IR2113 blowing ! help needed...

[alexclaire]

hi everyone,
I've been reading this forum for a long time but this is my first request...
I have made a half bridge class d amp not self oscillating, based on a 16MHz crystal and 4060.
the triangle wave is 700mV pp @ 250 Khz
the input opamp is TL072 (supply +-10V)
LM 319 comparator feeding a level shifter common base transistor

the next stage is 4070 and IR2113 to the N cannel Fet IRFP4227...
54uH and 470nF LC filter
post filter feedback

Power supply is +-80V 1kVA
IR2113 4070 are biased 12V over -80V

First power on no problem sound was great deep bass high frequencies clear but when i pushed it to clipping (symmetric at least !)

1 min after the 2 fet and IR2113 blow up

cooling was sufficient.

I don't know what happened please help...

Fets are not cheap and IR2113 blow every time fet do...

[Marc E.]

Hallo Alex,

the Problem seems to be the feedback. Feedback including the Outputfilter is not an easy Task, you have to deal with a 180 Deg Phase Shift.
I assume that your Amp oscillates with the Freq. of the Filter, which is approx. 30kHz using your Parts.
Pickup Feedback at the Mosfets, and don't forget to use a Zobel-Glied. Q-Factors of typical Outputfilters reach 200 and more. (Look at the GIF)

Best wishes
M.E.

[Workhorse]

Originally posted by alexclaire
hi everyone,
I've been reading this forum for a long time but this is my first request...
I have made a half bridge class d amp not self oscillating, based on a 16MHz crystal and 4060.
the triangle wave is 700mV pp @ 250 Khz
the input opamp is TL072 (supply +-10V)
LM 319 comparator feeding a level shifter common base transistor

the next stage is 4070 and IR2113 to the N cannel Fet IRFP4227...
54uH and 470nF LC filter
post filter feedback

Power supply is +-80V 1kVA
IR2113 4070 are biased 12V over -80V

First power on no problem sound was great deep bass high frequencies clear but when i pushed it to clipping (symmetric at least !)

1 min after the 2 fet and IR2113 blow up

cooling was sufficient.

I don't know what happened please help...

Fets are not cheap and IR2113 blow every time fet do...

Hi Alex,

Post your schematic first...
The problem might be in your PCB layout, Schematic design and even in the component value....
BTW:Try to increase the Bootstrap capacitor for the High side drive, your high side Mosfet might get little current to charge up during the clipping which inturn forcing it into the linear region of working and finally it damages......

regards,
K a n w a r

[Workhorse]

I think it would be much better for this thread if Phase_Accurate[Charles] would offer his invaluable suggesstions for this problem....The thread is waiting for you Charles!

[phase_accurate]

Did anyone knock ??

Some possible sources of the problem that come to my mind:

1.) Is your switching waveform clean ( i.e. without large overshoots) ?

2.) Do you use gate series resistors (two reasons: too fast a switch-off can kill your FETs and the increased dissipation can kill your driver IC though the latter doesn't seem to be more than 1/3 of what the IC is capable to dissipate) as one should ?

3.) Is the intrinsic deadtime of this driver IC enough to prevent shoot-through ?

4.) The gate-charge of this FET isn't very high (it is actually conveniently low for such a beefy FET !) but the input CAPACITANCE definitely IS !

5.) There is no dV/dt rating on the datasheet of this FET so maybe it is not meant to be used for class-d applications.

6.) Don't drive class-d amps that are using bootstrap-capacitor-supplied high-side drivers into clipping - unless special precautions have been taken !!!!!

Regards

Charles

[Tim__x]

They're almost never shown, but schottky protection diodes (from the IR2113's gate output to mosfet source) are often needed. Especially with very low, or zero, gate resistors.

[darkfenriz]

Originally posted by phase_accurate

6.) Don't drive class-d amps that are using bootstrap-capacitor-supplied high-side drivers into clipping - unless special precautions have been taken !!!!!

Why?

[Tim__x]

If it clips on the high-side for too long, the bootstrap supply may fall enough to put the high-side mosfet in it's linear region, = .

[luka]

Makes sence to me

[Gyula]

Hi alexclaire!

If your amplifier contains feedback loop, you should consider the output's delay time, what is normally 1/(2*Fs) is growing dramatically, when the FET are sticked by the clipping. This effect elongate the square signal's period time, thus divide the Fs, and multiplies the delay time. This causes insufficient stability parameters at the Nyquist criteria.
If your output filter is close to ideal. Maybe super wire, winds are far from each other, free from proximity, ideal caps... You can mitigate the filter effectively with a small serial R to the coil, or the cap. You can calculate the exact values with the transfer function.
Maybe your flying supplies at the top side are drained, and the FETs started to function as current generators, and were getting hot . One of them failed, the HV supply reached the 2113's output, and the 2113 pulled off his own hat. You should to place electrolitic caps also at the top-side supplies.

Have a nice day !

Gyula