electrolytic working voltage

[Walter Harley]

Followup to a question I posted on sci.electronics.components - trying for a
larger audience here. If you're interested in more details the s.e.c.
subject is "electrolytic mystery".

What is the consequence of exceeding the working voltage on an electrolytic,
if there is a current-limiting resistor in the circuit? I am testing a
circuit right now with a 48V supply driving a 15uF 25V capacitor through a
6.8k resistor. It's been running for a bit more than a day, with no
apparent consequences to the capacitor so far. I'm wondering whether the
cap's life is being reduced (and if so, by how much), or what.

 

[Charles Schuler]

Followup to a question I posted on sci.electronics.components - trying for a
larger audience here. If you're interested in more details the s.e.c.
subject is "electrolytic mystery".

What is the consequence of exceeding the working voltage on an electrolytic,
if there is a current-limiting resistor in the circuit? I am testing a
circuit right now with a 48V supply driving a 15uF 25V capacitor through a
6.8k resistor. It's been running for a bit more than a day, with no
apparent consequences to the capacitor so far. I'm wondering whether the
cap's life is being reduced (and if so, by how much), or what.

The capacitor will attempt to "reform" to the higher voltage. It might
overheat, explode, bulge, outgas, or leak. The resistor will limit the
current to 7 mA (worst case) for now, so the cooking process will be slow
and spectacular results are not likely. Remove the resistor for a more
spectacular display and keep your face away. Is it hot yet? The dielectric
film has to grow under these conditions. As to how long it will last,
that's complicated and will vary from one type to another.

 

[Roger Gt]

: Follow-up to a question I posted on sci.electronics.components -
trying for a
: larger audience here. If you're interested in more details the
s.e.c.
: subject is "electrolytic mystery".
:
: What is the consequence of exceeding the working voltage on an
electrolytic,
: if there is a current-limiting resistor in the circuit? I am
testing a
: circuit right now with a 48V supply driving a 15uF 25V capacitor
through a
: 6.8k resistor. It's been running for a bit more than a day,
with no
: apparent consequences to the capacitor so far. I'm wondering
whether the
: cap's life is being reduced (and if so, by how much), or what.
:


It is NOT the Capacitors life that is reduced! It's YOURS!

15uFD is a pretty small cap, and a cap will survive double voltage
and sometimes more at a stable voltage, but when you reach the
failure point the blast will alert the neighbors to call for the
coroner unless you have that CAP in a blast proof enclosure! I
have see LARGE caps take out a wall!




X X X X RIP!

 

[John Larkin]

: Follow-up to a question I posted on sci.electronics.components -
trying for a
: larger audience here. If you're interested in more details the
s.e.c.
: subject is "electrolytic mystery".
:
: What is the consequence of exceeding the working voltage on an
electrolytic,
: if there is a current-limiting resistor in the circuit? I am
testing a
: circuit right now with a 48V supply driving a 15uF 25V capacitor
through a
: 6.8k resistor. It's been running for a bit more than a day,
with no
: apparent consequences to the capacitor so far. I'm wondering
whether the
: cap's life is being reduced (and if so, by how much), or what.
:


It is NOT the Capacitors life that is reduced! It's YOURS!

15uFD is a pretty small cap, and a cap will survive double voltage
and sometimes more at a stable voltage, but when you reach the
failure point the blast will alert the neighbors to call for the
coroner unless you have that CAP in a blast proof enclosure! I
have see LARGE caps take out a wall!




X X X X RIP!

Less than 3 mA into a small cap will produce no explosion, no drama,
barely a perceptable temperature rise. I'd guess that the cap can
stand it forever. There's not enough energy available for a decent
pop.

John

 

[John Larkin]

Followup to a question I posted on sci.electronics.components - trying for a
larger audience here. If you're interested in more details the s.e.c.
subject is "electrolytic mystery".

What is the consequence of exceeding the working voltage on an electrolytic,
if there is a current-limiting resistor in the circuit? I am testing a
circuit right now with a 48V supply driving a 15uF 25V capacitor through a
6.8k resistor. It's been running for a bit more than a day, with no
apparent consequences to the capacitor so far. I'm wondering whether the
cap's life is being reduced (and if so, by how much), or what.

