I am confused as to how to calculate the Power which a flyback Diode will dissipate so that I can work out the correct sizing of the heatsink to keep it within limits.
My circuit consists of a 2 (two) ohm load (with my flyback diode in parallel), which I will control at a maximum PWM of 1KHz using an IGBT. Voltage is around 110V. So 55A Current Pulses. Duty Cycle will be between 0 and 70%.
Inductance will hopefully be less than 1uH.
SO the question is how do I calculate the heat which will be produced by the Flyback Diode so I can heatsink it appropriately.
When the current pulse is 55A, the flyback current will also be 55A. But it will be a short pulse and it depends on the speed of the IGBT and the inductance of the load. But also the inductance and resistance of the wires to the diode itself has influence. So it is very hard to calculate.
How fast will the IGBT turn off ? That depends on the IGBT itself, but also on the input voltage of the IGBT. The input is a gate with capacitance which slows it down.
That diode can do a few hundreds Amperes peak current and is 40ns. The heat in the diode depends mainly on the frequency of the PWM signal and the inductance of the load and the forward voltage. What frequency will you be using ? A normal value, below 1000Hz ?
The forward voltage is 1.2V or 1.8V or 1.55V (I don't understand the datasheet very well).
IF I PWM , it'll likely be at 1KHz, Decision to PWM will eb based on the amount of ripple I get. Otherwise I will switch on the load at a set voltage and off at a set voltage. Depending on how close they are will depend on the on/off frequency as the Load should draw double the current actully being produced by the generator.
I also didn't understand the datasheet well, so I was going to go for worst case 1.8V and see what figures it came up with. I could always parallel the Diodes if the heat was likely to be too great.
So Assuming no additional inductance / resistance in the resistor to Diode wires. The resistamce of the Load is 2 Ohm, the Inductance of the load is 1uH and the turn off time of the IGBT's being (worst case - 600V/ 20A, Rg= 10Ohm, Vge=15V and Tc=175C) Td(off) 345ns and tf 520ns. Input Gate Voltage will be 15V.
I couldn't work out how to include the various timings etc into a heat dissipated calculation.
The gate driver with 500ns also slows things down. But it should be okay for up to 1000Hz of course.
That IGBT is maximum 20A at 100 °C and 60A for a pulse, so I hope that the resistor will smooth things and they will work all the same. From what I have read, the resistor of 5 ohm or 10 ohm should be okay.
When one of the IGBTs is getting hotter than the rest, you can give that one the resistor that has a slightly higher value.
I like the speed of that diode of 40ns, so almost any diode with 40ns and 1200V and 55A will be okay.
The good thing about the flyback diode is that you can use as many as you like.
You can have one as close as possible next to each resistor and one near the IGBTs as well.
I don't know how to calculate Joules into Watts. But a rough estimation tells me that the diode gets less than 0.1Watt (I could be totally wrong).
When I look for high current switching circuits, I see sometimes LC filters for the power supply and RC snubber circuits. You might want to add that to reduce RF pollution.
I think you have to build a prototype and do some tests. Will the gate driver work well with 5 IGBTs ? Will the resistors mechanically break down with a PWM ? Is something getting to hot ?
I agree, it'll be trial and error , but trying to rule out as many issues before as possible.
If it's only 0.1 Watt, great. I'll see. Haven't switched 55Amps in the past, so never really worried about it.
I intend to heatsink all 5 together, but that isn't so easy with a passive heatsink. I looked at other IGBT's but had issues trying to find one's which switched reasonably fast and with ample leeway which I wouldn't have to parallel. I saw this one http://www.irf.com/product-info/datasheets/data/irg7ph50upbf.pdf
However as it doesn't have a RR diode built in , what kind of Diode would be appropriate for this and again, would I need to worry about dissipation if I'm now switching 55Amps?
OR (I know bad English to start with an or..), do I not need it if I already have the flyback diode on the load?
I also have a 9Amp and 12Amp driver to try if necessary but would need to add a separate optoisolator onto them. Any suggestions of a good opto-isolator for PWM work? Looking through datasheets on the web is a real pain. I prefered the old days when we had books...
Resistors should be fine. They are the welwyn WDBR2-12Ohm and will parallel 6.
At www.farnell.com search for "high speed optocoupler" and select the group with transistor output.
At www.mouser.com you can select the speed.
The 6N135 are called 'fast', but that is for serial data and so. Today they not so fast.
According to the datasheet of 6N135 it can do 1MHz, but it has a rather large propagation delay.
