IR 2110 - Full H bridge configuration at 40 Khz, IGBT output.

IR2110 HI-SIDE auto-on problem

I made an inverter at 40Khz, full H-bridge using two IR2110 to drive four IGBT.
An identical configuration is just working perfectely at 25 Khz, 1.5 Kw.
I set a large interleave time between lo and hi side to be sure to avoid cross conduction given by the t_on t_off of the IGBT and I set a fix 50% duty cycle to be sure about the bootstrap charge timing.
The problem is that the hi-side goes on itself and this auto-on is syncronised with low side command in this way:

- LO-SIDE=0 / HI-SIDE=1 -> at the end of the ON period of hi-side, I see the gate signal goes ON again after few us. The gate goes off as soon as the lo-side command goes to 1.

- LO-SIDE=1 / HI-SIDE=0 -> at the end of the ON period of lo-side, I see the gate signal of the hi-side goes ON itself.


I try to midify the gate resistance (usually 10 ohm with a BA159 fast diode in antiparallel) but nothig changed.

IGBT is IXYS model IXGH20N60BD1 (with fast reverse diode integrated) and , on the paper, can normally work at this frequency.

I have this auto-on not always but only sometime and I can see that the output signal of the inverter is good.

Of course I must be care because the cross conduction can easily distroy the output stage so I must do test with low main voltage (i.2 50-60 Vdc instead of 310 Vdc from a rectified main line) but I found the auto-on signal even with this low main voltage.

Have you some notice about??

Thanks in advance

Francesco

P.S: sorry for my english .. is really bad!

Hi, I'm no expert in power electronics but i was building a plasma speaker once with two N channel MOSFETs in half bridge configuration. I noticed that they were heating a lot so i checked the voltages across drain-source.

Apparently they were turning ON while the gate voltage input was supposed to be OFF.

I asked a professor and he told me that in high frequency applications, above 20kHz, (i was commuting at 40kHz) electro magnetic noise might induce the treshold voltage (3-5V) at the gate and therefore turn on the MOSFET and short circuit the half-bridge.

He told me that the way to fix this is to wire a 10kOhm resistor between Gate and Source of each MOSFET.

I hope this fixes your problem
Good luck :grin:

I would expect the false control timing present already at HIN and LIN. Another explanation for erratic IR2110 behaviour would be a violation of allowed voltage range, e.g. due to incorrect ground wiring. How do you probe the gate-source voltage?

A 10K Rgs is usually provided to prevent static charges building at an unconnected IGBT gate. It can't help against against wrong control signals sourced by a low impedance driver.

Neto06 said:

Hi, I'm no expert in power electronics but i was building a plasma speaker once with two N channel MOSFETs in half bridge configuration. I noticed that they were heating a lot so i checked the voltages across drain-source.

Apparently they were turning ON while the gate voltage input was supposed to be OFF.

I asked a professor and he told me that in high frequency applications, above 20kHz, (i was commuting at 40kHz) electro magnetic noise might induce the treshold voltage (3-5V) at the gate and therefore turn on the MOSFET and short circuit the half-bridge.

He told me that the way to fix this is to wire a 10kOhm resistor between Gate and Source of each MOSFET.

I hope this fixes your problem
Good luck :grin:

Click to expand...

Thanks Neto06

unfortunately, I just tried this way but nothing happens. Now I'm working about selction of IGBT cause I think that they can be at the limit of their SOA.

So kind!
Francesco

---------- Post added at 17:46 ---------- Previous post was at 17:41 ----------

I test it with a 10:1 probe 200Mhz digital scope. It's the same test I have done with lower freq. (25Khz) and all was Ok.
Note that the SAME circuit, changing the working frequency at 25 Khz, is working well so now I'm working about the IGBT selection (data sheet tell me that is all right but the circuit doesn't give me the same answer !!)

Thank for your attention.

Francesco