Uplight dimmer based on phase-fired controller: Noise Issue

[FvM]

The ultimate reason for the sound is core magnetostriction. It's intensity depends on core material, peak flux and di/dt. As said, an elastic potting material (PU, silicone rubber) can be expected to give better attenuation.

 

[c_mitra]

finally I potted the inductor as can be seen in the attached, unfortunately now the noise is even louder

You have to do some diagnostics:

1. How much is the noise when the load is removed? That will make the current though the inductor zero and the noise should disappear completely if the inductor is the culprit.

2. The body appears to be made of iron sheet (test with a small magnet). with the load connected, and the noise is maximum, touch the case with a insulated screwdriver (careful!) and listen if the noise is getting modulated. If yes, that suggests that the case is also vibrating with the choke and your potting has not touched the PCB and the inductor is still free.

 

[stenzer]

Hi,

Is it loudest at half sine wave trigger? and quietest at full power?

the "loud" noise starts at low light brightness, getting a little bit louder towards full load, and vanishes almost suddenly at almost full brightness/load. At full load the noise is pretty "silient" (one has to come very close to hear it).

in this case a hard potting compound only helps to provide a noise path - whereas a softer glue might have attenuated the noise.

UUnfortunately the compound is really hard.

1. How much is the noise when the load is removed? That will make the current though the inductor zero and the noise should disappear completely if the inductor is the culprit.

I haven't tested the dimmer without load so far. I will do this, and let you know how it behaves.

If yes, that suggests that the case is also vibrating with the choke and your potting has not touched the PCB and the inductor is still free.

I assume the potting has touched the PCB as quite a bit of the compound ran out on the lower side of the inductor (where the inductor touches the PCB), as I was not able to seal the PCB and the inductor fully. Further, the inductor is now permanently connected to the plastic housing. In the center of the inductor is a plastic cylinder (see picure in post #19) into which a screw is turned to thighten and close the housing. As you can see in the previous attached picture the inductor and thie cylinder are now connected by the compound.

BR

 

[c_mitra]

The ultimate reason for the sound is core magnetostriction. It's intensity depends on core material, peak flux and di/dt. As said, an elastic potting material (PU, silicone rubber) can be expected to give better attenuation.

The inductor appears to be ferrite-cored. Ferrites are not known to show significant magnetostrictive effect, if my memory is still good! di/dt should determine the frequency. Also this is difficult to explain that the noise disappears at full load. Is it possible that the frequency actually changes but is perceived as the intensity?

 

[stenzer]

Hi @c_mitra,

I performed the diagnostics as suggested.

1. How much is the noise when the load is removed? That will make the current though the inductor zero and the noise should disappear completely if the inductor is the culprit.

There is no noise at all if the load is disconnected.

2. The body appears to be made of iron sheet (test with a small magnet). with the load connected, and the noise is maximum, touch the case with a insulated screwdriver (careful!) and listen if the noise is getting modulated. If yes, that suggests that the case is also vibrating with the choke and your potting has not touched the PCB and the inductor is still free.

The inductor body is made of iron, I tested it with a magnet. By touching the inductor with an screwdriver the sound is not modulated.

BR