What's the cap voltage?

This is unlikely to damage the cap, in my opinion.

John

 

[Walter Harley]

What's the cap voltage?

This is unlikely to damage the cap, in my opinion.

The cap is rated 25V. The actual voltage on the cap is 48V, minus about 3mV
across the resistor (in other words, there's about 0.5mA of leakage).

 

[John Larkin]

The cap is rated 25V. The actual voltage on the cap is 48V, minus about 3mV
across the resistor (in other words, there's about 0.5mA of leakage).

Er, that's 0.5 microamperes.

Wow. In that case, power dissipation is zilch. I wonder if there's any
longterm thing going on, dendrites or gas evolution or something. Keep
it up and let us know what happens.

Maybe it's like 1N400x diodes: they make a huge batch of 400 volt
parts, then label them as various 50-400 volt parts and charge
accordingly.

John

 

[James Meyer]

Followup to a question I posted on sci.electronics.components - trying for a
larger audience here. If you're interested in more details the s.e.c.
subject is "electrolytic mystery".

What is the consequence of exceeding the working voltage on an electrolytic,
if there is a current-limiting resistor in the circuit? I am testing a
circuit right now with a 48V supply driving a 15uF 25V capacitor through a
6.8k resistor. It's been running for a bit more than a day, with no
apparent consequences to the capacitor so far. I'm wondering whether the
cap's life is being reduced (and if so, by how much), or what.

(Posted before reading any other replies.)

An aluminum electrolytic cap's dielectric is aluminum oxide that is
formed on one of its foil plates. The thickness of the dielectric along with
the area of the plates determines the capacitance. It also determines the
working voltage.

Electrolytic caps operated at a voltage high enough to cause some DC
current to flow through them will continue to form thicker dielectric coatings.
That has two results. The working voltage will go up to match the applied
voltage and the capacitance will go down compared to the value the capacitor
would have had if it had never been over voltaged.

Start with a new cap. Measure it's capacitance. Connect it to a
current limited voltage higher than its nominal working voltage. As soon as the
leakage current diminishes to a low level, increase the voltage a little.
Repeat the process a few times.

Then measure the capacitance of your newly formed capacitor. I'll bet
it is noticeably less than you started with.

Jim "That's my story, and I'm sticking to it." Meyer

 

[Walter Harley]

Er, that's 0.5 microamperes.

Hey, a zero here, a zero there...

Wow. In that case, power dissipation is zilch. I wonder if there's any
longterm thing going on, dendrites or gas evolution or something. Keep
it up and let us know what happens.

Maybe it's like 1N400x diodes: they make a huge batch of 400 volt
parts, then label them as various 50-400 volt parts and charge
accordingly.

Yeah, but electrolytics (as you know) have different physical sizes for
different voltages.

I'm pretty perplexed. Obviously the leakage current is very low; so at this
point (now that the cap is charged) it wouldn't make any difference whether
the 6.8k resistor was there or not. Perhaps it would make a difference if I
discharged the cap and then hit it with 48V with no current limiting. The
ESR of the cap is about half an ohm, so it would certainly dissipate a lot
of power for a little while. Maybe WV basically amounts to a surge power
rating - in this case, about 1kW for about half a usec.

I've not found much useful info on this topic on manufacturer web sites, so
far.

 

[Tam/WB2TT]

Followup to a question I posted on sci.electronics.components - trying for a
larger audience here. If you're interested in more details the s.e.c.
subject is "electrolytic mystery".

What is the consequence of exceeding the working voltage on an electrolytic,
if there is a current-limiting resistor in the circuit? I am testing a
circuit right now with a 48V supply driving a 15uF 25V capacitor through a
6.8k resistor. It's been running for a bit more than a day, with no
apparent consequences to the capacitor so far. I'm wondering whether the
cap's life is being reduced (and if so, by how much), or what.

The 6.8K seris resistor is not typical of the real world. We had a product
in beta testing that used 40 uf 20V Tantalum caps at 18V. After a while they
started to blow their guts out, and we had to change to 20 uf 40V (same case
size).

Tam

 

[John Larkin]

Hey, a zero here, a zero there...



