4
\$\begingroup\$

I have questions about the limitations of using a mains-rated power transformer (e.g. a small 6VAC 1VA PCB mount transformer) as a cheap differential probe, so that waveforms can be safely observed on an oscilloscope.

My questions refer to this schematic, which is not intended as a real circuit but to allow me to refer to the various elements:

schematic

simulate this circuit – Schematic created using CircuitLab

Vin1 and Vin2 are any two points in a mains-powered circuit across which the differential measurement is desired. R1 represents a standard scope input impedance. For now let's ignore R2, C2, C1, R4, and R5.

My understanding is that with an ideal transformer, this is a safe way to accurately display a typical mains voltage waveform. However, a real transformer will have some common mode leakage represented here by C1. This should mean that the transformer's "CMRR" (as it were) deteriorates as higher and higher frequency content appears across it.

  1. My main question is: will this common mode issue present a danger to an oscilloscope when probing a circuit like a phase angle controller with fast edges? Will the high voltage, high frequency content be capactively coupled straight across the transformer, resulting in a short via the shell of the scope's BNC connector?

  2. If this coupling could be a problem, would the addition of a simple filter like R2 C2 suffice to solve it, assuming that the user did not mind the loss in measurement bandwidth?

  3. Are resistors like R4 and R5 a sufficient way to minimize loading on the circuit being measured?

\$\endgroup\$
  • \$\begingroup\$ 1.) N but depends on all values and n ratio, 2.) maybe but depends on measurement BW needed. transformers at best , most do around 3 decades max. 3.) maybe but depends on Z ratio and acceptable load regulation error \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jan 8 '17 at 9:30
  • \$\begingroup\$ Are you referring to measuring the signal using two channels, inverting one channel, and adding them? \$\endgroup\$ – replete Jan 8 '17 at 16:20
  • \$\begingroup\$ @stretch thanks for the comment, just to clarify: I understand that solutions such as high voltage differential probes give better results here. The question is about the technical limitations of the simple transformer method ("Why can't we just do this?"), not about whether there are better methods. \$\endgroup\$ – replete Jan 8 '17 at 16:29
4
\$\begingroup\$

Firstly, you cannot expect ideal transformer operation unless large secondary current is drawn; you must make a LOW resistance load on the secondary to accomplish this (the transformer core otherwise sees large magnetic field fluctuations, and that causes distortion of the waveform).

There is no significant danger from common mode signal leakthrough, if the transformer is intended for AC power (the 'ground' pin is just as likely to give you glitches when lightning strikes nearby). Oscilloscope probe capacitance is comparable to (or greater than) the expected C1 capacitance. It may matter WHICH pin of the secondary is grounded, however; experiment with a function generator connected common-mode to the primary winding wires. C1 might be on the order of 50 pF. The secondary winding and load impedance, unless it is VERY high frequency, will be much lower and will dominate.

If you expect lots of feedthrough, though, there are isolation transformer designs which include shielding, and it is possible to ground the magnetic core of some common transformers (which will have a similar effect).

Filtering is usually a bad practice in measurement; most oscilloscopes handle transient overages gracefully.

R4 and R5 will reduce loading, but will also (slightly) filter the signal (more than you think, because at high frequencies, the transformer gets lossy). On the primary side, it wouldn't hurt to put a common-mode choke (since you only want the difference signal).

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
4
\$\begingroup\$

Assuming you have your transformer HV side to the circuit, 6v side to the scope ...

1) The current conducted via C1 should not cause a danger to the earthed scope. It is likely to add spikes to your readings. You can get mains transformers with an interwinding screen, designed to shield C1. Connect the screen to the scope ground, all the C1 current will flow into there, and none will turn up in the 6v winding.

2) R2 C2 will reduce the differential mode bandwidth, which is not necessary (see later).

3) R4 R5 will reduce the loading on the circuit, but are probably unnecessary. If the mains side is to the circuit under test, then there should be little need to use R4 and R5, as the impedance will be relatively high. A series capacitor might be useful for AC coupling, because although the AC impedance is high, the DC resistance is low, and the ability to make a measurement in the presence of the DC bias is useful.

An ordinary mains transformer is built on an iron core that's engineered to be low(ish) loss at mains frequencies, and very little more, that is, as cheap as possible. The frequency response / waveform fidelity may well be inadequate for components above mains frequency. Those sharp edges you mention will almost certainly not be visible, at least not in their original shape or amplitude.

People do use transformers to measure in the way you've drawn, but they tend to be ferrite or air-cored, if they are going for any real bandwidth.

\$\endgroup\$
  • 1
    \$\begingroup\$ I wonder if junkbox SMPS magnetics could be used for this job, given the extremely light loads involved... \$\endgroup\$ – ThreePhaseEel Jan 8 '17 at 15:29

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.