I've recently found the need for a ESR meter(My daughter and her boyfriend brought me 8 broken LED/LCD TVs and 6 have power problems).

I also decided to go all out and build a full blown ESR meter. The foundation for the meter is going to be the STM32F373 which has 3 16bit Sigma Delta ADC which are capable of simultaneous differential conversion. It also has DSP capabilities so I was going to use a DFT to determine impedance and go from there.

So I'll measure voltage across a test resistor, across the Cap under test and current. I want to use both variable voltage and frequency. I thought to use .2V to 1V for voltage and 100Hz to 25Khz for frequency. The test signal will be generated using one of the internal DACs filtered to a sine wave and buffered(for voltage and current control, including constant current(Edit to make it obvious)). I think I'll just output all this info via serial port and use a program to display and control. I can always add a interface later.

Sorry for all the leading information but I wanted make sure the question could be answered. With my brain though I've most likely forgotten something.

My question is under these circumstances which method of measuring current is better, using a shunt resistor or a current transformer. I've looked at a large assortment of various types of meters and just can't make up my mind.

I have a couple of IN Amps, a large assortment of op-amps and passive components available to me.

  • 1
    \$\begingroup\$ The foundation for any design is your design specification and test method, not the parts. An LCR meter uses a constant current source for sine wave at various frequencies to measure voltage and phase shift. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 May 11 '19 at 14:36
  • \$\begingroup\$ Fundamentally, the ESR is the quadrature component to the charging-current. Hence a precision ESR must sustain clean I and Q extraction. How to specify that? \$\endgroup\$ – analogsystemsrf May 11 '19 at 15:17
  • \$\begingroup\$ @SunnyskyguyEE75 My DAC output buffering will allow a choice of constant voltage or constant current. I thought I was clear in how I am testing with 2 voltage tests and a current test. My problem is choosing the best way to test the current. \$\endgroup\$ – GB - AE7OO May 11 '19 at 17:07
  • \$\begingroup\$ And please note, all 3 SD ADC can be triggered at the same time. I was planning on first running a small test to determine zero crossing and then take 10 measurements per frequency. I'll have choice of doing a single freq or a span. \$\endgroup\$ – GB - AE7OO May 11 '19 at 17:23
  • \$\begingroup\$ Usually phase and amplitude of output voltage with auto ranging sine current determines the mode of RLC then reactance in F or H and resistance in Ohms. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 May 11 '19 at 23:55

Sorry for the delay.

After testing both methods using various shunts(.01, .05(1%, .5%, no idea) and CT's(4 hand wound and 2 commercial (small value burden and tried with both with no gain and gain))), I'm going to go with a shunt. The CT's ended up having various phase errors that were in 1 case very non-linear(but that could have been my error, it was one of the home made ones).

The way I tested the them was:

  1. Using a Function Generator set to 100Khz and then setting up a sweep from 1Khz to 100Khz with slow steps and .25 to 3vpp.

  2. For a reference resistor I used a couple of old 1 and 220 ohm .5% mil-spec resistors (still on the reel) I picked up at a ham-fest years ago. Through hole, blue and lots of bands..:)

  3. The best I could do for the DUT were a some 5% film caps 104J and 163J. I was too lazy to make any inductors(The only ones I have are either low-Q crud or much better rf air core(I guess I could have taken one of the tank coils out of an amp, but that would have been overkill plus work), neither of which I thought would be helpful here)

Then I would observe the wave forms using my VDS1022(25Mhz) (I used a Kikusui 5060(60Mhz) to confirm that the DSO traces at least resembled reality) Then a screen cap was made.

After comparing them the shunts were all in line, meaning no surprises. The CT's were not bad, meaning they were fairly consistent within a device, but when compared to each other or the shunts, there was differences(Would the term jitter also apply here??).

Since I'm easily able to control by excitation and gain, the shunts just make it much easier without having to play follow the leader trying to tune a CT circuit or apply corrections in software.

Please note, the final application will end up being a LCR/ESR meter. I doubt I'll ever go over 50V or 1A. On the other hand I think if your dealing with trying to measure mains the inaccuracies of a CT really do not matter. The simple fact is that the chance of your voltage/current wave form being nicely symmetrical, of whatever shape, is so close to 0 as not to matter, meaning a little bouncing around really will not affect the measurements. And using a CT or Rogowski coil makes it much easier to measure your 120V+ at whatever current.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.