Yeah, but electrolytics (as you know) have different physical sizes for
different voltages.

I'm pretty perplexed. Obviously the leakage current is very low; so at this
point (now that the cap is charged) it wouldn't make any difference whether
the 6.8k resistor was there or not. Perhaps it would make a difference if I
discharged the cap and then hit it with 48V with no current limiting. The
ESR of the cap is about half an ohm, so it would certainly dissipate a lot
of power for a little while. Maybe WV basically amounts to a surge power
rating - in this case, about 1kW for about half a usec.

I've not found much useful info on this topic on manufacturer web sites, so
far.

I built a power supply once (actually, a lot of them) where we wound
up using 35 volt caps that got up to 42 volts at low loads. These were
big 'lytics. After a few years, they tended to blow up.

John

 

[John Fields]

Er, that's 0.5 microamperes.

Wow. In that case, power dissipation is zilch. I wonder if there's any
longterm thing going on, dendrites or gas evolution or something. Keep
it up and let us know what happens.

Maybe it's like 1N400x diodes: they make a huge batch of 400 volt
parts, then label them as various 50-400 volt parts and charge
accordingly.

Interestingly, 1N400X's from Digi-Key are all the same price for the
same quantity from the same mfg, from 1N4001 through 1N4007.

Looking up Diodes, Inc.'s stuff yields that all the finished
rectifiers are made from one diode each and that they all have exactly
the same characteristics except for Ft, which depends on the package.
Looks like they may all be 1N4007's with the different part numbers
applied to keep _everybody_ happy.

 

[Jim Thompson]

On Mon, 23 Aug 2004 19:11:28 -0500, John Fields

[snip]

Interestingly, 1N400X's from Digi-Key are all the same price for the
same quantity from the same mfg, from 1N4001 through 1N4007.

Looking up Diodes, Inc.'s stuff yields that all the finished
rectifiers are made from one diode each and that they all have exactly
the same characteristics except for Ft, which depends on the package.
Looks like they may all be 1N4007's with the different part numbers
applied to keep _everybody_ happy.

They're all made on the same line, and are then screened for BV and
labeled accordingly. (At least they were at Motorola in the 60's)

...Jim Thompson

 

[Spehro Pefhany]

On Mon, 23 Aug 2004 19:11:28 -0500, John Fields

[snip]

Interestingly, 1N400X's from Digi-Key are all the same price for the
same quantity from the same mfg, from 1N4001 through 1N4007.

Looking up Diodes, Inc.'s stuff yields that all the finished
rectifiers are made from one diode each and that they all have exactly
the same characteristics except for Ft, which depends on the package.
Looks like they may all be 1N4007's with the different part numbers
applied to keep _everybody_ happy.

They're all made on the same line, and are then screened for BV and
labeled accordingly. (At least they were at Motorola in the 60's)

...Jim Thompson

Motorola's 1980s data indicate that there were actually two models-
the 1N4001-5 and the 1N4006/7. The latter had lower junction
capacitance and worse reverse and forward recovery times.

Best regards,
Spehro Pefhany

 

[Walter Harley]

The 6.8K seris resistor is not typical of the real world. We had a product
in beta testing that used 40 uf 20V Tantalum caps at 18V. After a while
they
started to blow their guts out, and we had to change to 20 uf 40V (same
case
size).

It is typical of one tiny part of the real world, which is balanced audio
line outputs. They tend to have DC blocking caps (usually aluminum
electrolytics) rated around 25V; and these outputs tend to get plugged into
mixer inputs with phantom power. Phantom power is generally 48V in series
with 6.8k. I've been trying to figure out why this doesn't cause more
trouble than it seems to.

Based on answers and experimentation thus far, the answer would seem to be
that what happens is that the caps are getting re-formed to be higher
voltage but lower capacitance. So probably some of those audio outputs are
underperforming in terms of cutting off more of the lows than they are
supposed to; but since they are usually rated conservatively that is
probably not causing many problems.

They don't seem to be exploding, in any event.

