I got my grimy hands on a bunch of surplus caps and I want to measure them to make sure they work as advertised. I don't have an LCR meter and I can't buy one. So I figured I could measure them by discharging across a known resistor and counting the time to discharge. My problem is that I've had my head in the digital world for so many years that I don't know if I'm doing this the right way . 
Another thing that adds doubt is that nobody else seems to have done this, while there's been lots of talk about lcr meters.
here's how I did it: RC equals one time constant, which is roughly equivalent to 63% of discharge. I have a 12.5 v supply so if I charge to 12.5v then 4.6v would represent one time constant. So I charged it up to 12.5, then discharge using a 150 ohm 5% resistor, and count the seconds until it gets to 4.6v as measured on my multimeter. This gave some strange results as seen below in discharge time and calculated capacitance:
Cap 1: 12s = .08 F
Cap 2: 10.5s = .07 F
Cap 3: 11s = .073 F
Cap 4: 10.5s = .07 F
Sounds good except that the spec for these caps is 60,000 uF. How am measuring above spec? My multimeter could be off, but this is a 17% error. I did all of the readings at least twice, and found similar readings for other sets of caps. what gives?
thanks,
Evan
Another thing that adds doubt is that nobody else seems to have done this, while there's been lots of talk about lcr meters.here's how I did it: RC equals one time constant, which is roughly equivalent to 63% of discharge. I have a 12.5 v supply so if I charge to 12.5v then 4.6v would represent one time constant. So I charged it up to 12.5, then discharge using a 150 ohm 5% resistor, and count the seconds until it gets to 4.6v as measured on my multimeter. This gave some strange results as seen below in discharge time and calculated capacitance:
Cap 1: 12s = .08 F
Cap 2: 10.5s = .07 F
Cap 3: 11s = .073 F
Cap 4: 10.5s = .07 F
Sounds good except that the spec for these caps is 60,000 uF. How am measuring above spec? My multimeter could be off, but this is a 17% error. I did all of the readings at least twice, and found similar readings for other sets of caps. what gives?
thanks,
Evan
It is interesting that all your caps are measuring too high, by 13mF on average. It could be the caps or it could be a systematic error in your your measurment method.
Systematic errors are the accuracy of your stopwatch/clock and the accuracy of your DMM and the accuracy of your 150-ohm resistor. Additionally there is the display latency of the DMM.
Example using made-up numbers:
I would assume the accuracy of the DMM (once settled) is very high - 1% or better.
You need to check this, but your display latency may be up to 0.5s.
Then your personal latency (between adding the resistor and starting the clock and then seeing 4.63V and stopping the clock) could be 1s perhaps.
The 150-ohm resistor is rated at 5%. You didn't mention whether you measured its value using the DMM. So it could be as high as 158-ohms.
Looking at the worst case, Assume R=158 and the measurement delay is 1.5s, this would give a systematic error of +9.5mF. Add the 1% DMM error and this could be +10.1mF worst case.
Well, that still leaves 3mF (avg) unaccounted for. It is not untypical for such big caps to have tolerances of +/-20% and sometim,es skewed tolerances like +20%/-10%. Does it say anything on the caps? 20% of 60mF is 12mF - easily accounting for the error. But the error on all 4 is positive; this could be explained if they are all from the same batch or if their distribution is skewed.
I'm not familiar with possible ageing affects on capacitance. Perhaps a ye olde member can talk to this. 😉
Systematic errors are the accuracy of your stopwatch/clock and the accuracy of your DMM and the accuracy of your 150-ohm resistor. Additionally there is the display latency of the DMM.
Example using made-up numbers:
I would assume the accuracy of the DMM (once settled) is very high - 1% or better.
You need to check this, but your display latency may be up to 0.5s.
Then your personal latency (between adding the resistor and starting the clock and then seeing 4.63V and stopping the clock) could be 1s perhaps.
The 150-ohm resistor is rated at 5%. You didn't mention whether you measured its value using the DMM. So it could be as high as 158-ohms.
Looking at the worst case, Assume R=158 and the measurement delay is 1.5s, this would give a systematic error of +9.5mF. Add the 1% DMM error and this could be +10.1mF worst case.
Well, that still leaves 3mF (avg) unaccounted for. It is not untypical for such big caps to have tolerances of +/-20% and sometim,es skewed tolerances like +20%/-10%. Does it say anything on the caps? 20% of 60mF is 12mF - easily accounting for the error. But the error on all 4 is positive; this could be explained if they are all from the same batch or if their distribution is skewed.
