UCC3895 primary MOS burn out in light load or no load

Swing Jiang

Our customer find the primary bridge Mosfet would be burn out when the unit operated in no load or light load.however, increase the load, this issue will not happen.

I got some waveform from our customer, much appreciated your comments here.

Pic:ture A

Condition: Normal operation under no load

CH1: Id of Mosfet C

CH2: Vds of Mosfet B

CH3: Id of Mosfet B

CH4: Vgs of Mosfet B

Picture B:

Condition: First start cycle under no load

CH1: Id of Mosfet C

CH2: Vgs of Mosfet B

CH3: Id of Mosfet B

CH4: Vgs of Mosfet C

over 13 years ago

0 Swing Jiang over 13 years ago

TI__Intellectual 1230 points

We can see the Vds of Mosfet B stay keeping on according to picture A while Vgs of Mosfet B goes lower than the Mosfet turn off threshold.how can we to explain the phenomenon?

0 John Bottrill over 13 years ago in reply to Swing Jiang

TI__Mastermind 25130 points

Swing,

Can you provide the schematic?

John

0 Swing Jiang over 13 years ago in reply to John Bottrill

TI__Intellectual 1230 points

John,

This issue is found by the end customer, any our customer can not offer the schematic. i will try to convince our customer of the shcematic. is it possible to give some suggestion to our customer even we don't have the schematic, since this is the phase shift full bridge application.

0 John Bottrill over 13 years ago in reply to Swing Jiang

TI__Mastermind 25130 points

Swing,

My guess is that the converter under light loads is hard switching and the resulting losses are causing the FETS to fail. In most cases it is the A/B FETS that fail first but that can lead to the other FETS failing.

John

0 Swing Jiang over 13 years ago in reply to John Bottrill

TI__Intellectual 1230 points

John,

Thanks for your reply. Do you have some good suggestion to fix this issue?

0 John Bottrill over 13 years ago in reply to Swing Jiang

TI__Mastermind 25130 points

Swing,

Siince I don't have the schematic or any waveforms I would only be guessing. They could try putting a small series inductor between the junction of the A/B switches and the transformer with two diodes from the junction of the new inductor and the transformer to the rails. This should help the switching.Then they could play with the timing delays to see if that solves their problem.

If the unit is on under fell power almost all the time they could add a dummy load to the secondary that comes on when the load drops below a minimum level needed for soft switching.

They could also change the FET switches to more robust devices that can handle the hard switching.

These are all suggestions based on no information so none of them may work.

John

0 Swing Jiang over 13 years ago in reply to John Bottrill

TI__Intellectual 1230 points

John,

Thanks for your reply, much better understand now.

0 Swing Jiang over 13 years ago in reply to Swing Jiang

TI__Intellectual 1230 points

John,

DO you have more details discription for the soft start function? because of the failure is always happen at sart up, our customer want to know what our IC do during the start time, it's better we have some document or waveform to show to our customer.

0 Swing Jiang over 13 years ago in reply to Swing Jiang

TI__Intellectual 1230 points

John,

could you please help on the softstart question as show above? our customer want to know the what our IC will do during the whole softstart up time.

0 John Bottrill over 13 years ago in reply to Swing Jiang

TI__Mastermind 25130 points

Swing,

For this question I will need a schematic with values.

Regards,

John

Power management

Power management forum

UCC3895 primary MOS burn out in light load or no load