[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Terry's sisg paper

Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx>

Hi Russ,

At 10:30 AM 7/27/2006, you wrote:
I am reading your sisg.pdf paper. Great work! I have two questions.

1) If the primary freq is 250 khz and peak current is 500 A. you state that the rate of current increase is 500 A/us. If the primary current is modeled as a sin wave
near time = zero, we have:

I = Ipeak * sin(w*t)
dI/dt = w*Ipeak*cos(w*t)
dI/dt = w*lpeak                  near t = 0 so
dI/dt = (2*pi*250.*10^3)*500 = 785 A/us

Your right. I left out the leading "w" in the second line when I did the derivative quickly in my head at the time.

2) The 192 ns you calculate to charge the IGBT gate cap to 9 volts assumes 25 volts on C1. but it took 100 ns to charge C1 to 25 volts. So don't you have to add 100 ns to the 192 ns to get 292 ns of time before the IGBT turns on?

Yes again ;-))

Beware that these calculated numbers are "rough" since the IGBT's collector to gate capacitance feeds current back to the gate trying to keep the IGBT off. There are some scope pictures of this at:


Don't read the rest of the paper since it is completely out of date now :o)) But I think I will try 51 ohms for the gate resistor instead of 100 ohms to speed that up just a bit. The original value was just a guess. I added the updates:




Thanks, Russ