UCC28950: pmp8740 shim inductor heating

Hi Nidheesh,

It could be a couple of things.

1. power dissspation in the windings, measure the winding resistance and calculate the I^2 * Res loss in the winding. Note that the AC resistance will be higher due to skin effect and proximity effect. But it will give you a good first estimate.

2. Core loss, I cann find the AF139 material on the Web, we normally use 3F3, 3C96, or 3C95 materials. Please check peak flux levels by calculation to ensure the core is not saturating. You should also calculate the power dissipation in the core at the flux level of operation and frequency.

If the shim inductor saturates then there is only a low value of inductance in the resonant tank and ZVS may not be possible.

Regards

Peter

Hi Peter,

thank you for your response.

Actually core name was CF139 not AF139 ..sorry..

When I was using the conductor of 21SWG and 2 strands heating was severe. But 35 SWG 30+ strands heating got reduced but still heating.

Can you say anything about ZVS from the attached graph?

Actually gate pulse is a bipolar signal, so w.r.t common zero graphs attached. as per the graph, Gate pulse is getting positive only after VDS became zero. So is this correct ZVS. Or we have to look into the first graph? ( Both graphs are the same, But in second both have common 0V)

So can you judge anything on the ZVS from this?

Hi Peter,

Thank you for your valuable replay.

Now can you please say what was the max temperature in your shim inductor(in your evaluation board). It is very difficult to get the same part that you used.

Is that is having airgap between the cores?

Is there any isuue if I am using gapped cores to avoid heating due tocsatuarion?

Regards

Nidheesh.

Hi Nidheesh

I'm afraid I don't know what the temperature of the shim inductor on the EVM is.

If the inductor is getting hotter as the load current increases, then the increase will be due to copper losses because the core losses would remain more or less constant.

Assuming you have designed your shim inductor to run at 100kHz and with a flux swing of 200mT then the CF139 core material will have a power loss of 400kW/m^3. The PQ26/20 core has a volume of 5820mm^2 and this gives a core loss of 2.3W. Assume a similar loss in the copper for a total close to 5W. Please check these numbers.

Anyhow, the shim inductor does not have a DC current in it so a gap is un-necessary. If you have no gap then the permeability will be close to the ui value in the material data sheet or about 2100. The Al value of the PQ26/20 is 7020nH*T^-1 so you would need only about 3 turns to give 50uH.

Take a look at the information and procedures in the documents at

https://www.ti.com/seclit/ml/slup304/slup304.pdf

and https://www.ti.com/seclit/ml/slup132/slup132.pdf

Regards

Colin

Hi Colin,

With Litz wire I was able to reduce heating little bit.

 So issue is seems to be from the core material that I have purchased. So now I tried with differnet material from another manufacturer. Now heating is almost under control.

But it is noticed that R98 is heating. What is the purpose of this R98?

(my Output voltage is 60V and Vds of secondory MOSFET is found to be around 180V)

Regards

Nidheesh

Hi Colin, 

I have read an article in which u described how to calculate R98 value. So in that, you are calculating 1/2 L/N I^2 for leakage inductance w.r.t secondary.

I have Shim inductance of 10uH as leakage if so do I need to add this also to the calculation of 1/2 L/N I^2 or only transformer leakage inductance need to be considered?

Hi Nidheesh,

Sorry for the delay but Colin is out of the office this week, he will be back next week and can answer then.

I did look for R98 in the UCC28950 EVM but I did see it in the BOM?

Regards

Peter

Hi Peter,

Here I am attaching the schematic.R98 is marked with a red circle.