English

Measuring a rectifier with no load

Digital multimeters

By Chuck Newcombe

The 189 was the first Fluke handheld DMM that offered two ways to measure AC volts. In the Vac function, the measurement was essentially the same as most Fluke DMMs, but the 189 also allows you to measure ac volts while in the Vdc function. It's part of the system that allows the meter to display the True RMS value of a combined ac and dc signal. I knew it would only be a matter of time before someone compared the readings between the two AC functions, and it wasn't long before someone did.

It was in late 2000, shortly after the introduction of the Fluke 189, when a customer called in with a question about measuring ac ripple voltage at the output of a bridge rectifier circuit capable of producing currents of 20 Amps, or more. And, sure enough, he had taken the readings in both modes.

He found a small difference in the readings, and called us to see if something might be wrong with his meter. I told the person who took the call to find out if the difference in the readings was more than the combined specifications of the functions would allow. I also asked him to get the details of the circuit being measured, and the values of the measurements that he had taken, so we could review them and attempt to verify the differences in our engineering lab. The reading difference was much greater than normal tolerances would allow, so I had to put on my thinking cap.

It turns out the circuit being measured was the output of an inexpensive battery charger. I guessed that there was little more than a transformer and a bridge rectifier in the box. If that were true, and there was no load connected to the output, I was pretty sure we could get almost any reading we wanted. Why? Well, it turns out that there are leakage capacitances and resistances across the diodes that can present interesting results when the only load is a high impedance meter. We spoke again with the caller and asked him to connect a load (battery to be charged) across the output and then try his readings again. He did, and the results were more reasonable and consistent.

Now it was time to look at the differences in the two measurement modes and ferret out the explanation for the seemingly odd meter behavior. We did that, but since it's been nearly seven years, I have actually built a representative replica of the charger circuit (at great expense - about $25.00) and recreated the tests for this column - so I could have the test results fresh in my mind. I didn't have a 189 handy, so I used two other meters - an older Fluke 77, and my trusty 87-V.

Using the 87-V, I first measured the ac voltage at the output of the step-down transformer. It was 14.15 volts. I then measured the ac at the unloaded bridge rectifier output - about 0.7 volts - so far so good. Now, I connected the input of the 77 in parallel and set it to the V DC mode. The 77 read 4.4 volts DC - not what you'd expect for what was intended to be a 12 V charger, while the 87-V Vac reading dropped to about 0.4 volts. Then, when I added a 20 kohm resistor load to the rectifier output the readings all settled down to what one might expect. So, what's going on?

Well, the Vac input of most Fluke DMMs is a 10 megohm resistor in series with a large capacitor to block DC. The Vdc function of the same meters is the 10 meg resistor alone - no series capacitor, so it presents a different load to the rectifier circuit - one that allows a small amount of dc current to flow. That difference when presented to the leakage capacitance and resistance across the bridge rectifier diodes creates an interesting waveform - one that will apparently produce the strange results I saw.

The final part of my experiment utilized two other measurement functions of my 87-V. I disconnected the bridge rectifier from the transformer and measured the capacitance across the junctions. It was somewhere around 0.6 nF (after zeroing test lead capacitance.) Then, using the nS function of the same Ohms switch position, I found the leakage resistance of the back biased diodes to be about 1000 megohms. It was pleasing to me that I was able to make all of these tests with my favorite meter.

The lesson?

When you encounter strange or unexpected measurement results, first check for the obvious - shorts and opens - sometimes the associated wiring might be a factor. Then, think about the input circuits of your meter. Are there differences, as there were in this case, that might explain the strange results?

It pays to study your meter and its manual, and understand a little about how it works, when you encounter problems like this one.