IR2110 H-Bridge Inverter for Induction Heater

[h1mel]

Hi,

I was designing a High Power High frequency Induction Heater inverter. I've choose H-bridge configuration for the inverter
and IR2110 gate drivers to drive the bridge. As the IR2110 is a half-bridge driver i'll use two of them to drive the both sides.
I've done a rough circuit(just drawing). As i don't know much about power electronics i wonder if there is any mistake.

i am looking for advice, recommendation, and any kind of error that should be fixed in my circuit.

Also I've some questions:

1) what is the use of C3 n how i'll calculate the value?
2) Exactly which back EMF protection Diode should i use as D' ?
3) Do i need to use an opto-isolator for high side nMosfets?

The intended frequency is 150KHz and the Mosfet's are N-type.

 

[Antiphon]

I would replace the 1n5818 with something higher voltage and lower current. Even a 1n4148 would be better but not as good as say a STTH1L06. D2 and D3 will likely see 30v transients as the gate drive line rings.

In position D1 you would destroy the circuit the moment it powers up. When q1 closes, the inverse voltage is around 260v. White smoke and dead circuit with the 1n5818 rated for 30 volts.

C3 is a strait forward bypass cap. 100nf should do it. C2 needs some special consideration though as it holds up the high side driver during the time q1 is on. It's value will depend on the frequency and maximum duty cycle. The application note for the 2111 should give you this formula.

No opto needed, that's what the 2111 is for.

D' is built into the fet. You didn't specify it; that's something that will make or break your driver.

If you run the inverter correctly you will not need to pass large currents through the body diodes. You will instead close the fet and let the transistor do the work.

This is not a trivial design exercise. Expect a lot of learning to take place and a lot of blown prototypes.

 

[DragonPetter]

One suggestion I would make for diode D1 is to choose a much faster reverse recovery time diode. This also depends on your power supply that you are bootstrapping from, because if the diode does not switch fast enough, you have current go back into your supply and can ruin it.

I had this problem using a gate driver on a BLDC motor.

 

[h1mel]

I would replace the 1n5818 with something higher voltage and lower current. Even a 1n4148 would be better but not as good as say a STTH1L06. D2 and D3 will likely see 30v transients as the gate drive line rings.

In position D1 you would destroy the circuit the moment it powers up. When q1 closes, the inverse voltage is around 260v. White smoke and dead circuit with the 1n5818 rated for 30 volts.

I am not sure if i would find the STTH1L06 in my country. But still i'll search for it. Can you recommend another high power diode fast enough to charge the bootcap at 150 KHz switching speed??

 

[h1mel]

there is another cap between the Vdd and Vss in the datasheet of IR2110. is it mandatory in my circuit?

 

[Mike_In_Plano]

I'm not a fan of the 2110. For this application, I'd use a dual low side gate driver and a gate driver transformer. One of my favorites is the Pulse, P0584NL.

I drive the gates on through a resistor to delay turn on. Off through a schottky rectifier to ensure rapid turn off. I place back to back zener diode across the gate to source leads to proect the Mosfets.

Some Mosfets survive bridge applications better than others. As I recall, Infineon made good bridge parts.

 

[Antiphon]

I've had really good luck with STMicro MOSFETs. Unlike Infineon they include the gate Zeners in the package. Infineon has better overall specs on their fets though.

Also, I've used the IR2111 drivers wih great success. For a while my bridges were blowing up sporadically because I was running Vcc too close the max combned with inadequate gate damping. I dropped my Vcc to 12V and damped all the gate lines.

Haven't blown another fet since 2005 (have 2,400,000 unit-hours continuous with no failures).

There are lots of diodes that would work. You want an ultrafast rated for more than twice your peak supply voltage (a 600V diode would be fine). I think there's a diode common in Asia called SROM that would work for you.

 

[Antiphon]

I'm not a fan of the 2110. For this application, I'd use a dual low side gate driver and a gate driver transformer. One of my favorites is the Pulse, P0584NL.

I drive the gates on through a resistor to delay turn on. Off through a schottky rectifier to ensure rapid turn off. I place back to back zener diode across the gate to source leads to proect the Mosfets.

Some Mosfets survive bridge applications better than others. As I recall, Infineon made good bridge parts.

I use 1n4148 for the gate turnoff. Are you using a Schottkey because of a large gate capacitance?

 

[h1mel]

I've had really good luck with STMicro MOSFETs. Unlike Infineon they include the gate Zeners in the package. Infineon has better overall specs on their fets though.

Also, I've used the IR2111 drivers wih great success. For a while my bridges were blowing up sporadically because I was running Vcc too close the max combned with inadequate gate damping. I dropped my Vcc to 12V and damped all the gate lines.

Haven't blown another fet since 2005 (have 2,400,000 unit-hours continuous with no failures).

There are lots of diodes that would work. You want an ultrafast rated for more than twice your peak supply voltage (a 600V diode would be fine). I think there's a diode common in Asia called SROM that would work for you.

Thanx for the feedback! Now time to try out the whole thing in real! :)

 

[Mike_In_Plano]

In the nineties, we were using gate drive transformers with a high carrier frequency and rectification circuit to drive our motor IGBTs for motor drives. We attempted to use 2210's and another custom part, based around the silicon well technology.
In both cases, our failure rate during qualification went up remarkably.
After that, I've used either gate drive transformers or opto-isolator devices - each with great effect.
Of course, grounding is a big issue with the 2110's, but I don't believe that was our problem. We had a group of sell seasoned power electronics engineers on this.
I've also heard that packaging / contamination was a problem with the early devices.
This I'm more inclined to believe. I've seen these issues repeatidly - especially in rectifiers.

As to the reason I used shottky diodes, it probably started because we generally used them anyway and it was good to keep the bill of materails short. It also helped that they had a low forward drop and reacted well to the very short over current they experienced.

It was important to stick with lower voltage, moderate current, power schottkys. As they get larger, the charge transfer can be really high. The same for silicon power rectifiers that exceed 200 PiV.

As to MOSFETs, I designed bridges that did not have parallel rectifiers across the MOSFETs. From this, we learned that some MOSFETs hold up better than others. There's some odd physics involved in using the body diode in freewheeling mode. As I recall, the manufacturers said little about this, with the exception of infinion. They actually designed for this application.

I'm curious what the state of the art is now. I have friends that design resonant bridge converters, and I've never seen them use rectifiers across the bridge transistors, though I do think they use so sort of solid state high-side drive.

I'll have to get back on this.

- Mike