Switching transformer circuit produces horrible distortion and ringing

[Vilius_Zalenas]

Hi,

I am an analog/power electronics beginner. I have a task to generate nearly 200V AC voltage already having a 18.4V DC rail, but I faced massive distortion and transformer ringing problem...

In the beginning, I decided to go with a 1:10 pulse transformer and extremely simple switching circuit. I am using a pulse transformer (MPN: 750032050) and N mosfet (IRF 3205). I hooked everything up on my breadboard, and turned on an oscilloscope to see what is happening. So the first picture is what I saw immediately (picture 1): both windings have extremely weird higher order harmonics coupled into the 110 kHz square signal, also there was huge voltage spike every period. After some more research, I decided to put a snubber circuit for the mosfet and I decided to disassemble my breadboard and solder everything down to an example board (to ensure perfect contacts and minimize all parasitic parameters). I provide the formulas ( which I used to determine the RC snubber parameters) in the picture. So the second time (after soldering and adding a snubber branch) I saw this (picture 3): the voltage spike is still present and the signal is nothing similar to a square form (as it should be).

Some more additional info:

1. I checked my wiring many times (it is right)

2. I am driving the mosfet from PIC32 (I am using a high current PWM pin, which is set to 110 kHz and supplying 12-13 mA of current)

3. My 18.4V rail is capable of supplying up to 1A of current. I measured the transformer current (in the operating state), it was around 370 mA, so I dont think I have a supply rail problem (the mosfet gets only warm, but not hot...)

4. The datasheet of the transformer says its operating frequency is 100 - 400 kHz, so I picked 110 kHz PWM in order to keep the frequency as low as possible to minimize the turn on and turn off mosfet delay effects to the signal.

I am not sure if I am putting enough current into my mosfet gate, I was thinking to buy a mosfet driver, but then again, 110 kHz is not an extreme frequency... (maybe you can recommend me a good driver for the given parameters). Besides the mosfet gate current, I dont have any other ideas on how to make it work properly. So I am kindly asking for any advices on how to setup that magic pulse transformer circuit (to get something similar to square signals). I thank you in advance

Vilius

 

[KlausST]

Hi,

you talk abut generating pulses .... and at the same time you mention "breadboard".

Thus my comment on breadboard:

A breadboard is a no-go for generating (clean) pulses.

It simply is neither made for high frequency nor high currents.
For sure it´s additional capacitances and additional inductive impedances cause a mess.
And the built test circuit is not reliable.
Tell 5 people to do a simple "pulse generator" on a breadboard and you will see 5 totally different waveforms.
Some may be more clean, some may not work at all. For beginners this happens by accident.

When a professional tries the same on a breadboard the result will be more he/she knows about the impedances and how to go around them. Not all of them.

So many beginners find circuits like this : https://www.wumpus-gollum-forum.de/forum/thread.php?board=58&thread=234
a mess. But indeed it is not. It is way better than any breadboard can do.

Klaus

 

[Vilius_Zalenas]

Hi,

you talk abut generating pulses .... and at the same time you mention "breadboard".

Thus my comment on breadboard:

A breadboard is a no-go for generating (clean) pulses.

It simply is neither made for high frequency nor high currents.
For sure it´s additional capacitances and additional inductive impedances cause a mess.
And the built test circuit is not reliable.
Tell 5 people to do a simple "pulse generator" on a breadboard and you will see 5 totally different waveforms.
Some may be more clean, some may not work at all. For beginners this happens by accident.

When a professional tries the same on a breadboard the result will be more he/she knows about the impedances and how to go around them. Not all of them.

So many beginners find circuits like this : https://www.wumpus-gollum-forum.de/forum/thread.php?board=58&thread=234
a mess. But indeed it is not. It is way better than any breadboard can do.

Klaus

Like I mentioned, after my first try, I soldered everything solid to the board, unfortunately, the waveform is still horrible, do you have any other advices or observations?

 

[KlausST]

Hi,

You are worried about gate drive...
--> Show a scope picture of the gate voltage.
--> Show a scope picture of the supply voltage.

--> Show a photo of your circuit including scope connections. (~100kBytes)

Give links to capacitor datasheet and transformer datasheets.
When you switch ON ... you mainly see the leakage inductance effects.
When you switch OFF ... you see the total inductance effects.

Klaus

 

[Easy peasy]

your approach is misguided - if you want 200Vac square wave from 18VDC, a push pull with snubbers is the way to go.