EEVblog Electronics Community Forum

Electronics => Beginners => Topic started by: strawberry on May 17, 2018, 12:20:44 pm

Title: IGBT turn off delay and dead time
Post by: strawberry on May 17, 2018, 12:20:44 pm
FGH40T65SPD  Td off =  38nS   1370pF 

IKW40N65ES5 Td off = 184nS   2500pF

Do Infineon make slower transistors or is it the way they test them?

How to estimate "dead time" ?

How much gate drive current is sufficient for 15nF IGBT gate capacitance?
Title: Re: IGBT turn off delay and dead time
Post by: David Hess on May 17, 2018, 06:52:13 pm
The testing conditions are different but it looks to me like the Infineon part is slower.  There is a trade off between storage time due to minority carriers in the bipolar element and collector-emitter saturation voltage so the Motorola part is faster with a higher saturation voltage and the Infineon part is slower with a lower saturation voltage.
Title: Re: IGBT turn off delay and dead time
Post by: AQUAMAN on May 17, 2018, 06:56:53 pm
FGH40T65SPD  Td off =  38nS   1370pF 

IKW40N65ES5 Td off = 184nS   2500pF

Do Infineon make slower transistors or is it the way they test them?

How to estimate "dead time" ?

How much gate drive current is sufficient for 15nF IGBT gate capacitance?
How much gate current you need depends on the switching frequency
Title: Re: IGBT turn off delay and dead time
Post by: strawberry on May 17, 2018, 08:28:17 pm
Isn't that Infineon count antiparallel diode Trr into Td off?

For 100kHz (rise time 100nS), Vge = 12V, Cge = 6nF   estimated driver current 0.72A seems low
Title: Re: IGBT turn off delay and dead time
Post by: T3sl4co1l on May 17, 2018, 10:04:36 pm
By the way, note why Td(off) is slower in the first place: when Vge is rising, it's going from 0 to 15V (or whatever they tested at -- I didn't look, it's usually this), and the Miller step falls at, say, 6V.  So the gate charges through the total gate resistance (including driver, resistor and internal R_G) from a difference of 15-6 = 9V.  When Vge is falling, the Miller step falls at maybe 3 or 4V, so the gate discharges through only 3-4V, or taking about three times longer.

You shouldn't drive IGBTs with Vge(off) = 0V, anyway -- besides slow turn-off, they're prone to shoot-through during positive dVce/dt.  Better to use -5 to -15V Vge(off), which ensures sharp turn-off as well. :)

Tim
Title: Re: IGBT turn off delay and dead time
Post by: T3sl4co1l on May 17, 2018, 10:05:44 pm
Isn't that Infineon count antiparallel diode Trr into Td off?

For 100kHz (rise time 100nS), Vge = 12V, Cge = 6nF   estimated driver current 0.72A seems low

Cge?  Don't use that, that's a small signal parameter -- useless.  Use Qg(tot).  I(drive) = Qg * Fsw, simple as that. :)

(Note that Qg(tot) goes up if you drive more voltage swing, as I recommended above.)

Tim
Title: Re: IGBT turn off delay and dead time
Post by: strawberry on May 20, 2018, 07:57:15 pm
Qg(tot) is for loss calculation

turn off current must be 3*0.72A=2.16A

Is it possible to design dead time circuit without delayed signal?