EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: intabits on August 21, 2021, 08:14:31 am
-
A simple device I've made recently to aid with reforming electrolytic capacitors and testing their leakage current. It may be of interest to some.
It doesn't measure capacitance or ESR, or any of the other more exotic capacitor metrics. It's basically just a switch, 3 resistors and an ammeter. The switch simply eliminates the need to mess around with clip leads while doing these tests, and the built-in ammeter is a convenience.
I've already found it handy for testing a bunch of electrolytics for the power supplies in my PDP computers. And it has/will help with my understanding of capacitor behaviour for when I try to make an automated reformer tester for the large number of electrolytics I have out in the shed...
(https://i.imgur.com/ukgAnVv.png)
Its insides. The PCB over the meter includes meter protection diodes, and a switched shunt to provide a 50mA range, in addition to the 50uA range of the meter.
(https://i.imgur.com/UuJwNkM.png)
The schematic and some notes.
(https://i.imgur.com/fvVdgww.png)
Due to the higher resistance from the voltage source when the meter is in the uA range, the reform/charge current may be too high to stay in the uA range.
If I'm interpreting it correctly, the charge resistor and meter form a voltage divider with the (too low) leakage resistance of the capacitor. With the lower resistance of the mA range, the divider voltage still allows the capacitor to charge to near the power supply voltage. But in the uA range, the new divided voltage is lower, and the cap discharges down to it. This puts enough voltage across the meter to push it past 50uA.
With better capacitors that have less leakage, the meter can stay in the uA range, and the current drops to some final value of (I've seen around 10uA) as the cap reforms.
Video
https://www.youtube.com/watch?v=L-u8JhVgQlY (https://www.youtube.com/watch?v=L-u8JhVgQlY)
-
Looks nice. Each time I read/hear "forming a capacitor" (German: "einen Kondensator formieren") brings me back to engineering school in 1992, when we would call this "format c:" :-DD
-
Nice project! Such a device can be good for bringing back old equipment and avoiding smoke and damage on first power-up. That said, I've found that many/most very old caps will come back to life when reformed, but deteriorate much more quickly than they should, or never reach their original values. Beware old caps that are significantly above their marked value. It means that something has changed internally, though I don't know if it predicts a higher likelihood of failure. They seem to go for quite a while in that condition. Still, a handy and valuable thing to have around. Note that some manufacturers of aluminum electrolytics specifically recommend reforming caps that have been on the shelf a while, before measuring or using them.
-
Nice project! Such a device can be good for bringing back old equipment and avoiding smoke and damage on first power-up.
Thanks!
That said, I've found that many/most very old caps will come back to life when reformed, but deteriorate much more quickly than they should, or never reach their original values. Beware old caps that are significantly above their marked value. It means that something has changed internally, though I don't know if it predicts a higher likelihood of failure. They seem to go for quite a while in that condition.
Yes, something to watch.
After reforming, and checking leakage, I'd want to test the capacitance and ESR also. Ultimately, I want to make a more comprehensive automated contraption to do that as well.
I used the simple box shown here to check that the caps hold a charge after reforming. I've seen caps (very old ones in a tube radio) appear to reform OK, but would not hold a charge. So checking charge-holding should weed out a few duds that get past simple leakage testing.
Note that some manufacturers of aluminum electrolytics specifically recommend reforming caps that have been on the shelf a while, before measuring or using them.
Yes, even after as little as 12 months!
-
Looks good. I've never had to do enough reforming to put together something like this, but it would certainly be convenient to have.
I'm not sure that the 50 ua range is necessary. When reforming a capacitor, by the time you're down to the bottom end of the 50 ma scale, the process is complete - particularly if you're reforming at the capacitor's rated voltage. Any circuit you use it in will be at a significantly lower voltage.
-
I'm not sure that the 50 ua range is necessary. When reforming a capacitor, by the time you're down to the bottom end of the 50 ma scale, the process is complete - particularly if you're reforming at the capacitor's rated voltage. Any circuit you use it in will be at a significantly lower voltage.
Thanks. The uA range is a cheap and simple feature: just a switch, trimpot, and a few resistors.
I think it's of value because firstly, leakage current specs are nearly always quoted in uA.
