GZPA Class D 4000W with IRFP264

[B.I.G]

hello, i have a Ground Zero GZPA 1.4000D it is a class D mono using IRFP264 in the output ...

after some trouble with a 220nf capacitor on the input side i blew a irfp264 (leaving the amplifier on without a heatsink for a lot of time does that)

i replaced all outputs with irfp4229 and looking with a scope on the gate it looks like this it should be a square wave ?? it actualy looks like a triangle ...

the output of the fets is square....

i did apply power and loaded it on 1 ohm woofer playing music and it blew when i turned the volume up ... at about 1000-1500W ...

all irf3205 are dead and 2xirfp4229 ... i replaced the irf3205 and it works again ... any ideas on why did it die ?

the scope in the first pic is 2v/div ...

[savu]

irfp4229 seems to be better than irfp264, has faster body diode.
run it into 2 ohm load minimum to see what happens.

regards,
savu

[Pafi]

any ideas on why did it die

Because something was wrong.

If you showed a (partial) schematic and some other detail maybe I could be more specific.

Some interesting details:
- gate driver circuit (maybe it's partially damaged)
- supply voltage

I figured out that the configuration is half bridge with 5 MOSFETs paralelled (originally).

[nigelwright7557]

Interesting gate drive signal.
Could be they have a resistor on/ diode off circuit in which case the gate on drive will be a bit slower to stop cross conduction.

My class d amp has a square wave going to the mosfet gates but has deadtime programmable in the chip.

We really need a circuit diagram to help constructively.

[B.I.G]

i don`t have the schematics ... i was hoping that some one has a schematic of a similar amp or can tell me the gate drive signal is good or bad... i`m thinking is bad , all the amps i have seen use square wave to turn on and off fets.

this amp uses discrette driver i have attached it . the whole driver board is very simple (i was expecting a lot more). uses an resiston and a diode to turn on and off the mos-fet ... i will check tomorrow but i think it was a 100ohm resistor and a diode to turn it off . isn`t 100 ohm a little high value ? i was expecting 10ohm or lower...

this amp works on +/-91V regulated supply with minimum 1ohm load i`ve done 0,7ohm with no problem and only 8 output devices (IRFP264)

the amp uses an resiston and a diode to turn on and off the mos-fets.

[Pafi]

isn`t 100 ohm a little high value ?

Yes, absolutely!

Check low side gate driver against the high side one, because it seems to be correct!

[Eva]

Waveforms look perfectly normal for car-subwoofer grade class D. Were you expecting something better?

100 ohms is not a low value, but the first limiting factor for "switching speed" is PCB layout, MOSFET limitations (like body diode limits) only play a secondary role in that kind of amplifiers.

In fact, sad truth is that you may manage to make it work with IRFB4229 by further increasing gate resistors to 150 or 220 to reduce di/dt to the levels that the PCB can tolerate without EMI becoming self-destructive Reducing gate supply voltage a bit (keep it >9V) does have the same effect but it has to be done with care because other parts of the circuit could be disturbed.

Note that, although IRFP4229 is quite similar to IRFP264, it has 2 to 3 times higher transconductance than IRFP264 resulting in about 1.5 to 2 times higher di/dt (transconductance is higher but input capacitance is also a bit higher too).

Driving the gates with square-like waves only makes sense in resonant circuits (like in a push-pull SMPS with proper dead time) or when there is other means to control di/dt.

[mgm2000ro]

i did apply power and loaded it on 1 ohm woofer playing music and it blew when i turned the volume up ... at about 1000-1500W ...

mounting problem,your clips is to large !!!

[Pafi]

Originally Posted by Eva

Waveforms look perfectly normal for car-subwoofer grade class D.

Are you sure? Notice that falling and rising edge on output are very different! This can't be normal! If the slow one is normal, then the fast one is not.

edit:
With R192=220 ohms the gate waveform is understandable, but what kind of idiot place there such a high resistor there?

[Eva]

Originally Posted by Pafi

Are you sure? Notice that falling and rising edge on output are very different! This can't be normal! If the slow one is normal, then the fast one is not.

I was talking about the gate waveform. It's true that the fast rising edge in the switching waveform is telling "risk of overlap, more dead time on that side please", while the other looks more or less right (could use less dead time).

However, have you ever repaired one of these? They are intended for LF output only, >250Hz, switching frequency is usually in the 50khz range, and the discrete gate drivers they use are ugly. Many use electrolytic capacitors in the output filter, and nearly all use either pre-filter feedback or no feedback at all

So what I mean is that good symmetry is not even attempted, and at the low di/dt rates they operate (mandated by dumb PCB and junk MOSFET) overlap for a few dozen ns is not really a problem. They may take a whole microsecond of overlap to develop destructive currents.