I'm not familiar with possible ageing affects on capacitance. Perhaps a ye olde member can talk to this. 😉
OK, well I'm old. Two things: first, e-lytics do tend to measure high. Second, the capacitance is a bit higher before they are formed. You didn't say, but I'll bet they haven't had anything like their rated voltage across them for quite some time. And when e-lytic caps are used at less than WVDC for a time, the value does tend to drift upwards.
In any case, if you're going to use these for a PS application, some extra uF aren't going to hurt you.
In any case, if you're going to use these for a PS application, some extra uF aren't going to hurt you.
traderbam - good analysis, and you brought up one point that I hadn't considered, which is delay of the meter. In fact I'm using a cheap analogue meter and I have no idea how accurate it is. I think analogue may be better for this application because its easy to watch the needle and anticipate when its going to hit the right spot. My resolution is better than 1 second whcih I was able to show because when I repeasted the test for each I didn't look for the unit number on the cap until after I took the measurment (an aside is that I found all of the big caps have numbers printed on the top which makes it really convenient to do things like this. its a nice touch). so it was a blind test 😎
now to your point - I don't know how much delay there is in the needle speed, although it doesn't have to move very fast with this resistor. I could increase resistance and see if it changes anything, and that would take meter delay out of the equation. I measured the resistor and got spot on 150 as best as I could tell, but it fell on a part of the meter where its hard to interpolate exactly, and of course I'm sure the meter aint perfect.
20% error rating is nice to know too, and as long as my procedure is within reason, then I'm happy.
Sy - you are correct in that they've probably been sitting in the shop for a long time without a charge. I'm going to see if I can pick up a cheap powersupply today and charge em up better.
Its very interesting to hear that C creeps up if they haven't been charged for a while. Thats the kind of thing I was hoping for in this thread and this group never ceases to amaze 🙂
One more thing is that they are not from the same lot - they're the same manufacturor and same ratings, but slightly different model number (2 of each) and one set is heavier than the other. thankfully cap #1 & #3 belong to one set while #2 & #4 belong to the other (they have the most similar measured C) The fact that I got em for a buck fifty a piece makes the difference in model number not so significant
I also measured some others and this is what i got:
- at least one pair came in under spec but I don't know if I should be pleased or not
rated 14,000 uF:
1: measured 16,000 uF
2: measured 14,000 uF
rated 47,000 uF:
1: measured 44,000 uF
2: measured 42,000 uF
I'm happy because it seams at least that I haven't lost too much basic knowledge. 😎 just don't ask me to do anything more complicated unless its all 1's and 0's
I'd buy a better meter but I'm applying to grad school so why should I buy a meter when I can use one in the lab at school for the next seven years or so...
now to your point - I don't know how much delay there is in the needle speed, although it doesn't have to move very fast with this resistor. I could increase resistance and see if it changes anything, and that would take meter delay out of the equation. I measured the resistor and got spot on 150 as best as I could tell, but it fell on a part of the meter where its hard to interpolate exactly, and of course I'm sure the meter aint perfect.
20% error rating is nice to know too, and as long as my procedure is within reason, then I'm happy.
Sy - you are correct in that they've probably been sitting in the shop for a long time without a charge. I'm going to see if I can pick up a cheap powersupply today and charge em up better.
Its very interesting to hear that C creeps up if they haven't been charged for a while. Thats the kind of thing I was hoping for in this thread and this group never ceases to amaze 🙂
One more thing is that they are not from the same lot - they're the same manufacturor and same ratings, but slightly different model number (2 of each) and one set is heavier than the other. thankfully cap #1 & #3 belong to one set while #2 & #4 belong to the other (they have the most similar measured C) The fact that I got em for a buck fifty a piece makes the difference in model number not so significant
I also measured some others and this is what i got:
- at least one pair came in under spec but I don't know if I should be pleased or not
rated 14,000 uF:
1: measured 16,000 uF
2: measured 14,000 uF
rated 47,000 uF:
1: measured 44,000 uF
2: measured 42,000 uF
I'm happy because it seams at least that I haven't lost too much basic knowledge. 😎 just don't ask me to do anything more complicated unless its all 1's and 0's
I'd buy a better meter but I'm applying to grad school so why should I buy a meter when I can use one in the lab at school for the next seven years or so...
- Status
- Not open for further replies.