Homemade ESR meter

My ESR meter, assembled
Meter with "crude hand drawn scale." Unfortunately the calibration resistors I used made the scale look linear somehow.

assembled second version
Second/"Butterdish" 3d printed case ESR meter with crappy 200µA meter. "Tuning" scale markings are deceptively close if you take the number and divide by 2 (i.e. if you read 4, the ESR is actually 2 ohms).

The inductor in this one was self wound using the input filter of a dead ATX PSU, I used 90 turns of #30 for the primary (new) and 9 turns of #26 (scavanged from original winding on the core) for the secondary.

About the ESR meter

This is my high frequency (about 60KHz) AC ohmmeter, which is based off of Manfred's ESR meter. Since I get really frustrated at point to point on perfboard I designed a PCB for the circuit. Believe me I had enough wirewrap and perfboard for a while. (And I just can't get myself to do deadbug proto. Don't know why but it's just not my thing.) I suspect if you prefer using perf board, you can use the PCB layout as a guide. I found a way to lay out the board without jumpers.

Gerber files available are here, xpcb source is here.

The Op Amp

Note that I also used a TL082 instead of a TL062 or even TL072. It still works. Since it's in a socket I'll see what happens with other op amps such as LM358, LM1458, and maybe even an LM6142 RRIO which is probably overkill. Why did I use a TL082? because I have a bunch of them... And a lot of LM1458's... Maybe I should have used dual 741's because I have a boatload of those along with LF355 and LF356... I wish I had a few LM358 and LM324, these op amps should be pretty good. Only a few LM6142's unfortunately. I think I may also have a TS512AN's... I have a few RRIO TLV2374 too...
The LM1458 oscillated at a lower than expected frequency and apparently the amplifier stage did not work. It figures, it's an ancient dual unit. It probably does not like being run at 5V. After noticing the LM358 not working well, the LM6142 will likely have the same problem. It wasn't until I compared the datasheets of my op amp stockpile until it was clear - the input impedence of the rail-capable opamps is not nearly as high as the non rail-to-rail op amps. The LM6142's inputs are just atrocious. I suppose you can look for other JFET or CMOS op amps like LF353 or (I don't have any CMOS op amps to suggest), just don't try to sub in bipolar op amps without circuit modification.

You do know what, after studying the circuit a bit more, it leaves a bit to be desired. Due to the amplifier biasing I would imagine that there's no reason why any op amp wouldn't work here, except for the fact that the AC voltage in the original design seems to be biased strangely, though it still works for the TL062/TL082/TL084. The other issue is the fact it runs off of 5V: for many older op amps, this is on the low end borderline of the usable operation region, hence not getting enough voltage swing to drive the meter.

The Detector amplifier

I had to tweak the gain of the detection, R8, with using the probably same T1 from a broken PC ATX power supply. With the approx 8:1 transformer I pulled out, I needed to use a 10K resistor for R8 (originally 39K) to get reasonable response. The higher gain made low resistances invisible on the 200µA meter no matter what the pot setting was... Also, if you're staring at the board and notice I have the wrong values on the board, you're right. Good catch, I'm just using parts from my junkbox (a vast majority of the parts are recycled!) and deemed some of the values noncritical.

Circuit notes

Please, PLEASE put R11 (the potentiometer) somewhere easily accessible from the outside. The circuit is extremely temperature sensitive for me (hey wait, doesn't that sound like bad capacitors?) and tweaking the resistor is needed fairly frequently. However I've noticed a lot of bad caps are really bad, and the meter error isn't such a big deal. Just make sure you short the two leads occasionally and know exactly what your "0 ohm" reading should be, and take it from there. Bad capacitors tend to show up considerably different than the "0 ohm" reading - as ESR rises, losses from charge/discharge increase generating more heat. This heat causes more electrolyte to boil off, and causes a vicious cycle, destroying the capacitor.

Though I used a 200µA meter in the second meter, it does not work very well, full scale is hard to attain with the circuit as-is. I was attempting a 1mA meter which won't work at all due to the limited output voltage swing of the op amp. So please stick with 100µA or lower meter. Perhaps increasing R8 along with increasing C6 may help.
Then again, I don't know about the ESR of the caps C2 (coupling capacitor on primary side of transformer) and C3 (not marked on PCB, but it's the capacitor coupling the probe, connected to R7 and R6). If these have high ESR, then the scale reading won't go very high... So, time to test the ESR of these caps... uhoh... Now how do I test the ESR of these caps... :-)


This is the PCB I used in the black meter, sans the variable resistor.


XPCB screenshot of the layout

Notes

As I drained down its first PP3/9V battery, the meter behaves very oddly and claims infinite resistance for a dead short when the battery is weak. Therefore check the battery when its behavior is unexpected. I measured the current draw at around 7mA, which is the sum of the LED (I tried to target around 1mA), quiescent power of the TL082 (4.5mA), and the dynamic power of the oscillator plus power burned in R5. Using a TL062 would definitely reduce power consumption...

Other uses

The ESR meter can be used to measure the ESR of AA and AAA NiMH and NiCd batteries, which also should have very low ESR. If the cells have over 1Ω ESR, those are bad cells. I'd even venture that 500 mΩ is bad too... As the circuit will discharge batteries (due to the dual 10Ω clamping resistors), don't leave it connected too long. Do not measure the ESR of 9V batteries or anything higher than 2 volts.

Recycling

Well, I reused a lot of parts to build my ESR meter. Recycled parts used:
about half the resistors
all the capacitors
TL082
LED
Diodes
Transformer (did not rewind!)
LM78L05

New parts
other half of the resistors
Potentiometer R11
PCB (It is custom, after all.)
IC socket


Updates:
I made a smaller, 3D printed case for the old "butterdish" meter as I despised the old case. I relocated the LED on board as a meter side light so I can see if I left the meter on as well as able to read the meter in subdued lighting. With the slide on lid I no longer need a screwdriver to replace the battery.