[apa]

In my understanding the conventional rule of thumb is to replace
filter caps after some number of years (and the number varies
depending of who you ask). The same thinking says caps 30+ years
really should replaced on the grounds that if they're not failing now
they will soon.

I've seen a number of old caps (especially the cans in Ampeg V -series
amps but in other old tube amps as well) where there is no appreciable
ripple in the supply. Testing the caps on a Sencore meter shows them
to be well with leakage for their value. Sencore's dieletric
absorption test is to run the leakage test until the cap is showing
leakage in the acceptable range and then switch quickly back to
measuring the cap value. If the value creeps up slowly the DA is too
high, if it rises up quickly, it's fine. It doesn't define "slowly"
but the caps under test easily match the newer ones I have in stock.

So here's my question: If the caps are meeting these two tests and
there's no ripple problem in the supply, is there any advantage to
replacing them?

Thanks, Andy

[Scott Dorsey]

apa wrote:
In my understanding the conventional rule of thumb is to replace
filter caps after some number of years (and the number varies
depending of who you ask). The same thinking says caps 30+ years
really should replaced on the grounds that if they're not failing now
they will soon.

This is not really true... there are some kinds of electrolytic caps that
are very prone to failing, and there are some that are not. And for the
most part, ones that are in hot places will tend to fail faster than ones
that are kept cold.

I've seen a number of old caps (especially the cans in Ampeg V -series
amps but in other old tube amps as well) where there is no appreciable
ripple in the supply. Testing the caps on a Sencore meter shows them
to be well with leakage for their value. Sencore's dieletric
absorption test is to run the leakage test until the cap is showing
leakage in the acceptable range and then switch quickly back to
measuring the cap value. If the value creeps up slowly the DA is too
high, if it rises up quickly, it's fine. It doesn't define "slowly"
but the caps under test easily match the newer ones I have in stock.

Yes, but that doesn't mean that they aren't going to fail prematurely.

So here's my question: If the caps are meeting these two tests and
there's no ripple problem in the supply, is there any advantage to
replacing them?

Well, the thing is, sooner or later you're going to have a failure. You
don't really know where you are on the failure curve, but you DO know that
the failure curve is a bathtub distribution and sooner or later you get
to the end of it.

If it's mission critical equipment and someone is going to die if it fails,
I'd replace the caps, and I'd probably use twist-lock caps that match the
originals as well as possible (Antique Electronics Supply will stock them).

If it's a guitar amp and the consequences of failure are having to play
an acoustic set.... sheesh, just wait until something goes wrong.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[apa]

On Nov 17, 3:28 pm, (Scott Dorsey) wrote:
apa wrote:
In my understanding the conventional rule of thumb is to replace
filter caps after some number of years (and the number varies
depending of who you ask). The same thinking says caps 30+ years
really should replaced on the grounds that if they're not failing now
they will soon.

This is not really true... there are some kinds of electrolytic caps that
are very prone to failing, and there are some that are not. And for the
most part, ones that are in hot places will tend to fail faster than ones
that are kept cold.

I've seen a number of old caps (especially the cans in Ampeg V -series
amps but in other old tube amps as well) where there is no appreciable
ripple in the supply. Testing the caps on a Sencore meter shows them
to be well with leakage for their value. Sencore's dieletric
absorption test is to run the leakage test until the cap is showing
leakage in the acceptable range and then switch quickly back to
measuring the cap value. If the value creeps up slowly the DA is too
high, if it rises up quickly, it's fine. It doesn't define "slowly"
but the caps under test easily match the newer ones I have in stock.

Yes, but that doesn't mean that they aren't going to fail prematurely.

So here's my question: If the caps are meeting these two tests and
there's no ripple problem in the supply, is there any advantage to
replacing them?

Well, the thing is, sooner or later you're going to have a failure. You
don't really know where you are on the failure curve, but you DO know that
the failure curve is a bathtub distribution and sooner or later you get
to the end of it.

If it's mission critical equipment and someone is going to die if it fails,
I'd replace the caps, and I'd probably use twist-lock caps that match the
originals as well as possible (Antique Electronics Supply will stock them).

If it's a guitar amp and the consequences of failure are having to play
an acoustic set.... sheesh, just wait until something goes wrong.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Ok. So failure is my only worry.
There's not a subtle functional difference I'm not privy to.
Thanks as usual Scott.

-Andy

[Eeyore]

apa wrote:

In my understanding the conventional rule of thumb is to replace
filter caps after some number of years (and the number varies
depending of who you ask). The same thinking says caps 30+ years
really should replaced on the grounds that if they're not failing now
they will soon.

NO. Electrolytics fail due to loss of electrolyte. The primary cause of
this is heat, either in operation or storage.

There is no simple answer. But 30 years is going some.

Plus, if they're very old, they will need gradual 're-forming' to recreate
an Al2O3 layer on the aluminium film.

Graham

[Scott Dorsey]

apa wrote:

Ok. So failure is my only worry.
There's not a subtle functional difference I'm not privy to.

If the series resistance of the capacitor has changed or the value has
changed, it will change the sound of the amp. But it COULD change it
for the better, too. Whatever that means. If the cap checks out correctly
out of circuit on a dynamic tester, it's going to sound the same as a new one,
though.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[apa]

On Nov 18, 3:24 pm, (Scott Dorsey) wrote:
apa wrote:

Ok. So failure is my only worry.
There's not a subtle functional difference I'm not privy to.

If the series resistance of the capacitor has changed or the value has
changed, it will change the sound of the amp. But it COULD change it
for the better, too. Whatever that means. If the cap checks out correctly
out of circuit on a dynamic tester, it's going to sound the same as a new one,
though.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

What's a dynamic tester exactly? The Sencore I have tests for value
and leakage at rated voltage. It also has a vague method for testing
for dielectric absorption which is the only test it does which I'd
think could be considered dynamic. Is there more to it than that?

[Scott Dorsey]

apa wrote:

What's a dynamic tester exactly? The Sencore I have tests for value
and leakage at rated voltage. It also has a vague method for testing
for dielectric absorption which is the only test it does which I'd
think could be considered dynamic. Is there more to it than that?

There used to be two kinds of testers. One was a bridge that measured
the capacitance by comparison with a reference capacitance, and the
other was a relaxation oscillator combined with a frequency measuring
device.

The bridge type devices are generally better at finding electrolytics
with problems. The older tube-type Sencores were like this. The newer
ones I don't know about. The capacitance function on my cheap B&K VOM
is actually an oscillator and counter type.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."