LCR meters and ESR

[Floyo]

Some time ago I bought the Agilent U1731B LCR meter. And of course I would like to be able to measure ESR. I thought that measuring the resistance of a capacitor in series mode measurement would yield the ESR of the cap. After reading a post in the BK precision LCR meter review topic I wasn't to sure that my theory added up. So what do you do when you don't know weather something is correct or not, you experiment .

http://wiki.xtronics.com/index.php/Capacitors_and_ESR

On this page I found a formula that would let me calculate the ESR with the Q and impedance of a cap. So what I did is measure the Q of the cap, the value and the resistance in series resistance measurement mode.

Here follow the results

Code: [Select]

Q C F Zc ESR     Measured series resistance
14,4 9,44E-004 120 1,405 0,098 0,096
4,76 3,77E-003 120 0,352 0,074 0,065
7,4 4,04E-003 120 0,328 0,044 0,043
13 6,15E-004 120 2,157 0,166 0,166
11 4,13E-004 120 3,214 0,292 0,302
14,8 8,60E-005 120 15,418 1,042 1,008
9,4 9,14E-005 120 14,514 1,544 1,505
16,8 4,47E-005 120 29,678 1,767 1,746
31,7 4,52E-005 120 29,317 0,925 0,909
78,8 9,70E-008 1000 1640,773 20,822 18,03
172 1,04E-007 1000 1536,245 8,932 9,8
235 9,97E-005 1000 1,596 0,007 2,146

So as you can see, most of the measurements and calculations match up pretty well, until I get to the Film-caps of 100nF where the calculated ESR is extremely low (high Q) but the measured resistance is quite a lot higher.

So my questions are, Does the measuring of the resistance of a cap in series mode on an LCR meter give you the ESR?
And, of course, Why does the calculated ESR differ so much from the measured resistance once I measure 100nF foil type caps (I measured a few more then are displayed in the table, and they all give about the same result)?

Thanks in advance.

[tyblu]

What is the accuracy and precision of that meter for Ohms in that range?

[Floyo]

In the 20 Ohms range, the specs are as follows, 1.2%+40 counts at both 120 and 1Khz. This is in parallel mode, series mode is not included in the specifications.

So even in a worst case scenario the 2 Ohm's on the last cap is way out, 1.2% of 0,007 is 0,000084. The maximum value that the meter can display is 19,999 Ohms, so it should read a maximum of 0,047 Ohms. In practice the value is about 50 times higher. I don't expect it to read these very low values accurately, but I do think it is quite strange because it measures 1 Ohm 1% resistors just fine.

[tyblu]

I don't think the values would be the same. Series resistance (ohmmeter) would expect a frequency independent load, and may use a very low or DC voltage to probe it, while ESR assumes a cap or inductor, and will seek the proper frequency (~kHz for caps, ~DC for inductors). I am not sure, though, as your meter is made by people much quicker than I.