Just for fun, another reason I replace caps.....

H

HPMguy

Super Member
I've been reading lately about some technicians testing capacitors and putting them back in aging equipment (30+ years old) because they test "in spec" -- no test results listed.

My brother brought me a digital readout to look at the other day, stopped powering up a while back. No power on lamp, no activity.

I did a bit of digging, and what I found was that it has a power down function for the computer if the + and - 15V rails are not right, plus of course won't operate at all if the +5 rail is below 4.2V or so just like all digital computers of that age -- made in 1997. Suggested repair was replace the PS caps.

So I bought the recommended set (Nichicon UPW) and replaced them. All tested "in spec" although the 68uF 400V switching PS main filter cap read 51 uF, None of them were over 1.2 ohm ESR, leakages were all 2% or less.

However, after replacement it fired up perfectly. I'm thinking that means ALL of the caps less than 100uF I pulled from my second KA-80 last week are bad and were probably affecting the sound somewhat.

Moral of the story: Even if old caps test "within 20% of rated capacity" and have reasonable ESR, they can still be causing trouble!
 
Register to hide this ad
HPMguy said:
I've been reading lately about some technicians testing capacitors and putting them back in aging equipment (30+ years old) because they test "in spec" -- no test results listed.

My brother brought me a digital readout to look at the other day, stopped powering up a while back. No power on lamp, no activity.

I did a bit of digging, and what I found was that it has a power down function for the computer if the + and - 15V rails are not right, plus of course won't operate at all if the +5 rail is below 4.2V or so just like all digital computers of that age -- made in 1997. Suggested repair was replace the PS caps.

So I bought the recommended set (Nichicon UPW) and replaced them. All tested "in spec" although the 68uF 400V switching PS main filter cap read 51 uF, None of them were over 1.2 ohm ESR, leakages were all 2% or less.

However, after replacement it fired up perfectly. I'm thinking that means ALL of the caps less than 100uF I pulled from my second KA-80 last week are bad and were probably affecting the sound somewhat.

Moral of the story: Even if old caps test "within 20% of rated capacity" and have reasonable ESR, they can still be causing trouble!
Click to expand...
Yepper! Electrolytic caps of 30+ years of age are breaking down and there are no "ifs, ands, or buts" about it. ESR and capacitance are only two of the many characteristics of capacitors.
 
I tested a few power supply main filter caps, and some have higher then specked UF results (higher capacitance then specs). What to think about that?. Tested with a Peak Atlas ESR70.
 
It seems that if you went through the hassle of removing the old caps, you might just as well put new ones in there, since they are inexpensive parts. New ones may well perform better than the old ones in terms of parameters other than capacitance, such as leakage current, ESR and self resonant frequency. If these parts were costly, I'd change my tune and change out only bad ones, but they're not, so shoot the wad. Just be sure to go with recognized brands, and not some ebay no-names.
 
turboyam said:
I tested a few power supply main filter caps, and some have higher then specked UF results (higher capacitance then specs). What to think about that?. Tested with a Peak Atlas ESR70.
Click to expand...
Within 20% (if 20% tolerance) is "in spec"... ;) A 10k can be 12k, or 8k and be "within spec"...
 
egellings said:
It seems that if you went through the hassle of removing the old caps, you might just as well put new ones in there, since they are inexpensive parts. New ones may well perform better than the old ones in terms of parameters other than capacitance, such as leakage current, ESR and self resonant frequency. If these parts were costly, I'd change my tune and change out only bad ones, but they're not, so shoot the wad. Just be sure to go with recognized brands, and not some ebay no-names.
Click to expand...

Are you replying to my post?. I have a few amps but one of them is a SAE 2600. It has 4 mini Cola can sized caps, that cost about $200.00 CAD for the 4 of them so...
 
HPMguy said:
So I bought the recommended set (Nichicon UPW) and replaced them. All tested "in spec" although the 68uF 400V switching PS main filter cap read 51 uF, None of them were over 1.2 ohm ESR, leakages were all 2% or less.

However, after replacement it fired up perfectly.
Click to expand...

curious why it didn't work then. Unless the 68uf really needs to be closer to 68 than the measured 51. Typically thats not super critical though, it just flattens out the rectified wall voltage to DC so it can get chopped up again in the switcher. Unless perhaps it had just too much ripple to allow the switching circuitry to kick in and function.

If thats the case, it sounds like they didn't select a cap at a reasonable value to allow the circuit to function even at the extremes of tolerance.


and for what its worth I tend to agree with not re-installing old parts. I generally figure if I'm pulling it to measure, I'm probably going to change it. I will sometimes re-install if I'm diagnosing things just to not introduce too many variables though.
 
gadget73 said:
curious why it didn't work then. Unless the 68uf really needs to be closer to 68 than the measured 51. Typically thats not super critical though, it just flattens out the rectified wall voltage to DC so it can get chopped up again in the switcher. Unless perhaps it had just too much ripple to allow the switching circuitry to kick in and function.

