thingmaker3 -> Cboot calc.

Thread Starter


Joined Apr 11, 2008
This question is not just for thingmaker3 if anyone else has the answers I seek.
Cboot > (Qg + Dmax*Tsw*Idrv ) / ΔV

Qg is the gate charge of the MOSFET
Dmax is the maximum duty-cycle​
Tsw is the switching period​
Idrv is the average consumption of the driver​
ΔV is the ripple you accept across the bootstrap capacitor​

So I tried the calculation its much more simple then the one I was going to use from the application guide, but I'm getting a fairly large value.

So lets do one.
Cboot must be larger than: (Qg + Dmax*Tsw*Idrv ) / ΔV

(.00000017 + .75*20000*.2) / 2
is equal to 1500.000000085. This would appear to be 1500 farads?
What am I doing wrong?

Or, if I don't convert to actual values I could do:
(170nC + 75% * 20khz * 200mA) / 2v
and get a value of 235 which I could assume means 235 picofarads?

So a tally of the issues I'm having:
1) Did I do any of the calculations right?
2) Do I use the average current of my driver or peak pulse current?
3) What happens when you get to duty cycles of 1% and 99%?
4) A 2V ripple? Meaning I'm driving with 12v and will allow it to sag to 10v.



Joined May 16, 2005
The formula, by the way, comes from Christophe Basso, author of Switch Mode Power Supplies: SPICE Simulations and Practical Designs.

Thread Starter


Joined Apr 11, 2008
Why are you using a six hour switching period?:eek:

Oh, I see now, you've substituted the frequency for the period.:cool:

Mmk so I should use switching period instead?

The switching period is 0.00002 for 20khz? (I'm not sure on this).

When I do (.00000017 + .75*.00005*.2) / 2
I come up with .000003835 or 3.8uF.
Which seems closer to reality.

Thread Starter


Joined Apr 11, 2008
Ewps, yes, I appeared to get the correct answer that I used in my calculation but typoed the first one to .00002

Guess my application of math was good, but now that I know its the inverse multiple, that will help for the memorization.