4
\$\begingroup\$

I am looking at a full bridge rectifier from mains voltage, using MD5S-W.

First, I build my circuit for use at 20V AC for safety reasons. I used this adapter:

enter image description here

And this is my circuit:

schematic

simulate this circuit – Schematic created using CircuitLab

Now, using my probe at 1x attenuation, I see the output as I expect it to be:

ac 20v expected

But if I switch my probe at times 10 (x10), the output gets weird, every second 'rectification':

weird rectification

Now, the issue gets worse if I plug the 220Vrms mains, and probe it with x10 at the output of the rectifier, using this circuit:

schematic

simulate this circuit

(Larger resistors and I also used resistors on the probe + used the 10x probe attenuation since my oscilloscope/probes are not suitable for >250V use.

oscilloscope extra resistance

I see this:

ac mains rectification

So every second waveform, instead of just making it look worse, now its gone! It looks like half bridge rectifier.

My question:

  1. Why every second waveform/wave there is this issue with the attenuation, while the 'every first' waveform/wave looks just fine? Its just 50Hz after all, not a couple of Mhz or more.

  2. How could I probe my mains for more reliable results? My oscilloscope and probes:

Oscilloscope rated for:

oscilloscope voltage rating

Probes Rate:

probe rating

\$\endgroup\$
1
  • \$\begingroup\$ By the looks of it, that MD55 probably has an autotransformer instead of an isolation transformer. While it lowers the voltage a bit, you're still connected to mains and won't have the benefits of a proper transformer (among others, the voltage dropping significantly if the load is higher than supported). \$\endgroup\$
    – Mast
    May 28 at 21:09
5
\$\begingroup\$

You have two different problems.

First, the "mangled" output when using the 10X probes. As another answer has explained, the load of the 10X probe isn't enough to pull the output from the rectifier down to zero. A larger load across the bridge rectifier will take care of that. Even as little load as a couple of hundred kilo ohms would take care of this "problem."

Now to the more concerning case.

You have to realize that neutral and ground are connected together in your house, and that the ground clip of your oscilloscope is also connected to the ground of your house.

Your set up looks like this:

schematic

simulate this circuit – Schematic created using CircuitLab

Look over there to the left in the box labeled "Fuse box." The safety ground and the neutral are connected together.

Look over to the right at the box labeled "Scope." The safety ground of the scope is connected to the safety ground of the house. The scope safety ground is also connected to the scope probe ground.

Anything you connect to the probe ground clip is indirectly connected to the neutral of the house.

The junction of D3 and D4 is connected to neutral through your scope and the house safety ground.

That connection in effect causes a short to neutral on every other half cycle.

If you didn't have those resistors in there, you'd have a short circuit that would destroy the bridge and the scope and probably trip the circuit breaker - or maybe you'd be lucky and have a ground fault current interrupter (GFCI) in the house that will shut off the power to the outlet before anything gets damaged.

Another thing to keep in mind is that the ground clip of all channels of the scope are connected together. Clipping both ground clips to different points in your circuit that are not ground will cause a short circuit.


You have a similar problem to this question from last year.

This is the output of a full bridge rectifier where the scope and the source shared a ground:

enter image description here

The flat "bottoms" of your shorted bridge are slanted instead of horizontal, probably because of the 1M resistors playing with the capacitance of the circuit.


Be very careful when working with line voltage.

Things you don't know about (like the shared ground) can bite you really hard. Even things you do know but momentarily forgot can bite you. It only takes a moment's distraction, and you can end up connecting the scope ground to a hot wire - or worse, connect yourself to a hot wire.

If you must work on live line voltage, use an isolation transformer between the outlet and the device you are testing. It will prevent short circuits between the scope and the device, and between you and the outlet.

\$\endgroup\$
2
  • \$\begingroup\$ That's very odd that he'd be seeing such effects when he's powering the rectifier from a Class II (double insulated) AC-AC step down transformer...the metallic "short" path you describe doesn't exist in the OP's circuit \$\endgroup\$ May 28 at 21:59
  • \$\begingroup\$ @ThreePhaseEel: It doesn't exist in the first circuit through the transformer. The second circuit goes straight to the mains. No transformer involved. \$\endgroup\$
    – JRE
    May 28 at 22:03
3
\$\begingroup\$

The issue was that the oscilloscope was using the same AC mains wires with my system. I cannot explain this more, this hackaday tutorial helped me understand the issue:

hackaday page

hackaday page

So I used a AC-AC isolated transformer (From 220 to 130 Vrms) And It worked! I get the full rectified waveform now.

But a new question arises. As the tutorial says:

Why is every-other waveform transition a bit mangled? I had expected a consistent waveform through every transition.

But I wont bother with this, my system does what I want.

\$\endgroup\$
2
\$\begingroup\$

Short answer:
The bridge rectifier expects to see a much lower load resistance.


The most "sane" waveform is the one where the bridge's output side sees a 1-MEGohm load. It is almost what you'd expect. You can verify this by adding another resistor of smaller value in parallel with the oscilloscope.
The waveform should become the expected double-peak full-wave DC.

Why every second waveform/wave there is this issue with the attenuation, while the 'every first' waveform/wave looks just fine? Its just 50Hz after all, not a couple of Mhz or more.

This is a bit complex, because now the oscilloscope is AC-coupled instead of DC-coupled. So now a coupling capacitor inside the oscilloscope becomes part of the circuit, along with the oscilloscope 1-MEGohm input resistance. When DC-coupled, this capacitor is switched out.

It would appear that small differences between the bridge diodes causes this capacitor to retain some of its charge on one-half cycle. Again a smaller load resistor should give the expected waveform. But usually, you'd keep the 'scope DC-coupled when making these measurements.

\$\endgroup\$
2
  • \$\begingroup\$ Good approach. I did this and I did not get the result I expected. the lower the load I was using, the more the negative part of the AC was showing on the oscilloscope. In other words: I put a load, and the output of the rectifier was 100% same as the input. I managed to solve the issue though! Check my answer. \$\endgroup\$ May 28 at 15:12
  • 1
    \$\begingroup\$ I had assumed (apparently incorrectly) that your 20V AC source was a step-down transformer that did provide isolation via its magnetically-coupled primary-to-secondary windings. You should still use DC-coupling rather than AC-coupling in your 'scope. \$\endgroup\$
    – glen_geek
    May 28 at 15:20

This site is temporarily in read only mode and not accepting new answers.

Not the answer you're looking for? Browse other questions tagged .