Power Supply - Transformer Buzzing

[2X]

Hello,
I have a PCB from a power supply where I have a hearing buzzing problem. When I just have connected the power supply on the mains (without load) a transformer/inductor (Picture 1) makes a little hearing buzzing noise, but when I connect a device/load on it the buzzing noise is quite big (it gives the feeling that something is going to explode). I placed an oscilloscope on the output of the power supply and I saw some spikes on the output waveform (Picture 2), where I beleive this voltage ringing is the source of buzzing.

Also, I measured with an oscilloscope a same one power supply (this don't make a hearing buzzing when is working with or no load on it) and on the output waveform (Picture 3) there is no the existence of the voltage spikes like in Picture 2. One strange thing is that I observed is when I disconnected the C56 capacitor the buzzing stops and the voltage spikes dissapeared from the waveform. I took the same  good one capacitor from the good one PCB and I replace it but the same buzing again.

I made a schematic (Picture 4) from a part of the PCB and I can't understand what causes the voltage ringing. There is a chance that the buzzing problem caused by the transformer? What else in combination with the capacitor C56 can cause the voltage ringing?

[Jwillis]

Transformers will vibrate due to Magnetostriction. I most cases this isn't profoundly noticeable unless there is something loose within the transformer. Broken adhesives on the core or loose winding's can produce noise. Also the amount of noise depends on the resonant frequency of the material used, the frequency of the field applied and the strength of field. Loose core halves can have gaps which can cause spikes because of the electromagnetic energy stored in those gaps. It's really a mechanical issue that can't be remedied without changing the frequency and/or fixing any loose transformer core elements and winding's.   

[2X]

Transformers will vibrate due to Magnetostriction. I most cases this isn't profoundly noticeable unless there is something loose within the transformer. Broken adhesives on the core or loose winding's can produce noise. Also the amount of noise depends on the resonant frequency of the material used, the frequency of the field applied and the strength of field. Loose core halves can have gaps which can cause spikes because of the electromagnetic energy stored in those gaps. It's really a mechanical issue that can't be remedied without changing the frequency and/or fixing any loose transformer core elements and winding's.   

So theoretically if I replace the transformer from the good one PCB the noise will be stop (I will make a try). Also, could I check the loosing windings or the other defects that you said of the transformer with an Insulation Tester Multimeter (I don't have one yet).

The bad luck is tha the transformer resonates with this certain capacitor and if I remove it the noise stops. As I know this C56 capacitor is an X type capacitor which helps only to reduce EMI due to differential noise and the designeer selected this certain value for the optimal EMI reduction I beleive, so if I to try to put another value or to remove the capacitor at all is not acceptable in my opinion.

https://www.fluke.com/en-us/product/electrical-testing/insulation-testers/fluke-1587-fc

Thank you for the detailed reply.

[inse]

Where is the measurement taken?
The waveform looks like what “modified sine“ inverters put out.
Many electronic devices don’t like this waveform as it causes huge ripple currents in capacitors.

[2X]

Where is the measurement taken?
The waveform looks like what “modified sine“ inverters put out.
Many electronic devices don’t like this waveform as it causes huge ripple currents in capacitors.

I measured these waveforms with probe of the oscilloscope between P1 & P2 pins on the schematic.
The picture 2 is from a device that buzzing and the picture 3 is from another exactly same device that doesn't makes any noise.

[T3sl4co1l]

I don't see HTPN60C3D anywhere, are you sure it's an IGBT?

What the heck is that crystal doing there... is this an MCU driven power supply? In the 80s, really?  Legendary if true...

...Or is it just a "modified sine" inverter?  What's up with the time scale, that sure doesn't look like a "power supply" to me (converts mains to other voltages, usually DC).

That would better explain the circuit, then; you're simply seeing the output EMI filter, and yes some ringing is normal.  The resistors are there to dampen it, and both the inductor and capacitor (electrostriction) are expected to make noise.  You might need to replace the capacitor in case it's undergone too much self-healing or oxidation and is out of tolerance (check C and ESR; mind that small values aren't easy to resolve on most ESR meters, check the user manual), and it's possible the inductor has come loose; it's held together with tape and varnish, which is usually pretty strong, but loose parts can vibrate and amplify sound.

