milliohms or so.
Or not with my oscilloscope and pulse source at least.
Say 5 volts behind 50 ohms, a 100 mA pulse.
That needs 100 uV resolution to see 1 millliohm.
Your equipment measures this?
Best wishes
David
For such low ESR you might need a preamplifier and a higher voltage pulse gen.
I'm a happy owner of a W&G SNA-2 network analyzer (cost me about $800 which was a pretty spectacular price for it), so I can make direct impedance measurements.
I'm a happy owner of a W&G SNA-2 network analyzer (cost me about $800 which was a pretty spectacular price for it), so I can make direct impedance measurements.
You can easily measure ESR, ESL, and capacitance (bridge methods are required), but you can't easily measure reliability or consistency. restorer-john's advice here is gold.
I probably shouldn't have mentioned Jaycar as a supplier, it was only an example.
But your list is instructive, a while back people said not to buy "Japanese junk", stick with reliable brands.
Now, of course, Japanese quality is exemplary, as your list implies.
So the question is whether I can find equal quality products from new brands that are lower priced because people are, understandably, reluctant to switch from a known brand without a financial incentive.
That's the whole point of my search for some objective tests.
I do not plan to buy dubious quality to save a few dollars.
Best wishes
David
You are lucky, nice score.
But Australia has a much smaller population than Europe and an even smaller electronics industry so reasonably priced second hand equipment is practically nonexistent.
Instead we take pride to use minimal equipment to maximal effect.
And an ESR meter sounds a useful tool anyway so that's what I will probably buy.
Best wishes
David
Capacitance is fairly easy, does not require Mr Wheatstone, ESL is unlikely to be relevant for any normal value in a power supply.
Accurate low ESR is the question, what is your method?
Well, I can measure consistency of ESR and most of the evidence implies that ESR should be a reasonable way to track reliability.
Most failures are due to unacceptable rise in ESR, often with a run-away as the increased ESR increases the capacitor temperature which increases the ESR which...
So I am inclined to run a load test and monitor the ESR, see if the increase looks reasonable.
Best wishes
David
Thanks James but I want 100 V, as in my first post, it's for a fairly powerful amp and I plan a bit of a safety factor.
A friend has pointed out that capacitor prices rise disproportionately past about 63 V so I should rethink the whole idea and built push-pull amplifiers with 50 V rails. Hmmm.
Best wishes
David
When I need to do this, I cobble together a four point measurement. Real He Men get a good commercial bridge.
I don't see that as plausible. When you measure ESR, you capture it at one point in time. What will it look like after 100 hours of handling ripple current in a warm environment? 1000 hours? Unit-to-unit? Lot to lot? Good manufacturers use SPC methods and continuously monitor reliability through sampling. If someone has that data and will share it with you, that's a good sign (though it's not unknown for lowest-bid suppliers to "enhance" the data).
When I did commercial electronics, we got burned several times by acceding to the desires of the bean counters. I introduced the concept of Cost of Quality to get our CEO to see the wisdom of not allowing the accountants to drive the vendor qualification process.
I wrote "So I am inclined to run a load test and monitor the ESR, see if the increase looks reasonable.", obviously not clear, I'll be more explicit.
I plan to measure the ESR, run a load test for a reasonable time say 100 hours monitor if the increase in ESR looks reasonable and could be extrapolated to a tolerable value at say 1000 hours.
Make sense?
More later, impatient friend arrived.
Best wishes
David
I'd say buy AEC Q200 rated components and derate generously and you'll be out of trouble for years of load life.
you can safely run 58Vdc to 59Vdc with 63Vcapacitors.
Check the maximum output voltage at the smoothing capacitors for a 230:40-0-40Vac transformer.
The nominal 58 to 59V goes up to around 61 to 62V for worst case in UK with a 253Vac maximum mains voltage.
If you have already slowly reformed your capacitors to 63V and checked the leakage current, you are safe at the nominal voltages and short term increases to just under the 63V rating.
With 2 small transformers you can increase both excitation current and voltage response by a factor of 10. This way ground loop is also solved easily.
Degradation is strongly affected by temperature. I recommend making accelerated tests.
Extrapolation is risky, you can never know where the process start to speed up, unless you went to the end at least with one specimen.
The factories that are really serious, like Swedish RIFA, has, inte construction process, as well as in the production, put the caps into all kinds of tests and stresses, to pick out the bad apples, or bad design.
That guarantees, that the capacitors from those manufacturers, meet the specs in the data sheets.
The way they are tested and measured are mostly out of reach for the hobbyist, making a test with a DIY instruments made up by $10 of components almost a joke.
Maybe the simple stuff we may assemble can be used to sort cheap and not-so-good caps from China. But for testing good caps .... Noooooo😕
I must also add, that it is common as well ascommon sense practice NEVER use an electrolytic at higher voltage that 85 - 90% of the rated voltage.
That guarantees, that the capacitors from those manufacturers, meet the specs in the data sheets.
The way they are tested and measured are mostly out of reach for the hobbyist, making a test with a DIY instruments made up by $10 of components almost a joke.
Maybe the simple stuff we may assemble can be used to sort cheap and not-so-good caps from China. But for testing good caps .... Noooooo😕
I must also add, that it is common as well ascommon sense practice NEVER use an electrolytic at higher voltage that 85 - 90% of the rated voltage.
I must add that I didn't want to be insulting anyone, just trying to give a rather dry, objective look on the matter.
Why apply a further factor of safety?
The manufacturer states the maximum working voltage (WV).
He already has a range of failure voltages some way above that WV and apply their own factor of safety to cover for variations in materials and construction.
In addition they have a surge voltage rating that allows very short term transients without total failure.
I suspect this last, when used, is what shortens the life of the capacitor. Each pulse that exceeds some threshold destroys a tiny part of the foil. Gradually that foil gets used up in snubbing out the transient pulses. This shows up with X rated capacitors. They are exposed to transient pulses all the time and over the years their capacitance drops. I have some I have measured and a few are getting down towards 60% of printed capacity. Many are down around 80 to 95%.
- Status
- Not open for further replies.