How do I measure high capacitance values???

Discussion in 'Electronic Basics' started by royalmp2001, Apr 18, 2006.

1. royalmp2001Guest

I need to measure the capacitance of electrolytics or banks of
electrolytics ( say upto 50,000uF) fairly accurately.

I have a capacitance (and frequency) range on my dmm but it does not go
high enough...and I don't want to invest in an expensive Fluke meter
particularly.

I am getting hold of a used 20MHz scope soon, if that is of any help.

How can I do this without great expense.. maybe some simple circuit I
could build to enable me to do this?

Thanks

2. Tim WilliamsGuest

Wire up a constant current source using two diodes, two resistors and one
transistor. Connect the CCS to the capacitor, making sure it's discharged
first. Apply power and measure how many volts the capacitor gains in say,
10 or 100 seconds. If voltage just stops altogether, you've probably ran
out of voltage -- you can't charge it more than the volts the battery can
supply -- so pick a shorter time. There's nothing special about picking a
voltage or time limit. Capacitance is then:
I * t
C = -----
V
Where V is the voltage change (if it started at zero, then it's the voltage
you measured at the time t), I is the charging current (Google for a
constant current source diagram and how to set it for some current) and t is
the charging time.

For V in volts, I in amperes and t in seconds, C is in farads. If I is in
miliamps, C is in milifarads (1mF = 0.001F = 1000uF), and so on.

Tim

3. Greg NeillGuest

[Posting order rearranged]
Why not just charge the thing to a given voltage, connect a
known resistive load, and measure the time to reach (essentially)
zero volts. That should be about 5 time constants (R*C).

For more accuracy, measure the time it takes to drop to 1/2 the
starting voltage. The time constant R*C is then R*C = t/ln(2),
where ln(2) ~ 0.693.

4. JamieGuest

i bought a cap meter thats good up too very large for about \$39 bucks
if memory serves ? it has digital read out with zero and will go down
in the low PF's

5. redbellyGuest

Here's a relatively easy method that I've used for resistors that are
out of range of my DVM. It's not the most accurate, but it is easy to
set up.

1. Find a capacitor near the max reading your meter will handle, and
measure it. Let's call this value C1.

2. Next, put the unknown capacitor (C_unknown) in series with the
measured resistor, and measure the combined capacitance C_combined

3. The unknown cap is:

1/C_unknown = 1/C_combined - 1/C1

Notes:

You won't get the full accuracy of a direct measurement, and probably
lose a digit of accuracy for every factor of 10 that C_unknown exceeds
C1. So if C_unknown is 1000 times larger than meter's range, this
won't work very well.

To use this technique on resistors, the combination should be wired in
parallel.

Mark

6. tekamnGuest

Given a DMM with Ohm range, I use this method:

1. select a low range, e.g. 200 Ohm or 2k.
2. connect a cap with known capacity to the positve and negative input
lead of the DMM. For your caps I would take a 4700 or 10000 uF. Place a
short over the black and red lead. Note: Most DMM have "+" at the black
lead, and "-" poalrity at the red lead -- take care of this and check
first for your DMM (e.g. with a seprate voltmeter).
3. open the short and check time in seconds (your wristwatch is fine
for this) for the time it takes from 2 DMM diplay values: Example: I
start time measurement at a display "1.000" and stop at "5.000". Note
the time on a piece of paper, I refer to it as t3 ;-)
4. redo steps 2 and 3 with the unknown cap. Note time you get now: t4.
5. do the math: time noted from step 4 divided by time noted from step
3, result multiplied by your value of cap from step step 2:

C_unknown = ( t4 / t3 ) * C_known.

This works well for me for caps between 1000 and 100000 uF since many
years in field service. Method fails for caps with high leakage current
and/or shorts.

Another circuit I use in the homebrew lab is a simple astable
multivibrator (standard TL071 with discrete npn/pnp transistor follower
just to get 100 - 150 mA output current capability): Since
multivibrator frequency is 1/t, and t = const * R *C.

When you choose R to useful value, the time t is directly in uF (as can
be displayed on the digital freq/period counter switch to period time
display). Note: differences beteween time for astable output in pulse
or period ("high" resp. "low" voltage) gives a rough indication of caps
leakage current. All you need is a lab pwr supply, digital multipurpose
counter, opamp & a few transistors ;-))

hth,
Andreas

7. Rich GriseGuest

A calibrated power supply, a 1Meg resistor, a voltmeter, and a stopwatch. ;-)

Good Luck!
Rich

8. JeffMGuest

upto 50,000uF
Make that a calandar.

9. redbellyGuest

Maybe the best suggestion so far, except that I'd use a 1k resistor.
Why wait hours when the measurement could be done in less than a
minute?

Mark

10. Bill BowdenGuest

A calibrated power supply, a 1Meg resistor, a voltmeter, and a stopwatch. ;-)

A 1 Meg resistor is too high for a 50,000uF cap without a very high
supply voltage. The cap leakage current will be a large portion of the
total, and you will get a big error. Maybe use a 1K resistor for better
accuracy and also less test time.

-Bill