ESR meter

From Webshed
Revision as of 20:37, 9 January 2012 by WikiSysop (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

An Equivalent Series Resistance Meter

Some time ago I was trying to repair a switched mode power supply (SMPS) in my oscilloscope. I'd been quoted £800 for a new PSU, so trying to fix it first was well worth my while. It's pretty common for electrolytic capacitors to develop faults in an SMPS; specifically they develop an higher than normal internal resistance. So while the capacitor may hold a charge and measure as the correct capacitance, it will not behave correctly in a filter or PSU circuit. The ESR of a large high voltage, high capacity electrolytic capacitor should be fractions of an Ohm, smaller capacitors have ESRs of a few Ohms typically. As the capacitor degrades in use the ESR can easily climb to several hundred times the normal value, while exhibiting no changes to voltage and capacitance ratings

Suspecting the capacitors, and not having a way to measure their series resistance, I set out to design and build an ESR meter. We can't measure ESR with a normal DC Ohm meter, because the DC voltage would charge the capacitor. An AC signal is used, so that it passes straight though the capacitor with minimal reactive losses. As a ball-park figure, I decided my ESR meter design would operate at around 25kHz, where a 100μF cap has a reactance of about 60 mOhm (a 1μF cap has a reactance of about 6 Ohms at this frequency). I also decided to keep the test voltage below about 0.6 V, so as not to forward bias any semiconductors in the circuit under test. I want to be able to test capacitors in circuit where possible.


A three stage inverter RC oscillator starts much more reliably than a two stage oscillator. This provides a 25 kHz 0-5 V square wave, which is buffered by the other three gates in IC1, increasing the current output capability. R3,R4,R5 and R6 form a potential divider, so that only around 200mV is applied to the device under test - low enough that capacitors can be tested in circuit without turning on any semiconductors.

C3 provides DC isolation for the opamp, but don't rely on this to save you from a charged mains capacitor in a SMPS - always discharge before you test. You could also put a 10k resistor across the probes to help ensure capacitors are discharged - 10k is a much higher resistance than the meter can detect, so it will not affect ESR readings.

The opamp IC2, provides plenty of amplification for the AC signal passed though the capacitor under test, the output of the opamp is rectified by diodes D1 and D1 and applied to the meter. I use a 4.7k pot in series with the meter to set zero resistance with the probes shorted together for calibration. The exact value of the series pot will depend on the sensitivity of the meter used.

The circuit is powered by 4 AA cells, and draws <10 mA, so the batteries should last a long time. Board.png


Calibration I made a calibration standard from 10x 1Ohm 1% metal oxide resistors in series with two 10 Ohm metal oxide 1% resistors. With the meter zeroed with the probes shorted together, I measured first 1 resistor (1 Ohm) and marked the meter scale, then two resistors (2 Ohms), until I had marked up to 10 Ohms. Then 11 to 30 Ohms by adding 1 then two 10 Ohm resistors to the measurement.

Chops and Changes

There isn't anything difficult to obtain to build this meter, no coils or transformers to wind, no crystals to obtain. IC1 really should be a Schmitt trigger device, so the oscillator functions. Rectifier diodes D1&D2 would probably be better replaced with Schottky diodes, so there is less voltage dropped in the rectifier and more left to drive the meter. R8 and R11 set the amplification factor of the opamp, the values suggested work, but are not optimized, if you find you have insufficient gain to zero the meter, here is a good place to look (or use fresh batteries)


Zip archive containing Eagle CAD files for this project