If thats the case, it sounds like they didn't select a cap at a reasonable value to allow the circuit to function even at the extremes of tolerance.


and for what its worth I tend to agree with not re-installing old parts. I generally figure if I'm pulling it to measure, I'm probably going to change it. I will sometimes re-install if I'm diagnosing things just to not introduce too many variables though.
Click to expand...
Ever hear of leakage? Especially at operating voltage? That will stop some things dead in their tracks. Low voltage "testers" can't meaningfully measure leakage.
 
I'm guessing (since I didn't measure anything first) that excessive ripple resulted in erratic or low voltages in the supply DC to the board. It has a low voltage shut-off feature for the +15 and -15 volt lines to avoid burning up expensive encoders (it's used on high end milling machines and lathes for position control of the cutter).

The rest of the caps were OK except for higher ESR than stock, but I'm willing to bet they would fail in-circuit testing at actual use loads and voltages.

I suspect that both high and low capacitance values for old electrolytic caps really means internal shorts or dry electrolyte -- in either case, they will not charge properly. Shorts will increase the measured capacitance because the charge is leaking off while testing, and dry electrolyte doesn't permit electron migration so the cap charges up really fast. Either than have high resistance -- I've found one dead cap in the past that tested as a resistor at 6 ohms, no charging capacity at all.

Caps that have leaked out can also show "normal" capacitance on a low voltage tester, probably a combination of high leakage and dry electrolyte that by happenstance results in the "correct" value. Won't work at all though.....

The apparatus was produced in 1997 and has been energized pretty much 24/5 since, only get shut off on weekends, and not always then either.
 
HPMguy said:
.....Moral of the story: Even if old caps test "within 20% of rated capacity" and have reasonable ESR, they can still be causing trouble!
Click to expand...

I think another aspect here could be 20-30+ year old mechanical connections (solder) that was failed or failing providing an in circuit series resistance.
That being said for the cost of most of the e-caps, IMO, I'm updating these components with priority on the high stress circuits (thermal+electrical) like power supplies and then main amp and driver stages. If after testing I find noise or coupling issues those e-caps get updated also.
 
Testing caps is much like testing tubes.
The right gear helps.
Try most tube testers, then use a Hickok.
I am going to go look at the supplied links.
Hopefully they are informative.
 
Solder was fine, well sealed case (as is the rule in machine shops, lots of coolant mist in those places, and metal dust!). But old caps are old caps.....
 
Skywatcher said:
Ever hear of leakage? Especially at operating voltage? That will stop some things dead in their tracks. Low voltage "testers" can't meaningfully measure leakage.
Click to expand...

well aware of leakage. I mess with tube stuff. Higher voltage and much older, with far worse quality components. Its made me very not-sentimental towards old caps and other passive bits. I just like to have a solid reason when I replace parts. Don't get me wrong, I have definitely crossed my fingers and thrown a part at a problem but it bothers me when I don't know why I'm changing things besides not having any better idea.

HPMguy said:
The rest of the caps were OK except for higher ESR than stock, but I'm willing to bet they would fail in-circuit testing at actual use loads and voltages.
Click to expand...

thats probably what it is. Switch mode supplies are sensitive to ESR in the caps. They run at high frequency, and the higher the frequency, the higher the ESR will be. If the supply runs higher than the tester does, it will be even worse in-circuit than you measure.
 
Skywatcher said:
Ever hear of leakage? Especially at operating voltage? That will stop some things dead in their tracks. Low voltage "testers" can't meaningfully measure leakage.
Click to expand...

Wow. That seems a bit patronizing.

I suspect that he knows as much about the topic as you do, based on years of reading and learming from his posts.
 
spark1 said:
Wow. That seems a bit patronizing.

I suspect that he knows as much about the topic as you do, based on years of reading and learming from his posts.
Click to expand...
Not at all. Direct coupled Solid State units are more, I guess you could say, "sensitive" to leaky capacitors, especially in logic circuits, than tube gear is. Tubes will still operate at near, or in red plate conditions and you won't "hear" it. Transistors in the same state will puke and stumble, and sometimes go poof! in short order rather than degrade, albeit accelerated, over time. @gadget73 is a smart cookie, understands capacitors, and I believe he understood my comment on low voltage testers.
 
turboyam said:
Are you replying to my post?. I have a few amps but one of them is a SAE 2600. It has 4 mini Cola can sized caps, that cost about $200.00 CAD for the 4 of them so...
Click to expand...

I usually change all the little capacitors on the board, which are usually more failure prone, and leave those big ones in place unless someone is willing to get out their wallet. But one thing you may want to keep in mind, is a lot of those screw terminal computer grade capacitors are being discontinued, so if you don't replace them now, you may have a hard time finding replacements once failure occurs.
 
I'm amazed at how many times caps are to blame in stereo equipment failures. My tech found bad caps in my Adcom 535 amp that wasn't passing bass signals. My ham radio buddy fixed one of my radios by replacing all the caps in the hand-held microphone.
 
You must log in or register to reply here.
Back
Top Bottom