Tim

[2X]

I don't see HTPN60C3D anywhere, are you sure it's an IGBT?

What the heck is that crystal doing there... is this an MCU driven power supply? In the 80s, really?  Legendary if true...

...Or is it just a "modified sine" inverter?  What's up with the time scale, that sure doesn't look like a "power supply" to me (converts mains to other voltages, usually DC).

That would better explain the circuit, then; you're simply seeing the output EMI filter, and yes some ringing is normal.  The resistors are there to dampen it, and both the inductor and capacitor (electrostriction) are expected to make noise.  You might need to replace the capacitor in case it's undergone too much self-healing or oxidation and is out of tolerance (check C and ESR; mind that small values aren't easy to resolve on most ESR meters, check the user manual), and it's possible the inductor has come loose; it's held together with tape and varnish, which is usually pretty strong, but loose parts can vibrate and amplify sound.

Tim

There four of HTPN60C3D on the orthogonal heatsink (in blue boxes at the attached picture 1). The crystal is connected below on a chip where I beleive is the oscillator for the IGBT driving (Picture 5A). At the output it connected an AC power cord to a PC monitor where it takes this modifed sine wave of 230VAC.

[inse]

Confusing stuff removed

[2X]

Now, is it supplied with modified sine?
Does it also buzz when connected to regular mains voltage?

If i connect the power supply to the mains I hear a little buzzing but if I plugin the PC monitor on it the buzzing is very big. The same model of the power supply doesn't buzzing at all even without load on it.

[T3sl4co1l]

CD4060, just a big crystal divider chain lol, way more accurate than needed.

Tim

[inse]

As noted before: electronics don’t like the waveform.
Do a little research on ‚modified sine‘ and you will find more evidence on that.
I had been confused by the wording as the circuit is an inverter, not a power supply.
100% agree on t3sl4co1ls posting.

[2X]

As noted before: electronics don’t like the waveform.
Do a little research on ‚modified sine‘ and you will find more evidence on that.
I had been confused by the wording as the circuit is an inverter, not a power supply.
100% agree on t3sl4co1ls posting.

You have right the more accurate is inverter but why the one works fine without buzzing and the other have buzzing? As I know the modified waveforms create harmonics that will lead to EMI and RFI noise that don't like the electronics. There is a possibolitty that this inverter is a cheap in design from his designeer?

[inse]

Modified sine is a cheap solution and  has its limitations.
As t3sl4co1l noted, maybe a magnetic core has come loose or broken.

[2X]

You are both right. The problem was the transformer. As I know the most SMPS transformers are custom made for the specifications of the PCB designeer. Do you if there is a way to find a similar or if this can repaired (like rewind)?

[T3sl4co1l]

If there are loose parts, it can be re-varnished.  Urethane or alkyd varnish will work in a pinch, or superglue (if it doesn't run hot) or epoxy, or there is a proper electrical varnish that they use.

I would not recommend indiscriminately dipping it in goo, as it's not likely to change anything, loose material is likely to be visible, and you may simply make things worse.

You aren't likely to find an equivalent inductor, and rewinding would be full-custom, including counting original turns, and getting the air gap right.

If the design peak current, RMS, and inductance are known, it might be replaced by a lower-magnetostriction material, but that leaves several unknowns.

Tim

[2X]

If there are loose parts, it can be re-varnished.  Urethane or alkyd varnish will work in a pinch, or superglue (if it doesn't run hot) or epoxy, or there is a proper electrical varnish that they use.

I would not recommend indiscriminately dipping it in goo, as it's not likely to change anything, loose material is likely to be visible, and you may simply make things worse.

You aren't likely to find an equivalent inductor, and rewinding would be full-custom, including counting original turns, and getting the air gap right.

If the design peak current, RMS, and inductance are known, it might be replaced by a lower-magnetostriction material, but that leaves several unknowns.

Tim

All I have is to find where there is the loose part and drip some drops of varnish or to cover all this loose area with varnish? Do you have any brand of varnish to suggest me that you trust? As I know Polyurethane is much tougher from the side of weather conditions, oil, water than alkyd varnish (I don't know how varnishes stand against heat). The factory uses these classic types of varnish or there some special varnish type for transformers where have higher insulation resistance and thermal stress? Thanks for your detailed answer.