And secondly, it enables some discrimination between "very good" and "fairly good" (for want of better classifications). Both VG and FG will show zero leakage in mA. And when the meter is switched to uA, they both might for example, initially show say 30uA, but the FG will quickly rise to beyond 50uA, whereas the VG will gradually drop to 10uA or less. (At least that's my interpretation of what's happening)
-
I reform caps on a GR 1617-A bridge. The lowest current scale is 60 uA, which I find is about perfect. The best caps will be down near the peg and most can get below 20 if you wait a while. I'd definitely miss that scale if it wasn't there. (I hate to use "u" for micro, but I don't know how to get the right character."
-
I'm not sure that the 50 ua range is necessary. When reforming a capacitor, by the time you're down to the bottom end of the 50 ma scale, the process is complete - particularly if you're reforming at the capacitor's rated voltage. Any circuit you use it in will be at a significantly lower voltage.
Thanks. The uA range is a cheap and simple feature: just a switch, trimpot, and a few resistors.
I think it's of value because firstly, leakage current specs are nearly always quoted in uA.
And secondly, it enables some discrimination between "very good" and "fairly good" (for want of better classifications). Both VG and FG will show zero leakage in mA. And when the meter is switched to uA, they both might for example, initially show say 30uA, but the FG will quickly rise to beyond 50uA, whereas the VG will gradually drop to 10uA or less. (At least that's my interpretation of what's happening)
Yes, I agree with what you've said. If I had buit your circuit I probably would have added the 50 ua scale too. However,
First, if the capacitor is old enough that it needs to be reformed, it's not appropriate to use datasheet leakage values. There are various things that can cause low levels of leakage that are related to aging and are unaffected by reforming.
Second, a "fairly good" capacitor will likely become a "very good" capacitor with use as the last tiny bit of the reforming process completes. Even if it doesn't, it won't affect the operation of the circuit.
Third, at current levels in the ua range, you're not actually reforming at all, you're just seeing dielectric absorption which is a totally seperate phenomena that is inherent to all capacitors at differing levels. To quote from Wikipedia ( https://en.wikipedia.org/wiki/Dielectric_absorption ) :
Another effect of dielectric absorption is sometimes described as "soakage". This manifests as a component of leakage current and it contributes to the loss factor of the capacitor. This effect has been known of only recently: it is now a proportionately greater part of leakage current due to the significantly improved properties of modern capacitors.
Remember that the purpose of reforming is to regrow the oxide layer that allows the capacitor to tolerate the applied voltage without being destroyed. After reforming, you don't have a new capacitor - you have an old capacitor that can be reused for general-purpose applications. I would never use a reformed capacitor in a precision application where leakage current, ESR, dielectric absorption, etc. were significant factors. Yes, I realize that electrolytic capacitors would rarely be used in situations like that, but it does happen.
Ed
-
The uA range is a cheap and simple feature...
Duh! Just realized that I didn't add the µA range, I added the mA range. Depends on what meter you have.
But I wanted to see µA, so I used a 50µA meter, and then adding the mA range becomes almost mandatory.
Conrad: See what I did there? You just hold Alt and type "230" on the numeric keypad.
if the capacitor is old enough that it needs to be reformed, it's not appropriate to use datasheet leakage values
That seems reasonable for truly old caps, but mind that even a year is recommended as old enough for reforming.
That wikipedia quote is interesting. I'm surprised to see that DA causes a continuous current. I would have thought that once the dielectric had "absorbed the charge", its contribution to current flow would stop.
BTW: Just had a Sprague "powerlytic" 16000µF 15V cap that looks to be from 1979 on "fast charge" (via 47R) overnight.
It was reformed a few months ago, and I'm just checking it again.
Switched to "slow charge" (via 3.33k) to see the current, and it was initially zero µA!
It has slowly crept up to 4µA, and stayed there. I'll put that one down as "very good".
-
... µA ...
Conrad: See what I did there? You just hold Alt and type "230" on the numeric keypad.
now that is an interesting piece of trivia. The German keyboard has the µ right there on the M-key. You access it by <Alt>-<Shift>-m or <AltGr>-m. Somehow it is still ingrained in my brain that foreign computers have issues with anything beyond the 7bit ASCII characterset - so I tend to avoid those special characters in source code comments or international exchanges. At the same time I am quite comfortable with switching my Görmän keyboard to Japanese IME and prodücing UTF-8-encoded texts - 不思議な世界ですよね...
Sorry for not contributing to the main topic, but I find this kind of stuff remarkable.