Question I MUST ask: Caps - can they charge on their own?

Thread Starter

Tonyr1084

Joined Sep 24, 2015
7,853
I have a 1500µF 200V cap just sitting on the bench. I know that when I test a cap using a DVM the cap will slowly charge because the DVM provides a minute current. I've seen that for many years. However - and here's the part I find hard to believe - I have this cap sitting on the bench. Not connected to a darn thing. Leads face up (away from the conductive ESD mat). I just connected my oscilloscope to the cap. Immediately the line diverged from zero to off screen. It came back to zero in about 3 seconds. Scope was set to DC on the 0.1 scale. Was definitely surprised to see this happen.

The probe I'm using is a video balun connector with a cat-5 twisted pair of wires. Not your normal test leads. In fact, the whole reason for touching this to the cap was to see what would happen. I've already connected this test lead rig to a coin cell battery sitting at 2.9 volts. The scope reading was 2.9 volts as well as the DVM. I'm actually wondering about using this balun device as a means for connecting to my breadboard to test circuits - to see how it compares to regular scope cables. I have channels A & B so I can set up a test where I feed the same signal into both and observe any changes - IF any. I'll let you know how that turns out later. For now I'm stumped as to understand how a cap can develop some sort of residual charge. The last time I put this cap in circuit and removed it I made sure it was shorted. It's sat on the bench for about 9 months untouched. So any residual charge SHOULD have dissipated by now. So I'm stumped! Anyone know something I don't - regarding this phenomena. I know many of you know things I don't.

Anyway - thanks.

[edit] Following up on my balun cable test: The coin cell shows 3.2 volts via the standard scope probe and 3 volts via the bodged together balun cable. Doesn't look like I can expect to see any kind of accuracy. However, as far as seeing a signal (amplitude not withstanding) it should be a fair reading. Maybe I should pull the balun apart and see what's inside it. Likely some sort of line matching transformer. But then why the steady reading when checking a battery?
 
Last edited:

Thread Starter

Tonyr1084

Joined Sep 24, 2015
7,853
@OBW0549 Thanks. I'm going to connect a jumper to the terminals and let it sit for a few days to make sure it's fully discharged. Then I'll disconnect the jumper and just let it sit for a few months, and then get back to checking to see if it has developed a charge on its own. The link is helpful. Thanks.

Thanks @AlbertHall. Learned something I never knew before.
 

OBW0549

Joined Mar 2, 2015
3,566
I'm going to connect a jumper to the terminals and let it sit for a few days to make sure it's fully discharged. Then I'll disconnect the jumper and just let it sit for a few months, and then get back to checking to see if it has developed a charge on its own.
It will. It might only be a few dozen millivolts, but you'll see something.
 

AnalogKid

Joined Aug 1, 2013
10,987
CRTs were bad for this - the glass-aquadag could be painful even after being discharged.
Learned that in high school, working after school in an all makes/all models TV repair shop (that tells you how long ago this was). No one warned the new kid about this, so I "learned" it the old fashioned way. Same for high voltage electrolytics. Later in life, two new, large value computer grade electrolytics arrived with jumpers across the terminals. Didn't have to ask why.

ak
 

crutschow

Joined Mar 14, 2008
34,283
Maybe not for you, but we have more stories to tell ...
A (perhaps) interesting one.
A number of years ago I was tasked with identifying why a high resolution (12-bits if I recall correctly) A/D converter was sometimes giving small errors in its conversion output.
After some testing, I noticed that the output digital word was changing slightly for a few ms until it stabilized after a step change in the input voltage, and the change was larger for a larger step.
As you would expect, from the subject of this thread, the problem turned out to be dielectric absorption in the A/D input anti-alias filter.
A high dielectric-constant ceramic cap, (likely X7R) was being used, and they have a significant dielectric absorption, with a many ms time-constant release of the trapped charge, causing the error.
(A high dielectric material in a ceramic capacitor comes at a price.)
Changing to one with an NPO dielectric, which has a low dielectric-absorption, solved the problem.

So the lesson is, you don't use a cap with significant dielectric-absorption in the anti-alias filter of a high resolution A/D with less than several ms between conversions.
 
Last edited:

Thread Starter

Tonyr1084

Joined Sep 24, 2015
7,853
Maybe not for you, but we have more stories to tell ...
And more fun to be had!

Just for kicks, I'm gonna do some devious experiments on innocent electrolytic capacitors...
Would be interested in hearing your stories and experiments. I'm going to go disconnect the jumper from my big electrolytic cap (which I don't think I mentioned its type before)
 

MisterBill2

Joined Jan 23, 2018
18,176
Bob Pease, that incredibly brilliant engineer, wrote quite a bit about dielectric absorbtion, storage, and recovery. My recollection is that the capacitor will recover charge a bit if it was ever charged. I trust that Pease was correct.
 

OBW0549

Joined Mar 2, 2015
3,566
Would be interested in hearing your stories and experiments. I'm going to go disconnect the jumper from my big electrolytic cap (which I don't think I mentioned its type before)
I've got a 100,000 μF 35 WVDC electrolytic which has been "soaking" on the output of a 35 volt power supply for the last 2 hours. I'm going to let it go for another 3 hours and then try discharging it for 15 seconds through a 10 Ω resistor, which should bring its voltage down well below a millivolt. Then I'll disconnect the resistor and watch as the voltage (if my expectations are correct) climbs back up. If anything dramatic occurs, I'll post it here.
 

OBW0549

Joined Mar 2, 2015
3,566
Bob Pease, that incredibly brilliant engineer, wrote quite a bit about dielectric absorbtion, storage, and recovery. My recollection is that the capacitor will recover charge a bit if it was ever charged. I trust that Pease was correct.
Here's one of Pease's columns in Electronic Design on dielectric absorption, a.k.a. "soakage."
 

OBW0549

Joined Mar 2, 2015
3,566
If anything dramatic occurs, I'll post it here.
Not exactly dramatic, but an hour after being discharged through the 10 Ω resistor for 15 seconds, the voltage is now back up to 2.2 volts.

EDIT: After sitting overnight, the voltage has risen to 2.6 volts-- 7.5% of the voltage it was originally charged to.
 
Last edited:
Top