Transformer high voltage spikes

Ian0

Joined Aug 7, 2020
9,677
Generally, the two input clips are connected together at the scope input, and a probe will operate with its clip disconnected, provided that the other one is connected, but it will be very susceptible to pickup, so I wouldn’t trust anything at HF that I see on that channel.
 

Thread Starter

Burak HANÇERLİ

Joined Jan 4, 2016
9
@MrChips Thank you for the reply.

So either (as you suggested) I'm a lack of oscilloscope usage know-how or I still couldn't express myself clearly. Let's forget about the DC-DC converter circuit for a moment and focus on the AC measurement.

1598274929096.png

This is the way that I measure the signal on the secondary of the transformer. As expected the result is Vp+/-=+/-35V, Vpp=~70V and Vrms=~24V. (GND clip is already connected to one of the outputs of the secondary side. Are you referring to this ground since the beginning? Why on the earth should I try to measure anything without the GND clip is connected?)

As the transformer is isolated from the main I don't clearly get the idea of "grounding" in your terms. I, of course, know the necessity of differential probes with the reasonings if one will measure the main AC directly but since I'm not dealing with such a situation I'm using regular 10:1 scopes.

If you believe that I'm still making a mistake up to this point, I'd be glad if you can point me out some references and I'll be back after reading them.

So while I measure the transformer in this connection, I capture spikes if I turn it off. Here is the capture;

SCR14.PNG

Then I decided to test it by connecting a 10ohm resistor as a load to see if it can suppress the spikes by consuming a fairly enough amount of energy. And it worked, spikes did not happen until I remove the load. (This one is not a solution, just wanted to test it)

If you still say that spikes on the screen are because the probe is picking up noise from some other place, I definitely don't understand that part and need more explanation about it.
 

Ian0

Joined Aug 7, 2020
9,677
Is the transformer running (i.e. Primary connected to the mains) at the point you took this waveform? In other words, are we looking at a small portion of 50Hz sine wave? Possibly the point at which the rectifier starts or stops conducting?
 

Thread Starter

Burak HANÇERLİ

Joined Jan 4, 2016
9
Is the transformer running (i.e. Primary connected to the mains) at the point you took this waveform? In other words, are we looking at a small portion of 50Hz sine wave? Possibly the point at which the rectifier starts or stops conducting?
This is a capture from the exact time of turning-off the transformer. And yes, indeed, it is a very small portion of 50Hz.

1598278280884.png

There's nothing connected to the transformer at the time that I captured those spikes. It's just oscilloscope probes. And I get similar spikes on every turn-off moment (also there are other spikes at the turn-on moment and they're similar in between them)
 

MrChips

Joined Oct 2, 2009
30,720
What is switching off the power to the transformer?
Is it a mechanical switch?
If it is, then what you are seeing is called switch bounce. This is expected.
 

Ian0

Joined Aug 7, 2020
9,677
as it only lasts half a millisecond, a varistor or a Transorb will deal with it, or perhaps even a snubber across the switch, or across the either side of the transformer.
 

Thread Starter

Burak HANÇERLİ

Joined Jan 4, 2016
9
Thanks again.

Yes, it's a mechanical switch. As you suggested we were already discussing the switch bounce but I was expecting a couple of tens of microseconds and 10 or 20 spikes at most, not nearly 1ms of spikes with hundreds of peaks, that's an interesting discovery for me. And what I don't get is how it can go up to 70V. I instinctively expect a decreasing amount of energy after turning-off the transformer, so which should result in a set of spikes which are getting smaller and smaller each time.

If I connect the SMPS PCB those spikes are somehow going through the filtering circuit and triggering the crowbar circuit.

I'll try to test it by connecting the transformer to the variac to eliminate the mechanical switch bounce.
 

MrChips

Joined Oct 2, 2009
30,720
An inductor stores energy. When you break the current you release the energy stored in the inductor. This produces a back e.m.f.
 
Top