Reducin Transforme Audible Noise

[sabu31]

Hi all,

I am developing a pulse transformer with input 0-500V and output 0-4kV at 78 ohms load. The pulse width maximum is 10us. And the frequency is 250Hz. The core used is amorphous metal core (UMCCC-400). Cross-section area is 12.5cm^2. The approximate magnetizing inductance is nearly 112micro-henry.

I tried two configurations of the transformer. And the winding diagram is attached
In the interleaved case, I am getting lower ringing, but the audible noise is higher (around 60-70db). In the configuration where the primary is parallel and secondary is series, the audible noise is lower (still not comfortable) but the ringing is higher. Surprisingly the leakage is nearly same around 900nH. My queries are

How can I reduce transformer Audible noise? I covered each layer with Kapton tape.

Will keeping one terminal of secondary grounded reduce the ringing

 

[wwfeldman]

the time scale of the 'scope traces is 4 us, with the ringing (at least) being 4x, or more greater.
so the ringing frequency is about 1 MHz. since you can't hear that, the noise is due to the 250 Hz
repetition rate.

the noise is due to domain shifting and core material movement
maybe copper windings attracting

what is the mechanical configuration / what type kind of core are you using?
did you vacuum impregnate or pot the transformer?

 

[mtwieg]

Hard to say whether the sound is due to
1. Lorentz forces on the wires,
2. Magnetostriction in the core, or
3. Attractive forces between the two core halves.

The means by which the cores and windings are fixed in place will also matter tremendously. Increasing the clamping force between the core halves can mitigate #3. Potting the windings can mitigate #1. #1 is basically impossible to mitigate without changing cores entirely, or reducing the magnetizing flux.

Also, a rough calculation estimates your magnetizing flux will swing above 0.6T. Consider pre-biasing the core with around -0.3T, either with another winding or with a bipolar pulse shape. A flux swing of +/-0.3T might produce much less acoustic noise than a flux swing of 0/+0.6T.

 

[Easy peasy]

when it's working OK pot the whole thing in 2 part epoxy resin, aka Araldite, to make it quieter ...

 

[FvM]

Not shown in the oscilloscope waveform, but according to previous threads, the pulse has alternating sign, +/- 0.6T following your numbers.

As already pointed out, the winding configuration has no effect on core flux and magnetostriction, winding vibration may change. All in all, you need impregnation and/ or potting to reduce noise.

 

[Easy peasy]

Just to answer your question, the core halves will attract each other as the mag field goes to max each switching operation - if you can put some thin card or silicon sheet between the core faces ( the Lmag will reduce - but not a biggie ) you may soften the blow and reduce the noise.

The freq of the noise is directly related to the repetition frequency of the pulses ...

 

[mtwieg]

Yes, even if you do not have an air gap, adding a thin spacer material between the halves may help a lot with sound due my reason #3. Try different materials, like silicone, phenolic, FR4. Also experiment with different amounts of clamping force, and also add similar layers of material between the core and the clamping fixture.

A picture of your construction might help a lot.

 

[sabu31]

I have not potted as yet as I was trying different winding configurations , so had to remove the wires . As suggested I will use epoxy coating and update results.

 

[Easy peasy]

epoxy may do nothing if the core faces are hard and butted up against each other ...

 

[sabu31]

Then what material should i place between the core faces as I do not want to reduce the magnetising inductance

 

[sabu31]

Hi all,
I had put Kapton tape on the faces of the core. It helped reducing the noise but at cost of reduced magnetising inductance. Should I wrap the entire core with the tape . Any material better than Kapton tape for putting in between the faces (say common insulation tape etc.) Thanking you.

 

[mtwieg]

No sense in covering the entire core with tape, only places where a lot of pressure is felt on the surface (where the core halves meet, and maybe where they contact the clamping fixture holding them together).

One 0.001" piece of kapton should not significantly affect Lm of such a large transformer. Maybe you're using several layers though.

Material depends on how much gap you are willing to introduce.

 

[sabu31]

Thanks mtweig for the reply. I have put two layers on face of one U core section. The Lm reduced from 140uH to around 78uH. I am now increasing the primary turns from 5 to 6 to bring the Lm back to above 100uH. Should the windings be also wound by tape. I am presently using Kapton for that also.

Is there any other tap which has strong adhesion along with insulation. Will the following be of help

https://www.amazon.in/Tape-69-Fusin...otton+tape&qid=1639660886&s=industrial&sr=1-5

https://www.amazon.in/ETI-Teflon-Sealing-Machine-Machine/dp/B07FCHTRV2/ref=sr_1_9?keywords=glass%2Bcotton%2Btape&qid=1639660886&s=industrial&sr=1-9&th=1

 

[Easy peasy]

the interface material between the core faces should be durable but also able to absorb energy to reduce the noise - some sort of silicone sheet might be best - or hard rubber compound - you cannot escape the drop in Lmag - so best design for it.

 

[dick_freebird]

Might consider a urethane adhesive like 3M 5200
or E5000, they have good adhesion strength but
will dampen acoustics as it cures "rubbery".

 

[sabu31]

Might consider a urethane adhesive like 3M 5200
or E5000, they have good adhesion strength but
will dampen acoustics as it cures "rubbery".

Is it for complete potting or for use between the amorphous iron core faces

--- Updated Dec 23, 2021 ---

Hi all

Will silicone sealant material as shown in link

https://www.amazon.in/Dr-Fixit-501-Silicone-Sealant-Clear/dp/B07X89WDYG/ref=sr_1_7?crid=RC5S9C9V91NY&keywords=silicone+sealant&qid=1640258350&sprefix=silicone+sealant%2Caps%2C342&sr=8-7

be useful for complete potting or use for between windings.