UC3879: Phase shifted full bridge converter

Dear Mr sun

I am able to achieve ZVS in Leading Leg in small Load current but when current is increasing an oscillatory behavior is starting to take place in VDS
please find below design details

Mosfet:IXFN38N100Q2 Coss=950pF

L_leakage of transformer=11.3uH

L_magnetising of transformer=17.793mH

Vout=24V

Pout=2.5Kw

Vin_nominal=600VDC

Removed snubber but response is not changing.

Hi, HARIKRISHNAN,

I am working on that, will get back to you soon.

Regards,

Bosheng

Hi, HARIKRISHNAN,

Can you show some VDS waveform? Did you try to tune the control loop?

Regards,

Bosheng

Dear Bosheng

Please find waveforms

1) Low output current ZVS is visible

2) High output current VDS is oscillating

Thank you

HARIKRISHNAN

Hi HARIKRISHNAN,

Please try to reduce the deadtime.

What is the value of your resonant inductor?

Regards,

Bosheng

 

Lresonant=11.2uH

not added shim inductor , only leakage inductance of transformer used

Dear Bosheng,

I have added a 10uH inductance in series with the primary of the transformer and checked the wave form
now the inductance is 11.2uH leakage plus 10uH inductance
even though lagging leg is not going into ZVS please let me know any other corrections to do

HARIKRISHNAN

Hi Harikrishnan,

I have asked an expert to look at this, he will help soon.

Regards,
Bosheng

Dear Bosheng,

After some mathematical calculations I have concluded that energy available for lagging leg to go into ZVS is not sufficient at my current

please look into the following difficulties I am facing

The lagging lag is not going into ZVS but the gate voltage is reducing to zero after turning it ON why?

Can you suggest some thing

HARIKRISHNAN T

Hi HARIKRISHNAN,

I believe this is more related to PCB layout and drive capability of the controller rather than topology of the converter. It seems that the glitch was happened at the Miller Plateau (Vds falls) so the we can't say the MOSFET was fully turned on at that time. 

Thanks
Oliver

Hello Harikrishnan

The gate voltage is dropping almost to zero due to miller capacitance effects. In the absence of ZVS the high dv/dt at the MOSFET drain partially discharges the gate / source capacitance of the device. What you are seeing is a severe form of the 'Miller Plateau'. If you achieve ZVS, then the drain voltage is already zero at the time that the MOSFET is turned on and there is no dv/dt to discharge the gate/source capacitance.

The ZVS transition happens much more quickly when the system is at full load. In your case you should try to reduce the dead time at full load.

The UCC3895 device and the newer UCC28950 device offer an adaptive delay function so that the delays are modified automatically to take account of the way the optimum dead time varies as a function of load current.

Regards

Colin 

Hello Mr.Colin

After reducing the dead time I have experienced cross conduction in leading leg.

I have added 20uH inductor in series with the transformer and this issue got resolved

Leading leg is now switching in full ZVS and lagging is in partial ZVS and I got nearly 3% more efficiency than normal full bridge. I have conclude my primary current which is nearly 4A is not enough for lagging leg to go into ZVS

But 30uH leakage inductance is not an acceptable solution as oscillations in the drain current is too high and my devices are blowing while running in full load for some time.

Please find below design values

Input 600 VDC

output 110V DC

Power 2.5KW

transformer magnetizing inductance-17.793mH

Transformer leakage inductance-11.3uH

Coss=950PF at 25V

Transformer ratio=34/9

Lout=230UH

Regards

HARIKRISHNAN NAIR

Hello Harikrishnan

I'm glad to see that you have made some progress -

There are a couple of things that could be causing the over current oscillations 

1/ There is a possibility for an over voltage condition to occur as the transformer comes out of freewheeling. This can be clamped by adding some diodes as shown in Figure 48 of the UCC28950 (also a PSFB controller) - copy shown below.

2/ It may be that there is some un-clamped inductive ringing due to a PCB layout issue

3/ Your transformer has a very low leakage inductance in proportion to its magnetizing inductance. This is normally a good thing but it also implies that the transformer has a very high inter-winding capacitance. Switching action can drive fairly large currents through this pri/sec capacitance leading to ringing on the current sense signal. 

Could you post a schematic, a plot of the CS signal and a plot of the drain voltage waveforms. If you wish you can send them directly to me at colingillmor@ti.com -

Regards

Colin