TinySwitch III temperature issue when running at 240VAC. PLS PLS Help... :)

Posted by: lbaerd on 06/22/2010

Dear Experts,

I have designed a 85VAC-265VAC LED Driver(8W) using TNY279. This is an isolated design. All the parameter is designed according to PI Expert recommendation.

When I run the design at 100VAC @ 25 Degree Celsius ambient, the switcher temperature is approximately 48 Degree (with small heatsink attached to the Source pin of TNY279). This temperature is pretty decent.

However when I run the design at 240VAC @ 25 Degree Celsius ambient, the temperature of switcher goes up to 58 degree which is 10 degree more than 100VAC input. All the design is the same.

As per my observation from reference design , the temperature for switcher should be quite close. However for my case, the difference of switcher temperature when running at 100VAC and 240VAC input voltage is about 10 degree celcius!

Do you guys have any idea where should I start to look at . Pls advice
TQ very much

John

Comments

Since the device is hotter at high line voltages, this indicates the losses are switching losses. I would look at the drain voltage and try to optimize the clamp circuit.

Z. Cochrane

Hi Experts

I am using RCD clamp, which is suppose to be quite efficient among other clamp circuits. Attached is the clamp circuit. I have tried other resistance value but to no avail. Any other recommendation
BIG THANKS

John

The best way to optimize the clamp is to look at the drain voltage and current. This way you can see any improvements as you change the circuit. the first thing I would lower the clamp's series resistor. This should lower the leakage spike on the drain when the FET turns off.

Z. Cochrane

Thanks for the recommendation. I have tried with lower series resistance and it seems to help abit. More evaluation has to be done.

I have captured the waveform of Drain current for both 100VAC and 240VAC input voltage. It seems that the Id waveform for 240VAC is very spiky and noisy. I am not sure if that is normal. Attached is the waveform for both input voltage.

Pls advice. TQ in advance

I do not know why there is more ringing at 240 Volts. How are you measuring the current? Current probe on the drain or resistor on the source? Also please take photos of the drain voltage waveform.

Z. Cochrane

I am using 0.5 Ohm resistor on the source to measure the current. I will have to get a high voltage probe in order to probe the drain voltage waveform. Will try to take the waveform.

So can I say that it is not normal to have so much ringing on the waveform?

Thanks

I use a current probe on the drain. The ringing may be due to the gate charge when the FET switches. Also make sure that the 0.5 Ohm sense resistor is non inductive. Do not use a standard wire wound resistor.

Z. Cochrane

Hi,

FYI, I am using thick film resistor. Do we expect more ringing when using wire wound resistor? For my case, the ringing is much more serious for 240VAC input voltage.

If it's due to gate charge when FET switches, is there anyway I could improve the situation

Thanks
John

The film resistor is OK. I was concerned that you were using a wire wound resistor.
Since I use a current probe on the drain, the gate charge is not observed. If the ringing is gate charge, then the ringing is not the issue. Also the photos are at different time scales. Taking that into account, the ringing is not much worse at 240 Volts. If this is the case, then this is not the issue.
The temperature increase may just be switching losses at higher line voltages. You may have to live with that. Optimizing the snubber may improve the device loss. If you can send me some photos of the drain voltage and source current (on same display), I would have more information to determine if the thermal issue can be improved.

Z. Cochrane