 

[Roger Gt]

: : > The 6.8K series resistor is not typical of the real world. We
had a product
: > in beta testing that used 40 uf 20V Tantalum caps at 18V.
After a while
: > they
: > started to blow their guts out, and we had to change to 20 uf
40V (same
: > case
: > size).
:
: It is typical of one tiny part of the real world, which is
balanced audio
: line outputs. They tend to have DC blocking caps (usually
aluminum
: electrolytics) rated around 25V; and these outputs tend to get
plugged into
: mixer inputs with phantom power. Phantom power is generally 48V
in series
: with 6.8k. I've been trying to figure out why this doesn't
cause more
: trouble than it seems to.
:
: Based on answers and experimentation thus far, the answer would
seem to be
: that what happens is that the caps are getting re-formed to be
higher
: voltage but lower capacitance. So probably some of those audio
outputs are
: underperforming in terms of cutting off more of the lows than
they are
: supposed to; but since they are usually rated conservatively
that is
: probably not causing many problems.
:
: They don't seem to be exploding, in any event.

Be patient, you won't be disappointed!
When it goes, it'll sound like a Firecracker!

 

[Pooh Bear]

Followup to a question I posted on sci.electronics.components - trying for a
larger audience here. If you're interested in more details the s.e.c.
subject is "electrolytic mystery".

What is the consequence of exceeding the working voltage on an electrolytic,

Leakage current increases.

if there is a current-limiting resistor in the circuit?

A measureable voltage may be developed across that resistor.

I am testing a
circuit right now with a 48V supply driving a 15uF 25V capacitor through a
6.8k resistor.

Audio phantom power ?

It's been running for a bit more than a day, with no
apparent consequences to the capacitor so far. I'm wondering whether the
cap's life is being reduced (and if so, by how much), or what.

You'll get away with it for a while I guess. I use 50 or 63 Volt parts in that
application.


Graham

 

[Pooh Bear]

It is typical of one tiny part of the real world, which is balanced audio
line outputs. They tend to have DC blocking caps (usually aluminum
electrolytics) rated around 25V; and these outputs tend to get plugged into
mixer inputs with phantom power. Phantom power is generally 48V in series
with 6.8k. I've been trying to figure out why this doesn't cause more
trouble than it seems to.

That's because only *microphone level* mixer inputs have phantom power.

You don't normally plug *line level* outputs into *mic level* inputs !

The line level inputs on mixers don't have phantom power.

Also - there's normally a switch to remove the phantom power when not required.

Based on answers and experimentation thus far, the answer would seem to be
that what happens is that the caps are getting re-formed to be higher
voltage but lower capacitance. So probably some of those audio outputs are
underperforming in terms of cutting off more of the lows than they are
supposed to; but since they are usually rated conservatively that is
probably not causing many problems.

They don't seem to be exploding, in any event.

I recommend a study of good audio practice too.


Graham

 

[Spehro Pefhany]

Be patient, you won't be disappointed!
When it goes, it'll sound like a Firecracker!

Worst case it's 84mW. I don't think anything is going to be blowing
up.

Best regards,
Spehro Pefhany

 

[John Larkin]

On Mon, 23 Aug 2004 19:11:28 -0500, John Fields

[snip]

Interestingly, 1N400X's from Digi-Key are all the same price for the
same quantity from the same mfg, from 1N4001 through 1N4007.

Looking up Diodes, Inc.'s stuff yields that all the finished
rectifiers are made from one diode each and that they all have exactly
the same characteristics except for Ft, which depends on the package.
Looks like they may all be 1N4007's with the different part numbers
applied to keep _everybody_ happy.

They're all made on the same line, and are then screened for BV and
labeled accordingly. (At least they were at Motorola in the 60's)

...Jim Thompson

Motorola's 1980s data indicate that there were actually two models-
the 1N4001-5 and the 1N4006/7. The latter had lower junction
capacitance and worse reverse and forward recovery times.

Best regards,
Spehro Pefhany

I suspect thet there are two major recipes, sliced about as you
describe. The higher-voltage guys seem to be PIN structures, and have
more forward drop and different charge-storage behavior. But from one
mfr, all the LV parts are generally the same, and ditto the HV parts.

Speaking of recipes, did I mention that cranberry beans are back in
season? Yay! I have a huge pot of beans and onions and bacon and
sausage simmering even as we speak. Or type. Whatever.

John