Page 1 of 7
Dielectric Absorption
Posted: Sun Mar 25, 2018 4:03 pm
by mediatechnology
Dielectric absorption is the name given to the effect by which a capacitor, that has been charged for a long time, discharges only incompletely when briefly discharged. Although an ideal capacitor would remain at zero volts after being discharged, real capacitors will develop a small voltage from time-delayed dipole discharging, a phenomenon that is also called dielectric relaxation, "soakage", or "battery action." https://en.wikipedia.org/wiki/Dielectric_absorption.
Perhaps this post should be under JR's Entropy thread...
Recently, while checking a 48V switching power supply, I charged a 47µF/63V Panasonic cap to 48V in order to bypass the supply output.
It needed a simple RC filter to be useful.
Before putting the cap away I discharged it for a few seconds into a dead short.
Being curious I measured it a few seconds later and noticed it recharging.
This is due to dielectric absorption...something I've had to deal with before in mic preamp phantom coupling caps.
It was charged and "discharged" about a week ago.
I checked 2-3 days ago and it was at about 1.5V
Today its still at 1.5V with no sign of discharging.
I'll hold onto the cap awhile longer and measure it. (Until I get bored with it.)
I think we'll find it will hold the 1.5V for a very,very long time.
Re: Dielectric Absorption
Posted: Sun Mar 25, 2018 9:32 pm
by JR.
I wrote about DA in my old audio mythology column back in the '80s, and wrestled with the audiophools that thought it would be audible.
It might be amusing to hang a sensitive micro-ammeter across the mostly discharged cap and see how long it trickles out charge.
JR
Re: Dielectric Absorption
Posted: Mon Mar 26, 2018 5:11 am
by mediatechnology
If I hold this one to my ear I hear the sound of the sea.
![Laughing :lol:](./images/smilies/icon_lol.gif)
Re: Dielectric Absorption
Posted: Mon Mar 26, 2018 7:17 am
by emrr
There was a type of grid bias battery used in the 1930's that looks a lot like a hearing aid battery, clipped in near the tube cathode. I found one awhile back, and it still has a charge.
Cap tangent; I rebuilt another 1930's preamp last year which still had it's original bathtub type film caps. They measure fine for C, yet series R is very high compared to a new or even unused similar vintage cap. They sound fine, perhaps because the amp is a high Z circuit anyway, their increased R is of little consequence. They're staying in as the piece is of the collectors status which would be harmed by modern parts replacements.
Re: Dielectric Absorption
Posted: Mon Mar 26, 2018 10:56 am
by mediatechnology
I don't recall ever seeing a grid bias battery except maybe in an instrument. Was this one Mercury or Silver Oxide? Was it growing any crystals?
I think I've had to deal with dielectric absorption, "DA" in circuit design one time and that was input switching ahead of a mic preamp.
Getting the charge off and keeping it off to prevent click-free switching from mic to line is challenging.
We discussed DA here in the RMS detector thread. One of us hypothesized that a low DA film cap might make a more accurate RMS detector (compared to a low-leakage Tantalum) because it might have less errors for small gain changes.
As a practical matter (staying alive) I remember that in my RF days we kept the transmitter's spare oil-filled plate capacitors shunted. They work at 7 kV and a little bit of recharge is deadly. The operating ones are continuously shunted when the interlocks are activated for an open door.
Today's cap voltage is 1.4 V when initially connected.
The DVM begins to discharge it a little but it bounces right back.
Re: Dielectric Absorption
Posted: Mon Mar 26, 2018 12:47 pm
by JR.
mediatechnology wrote: ↑Mon Mar 26, 2018 10:56 am
I don't recall ever seeing a grid bias battery except maybe in an instrument. Was this one Mercury or Silver Oxide? Was it growing any crystals?
I think I've had to deal with dielectric absorption, "DA" in circuit design one time and that was input switching ahead of a mic preamp.
Getting the charge off and keeping it off to prevent click-free switching from mic to line is challenging.
Perhaps a leakage issue (too)?
We discussed DA here in the RMS detector thread. One of us hypothesized that a low DA film cap might make a more accurate RMS detector (compared to a low-leakage Tantalum) because it might have less errors for small gain changes.
I don't recall that discussion but DA is mainly expressed when there are different charging and discharging time constants or impedances. IIRC the THAT RMS detector charges with low Z and discharges high Z so indeed DA could cause an error, but it is hard to imagine this being a serious problem for a single ended dynamic processor.
Perhaps if that RMS detector was used in an encode/decode companding NR (like the old DBX nr) using different DA caps for encode and decode could cause tracking errors.
As a practical matter (staying alive) I remember that in my RF days we kept the transmitter's spare oil-filled plate capacitors shunted. They work at 7 kV and a little bit of recharge is deadly. The operating ones are continuously shunted when the interlocks are activated for an open door.
Today's cap voltage is 1.4 V when initially connected.
The DVM begins to discharge it a little but it bounces right back.
Discharge current?
-----
OK this old story is starting to bore even me, but back when I designed my CX decoder kit, I intentionally used a tantalum cap for the CX playback side-chain time constant to symmetrically reflect the (Urie) encoder that used a tantalum cap in its side-chain. In a side chain circuit with different attack and release time constants it is easy to imagine tracking errors between high DA and low DA caps, while these errors should be slow changing and subtle. Depending on which is used where the noise floor could suffer some (slow) modulation, decaying too soon or not soon enough.
JR
Re: Dielectric Absorption
Posted: Mon Mar 26, 2018 3:55 pm
by mediatechnology
Leakage current is a factor with phantom on.
DA is problematic when phantom is off.
THAT story plays out here beginning on page 7:
http://www.thatcorp.com/datashts/dn140.pdf
I found that post where we talked about DA in the RMS:
www.proaudiodesignforum.com/forum/php/v ... 6&start=20
My thinking is that DA might cause low-level gain changes at times it shouldn't.
And that big honkin' film is way more impressive than a tiny bead Tantalum.
Re: Dielectric Absorption
Posted: Tue Mar 27, 2018 7:15 am
by mediatechnology
This morning's voltage read was 1.546V.
Higher than yesterday and could be measurement error. I only leave the meter connected for about a second since I don't want its 10MΩ load to discharge it too much.
Doug: Do you suppose that the increased series R of the vintage capacitors might be due to a combination of heat, voltage/current and epoxy/Bakelite reacting to form an oxide or crud on the plates? Similar to cathode crud? Maybe lead attachment to plates junction?
Re: Dielectric Absorption
Posted: Tue Mar 27, 2018 7:48 am
by emrr
It had not grown anything. Here's an article about those bias cells:
http://www.antiquewireless.org/uploads/ ... s_cell.pdf
I really don't know the mechanisms at play in the old caps. That particular piece I would easily believe was powered on most of 1939-1988 in a radio station. Then it sat in an abandoned building with no HVAC for another decade.
Re: Dielectric Absorption
Posted: Tue Mar 27, 2018 8:24 am
by mediatechnology
Thanks Doug. Fascinating read. I've never seen one.
Amazing that Carbon/Zinc chemistry could last that long without eating away the container.
500 mV or so across the seals in my kitchen faucet are enough to eat it alive in 2-3 years. (
viewtopic.php?f=11&t=406&p=11033)
Getting C cells to last in the car Mag-Lite (which fortunately is never used) to last more than a year or two is tough.
I need a Duracell 504 15V and found that Exell were making a lot of vintage battery replacements.
I don't suppose there is enough of a demand for those but Exell is a good resource.
http://www.exellbattery.com/