Gate Resistor Calculation

[coil1002]

Hi all I am new to the forum. I have built quite a few Spark gap style coils throughout the years and am now trying my hand at my first DRSSTC. I am going to use loneoceans UD2.7 c for the driver and his 80mm Full bridge board for the bridge. I will be using FGA60N65SMD for the igbt's. I am having trouble wrapping my head around calculating the gate resistors values. Could someone point me in the right direction to a good place to read up on this ? I tried using the calculator but I am at a loss as to how to figure out the maximum gate drive current the driver is capable of. Better stated what do I calculated it off of? My assumption is the current of the mosfets that drive the bridge and if so is it x1 since there is a gate resistor per igbt or total current x4?

Any help would be greatly appreciated. Thank you

 

[profdc9]

I will venture an answer to this...

The purpose of the gate resistor is to suppress oscillations of the series LC circuit formed by the IGBT gate capacitor about 3000 pF for the IGBT you have chosen) and the inductance of the connection to the gate.  This is one reason why the connection to the gate should be kept as short as possible, and the signal is carried to the gate using a transmission line of some sort, for example, a twisted pair of a cat 5 cable which has a characteristic impedance of about 100 ohms.  Many gate drive transformers already use cat 5 twisted pairs so this is convenient.  The gate is a load on this transmission line that is very much like a short circuit being a large value capacitor, and the capacitor formed by the two conductors of the transmission line adds only a little extra capacitance to this.   If the gate resistor resistance is too large, the RC circuit formed by the gate resistor and the gate itself has a rise time that is lengthened and therefore slows down the transition of the gate.  If it is too small, you get the oscillations.  So for example, lets say you have 3 cm of trace which has a inductance of perhaps 20 nH.  The impedance of the LC circuit formed by the trace and the gate is sqrt[20 nH/3000 pF]=3 ohms.  To critically damp this LC circuit, you would use a 3 ohm resistor.  This would result in a RC time constant of (3 ohms)(3000 pF)=9 ns.  This does not increase the transition time of the gate that much.  Typically you err a little on the side of a higher resistance, so that 5 ohms is a typical series resistor value.  The value of the resistor is not that critical because in practice it is difficult to measure some of these quantities in-circuit.

Dan

Hi all I am new to the forum. I have built quite a few Spark gap style coils throughout the years and am now trying my hand at my first DRSSTC. I am going to use loneoceans UD2.7 c for the driver and his 80mm Full bridge board for the bridge. I will be using FGA60N65SMD for the igbt's. I am having trouble wrapping my head around calculating the gate resistors values. Could someone point me in the right direction to a good place to read up on this ? I tried using the calculator but I am at a loss as to how to figure out the maximum gate drive current the driver is capable of. Better stated what do I calculated it off of? My assumption is the current of the mosfets that drive the bridge and if so is it x1 since there is a gate resistor per igbt or total current x4?

Any help would be greatly appreciated. Thank you

[Uspring]

The purpose of the gate resistor is to suppress oscillations of the series LC circuit formed by the IGBT gate capacitor about 3000 pF for the IGBT you have chosen) and the inductance of the connection to the gate.

Wouldn't you have to add the GDT leakage inductance to this?

[Mads Barnkob]

Which gate resistor calculator are you talking about? I wrote this to aid myself in playing with figures for DRSSTC: http://kaizerpowerelectronics.dk/calculators/igbt-gate-drive-calculator/

[coil1002]

Which gate resistor calculator are you talking about? I wrote this to aid myself in playing with figures for DRSSTC: http://kaizerpowerelectronics.dk/calculators/igbt-gate-drive-calculator/

That would be the one Mads. I ended up using the current of the mosfet to calculate to. I came up with 3 ohms with the calculator so I bought 5 ohm resistors for tolerance

[profdc9]

If you look at the transformer with an equivalent "T" network, there is a shunt arm that is in parallel with the secondary and therefore isn't in series with the gate, and a series arm that is.  I would think the leakage inductance from the series arm is something like A_L N^2 K where A_L is the inductance of a single turn, N is the number of secondary turns, and K is the flux coupling constant (0 to 1).  I would think that a bifilar winding on a high permeability core has a K pretty close to 1, but it could be an issue I suppose.

Dan

The purpose of the gate resistor is to suppress oscillations of the series LC circuit formed by the IGBT gate capacitor about 3000 pF for the IGBT you have chosen) and the inductance of the connection to the gate.

Wouldn't you have to add the GDT leakage inductance to this?

[Hydron]

In the services of simplicity, I'd suggest trying something one of the following options:
- a single resistor in the region of 3-5 ohms (corresponding to the datasheet test spec resistance)
- a resistor as above, with an additional combination of another resistor for slowing turn-on (thus increasing effective dead-time) with a parallel diode for fast turn-off
If you're driving the gates with more than 15V then I'd lean towards 5+ ohms rather than 3.

In both cases scope the Vce and Vge waveforms - if there is Vge oscillation or excessive Vce turn-off spikes then increase the resistance. These IGBTs are pretty darn fast with low Qg, so I doubt that you're going to run into too-slow switching unless you use really large gate resistors.

[coil1002]

In the services of simplicity, I'd suggest trying something one of the following options:
- a single resistor in the region of 3-5 ohms (corresponding to the datasheet test spec resistance)
- a resistor as above, with an additional combination of another resistor for slowing turn-on (thus increasing effective dead-time) with a parallel diode for fast turn-off
If you're driving the gates with more than 15V then I'd lean towards 5+ ohms rather than 3.

In both cases scope the Vce and Vge waveforms - if there is Vge oscillation or excessive Vce turn-off spikes then increase the resistance. These IGBTs are pretty darn fast with low Qg, so I doubt that you're going to run into too-slow switching unless you use really large gate resistors.

I ended up  getting 5 ohm I will check the waveforms when its built. Thank you all for the responses they where incredibly helpful.