Guru
Joined: 08/02/2015
Location: New ZealandPosts: 1252
| Posted: 07:40pm 27 Mar 2017 |
Very interesting, I like your detailed scope examination.
When we parallel many mosfets, eg 6 per leg of the bridge, I think your scope would show a different story. The typical IRFP4110BPF has a combined gate charge of 180nC, so 6 paralleled have a combined charge of 1080nC. The typical driver chip IR2110 can only supply 1.6 amps, but as it has to drive 6 devices then the max drive current per device for even current sharing is 1.6/6 = 0.267 amps, so the mosfet turn on time is now Q/I = 180nS/ 0.267 = 674 nSec, a lot more.
In this case the gate resistor should be approx 40 ohms not the normal 5.6, ie 12v/0.267a, to prevent damage to the driver.
With this driver chip if the resistor was made smaller then not all mosfets would turn on evenly as the driver cannot source enough current, so some mosfets will draw more current than others and perhaps self destruct in doing so.
If a diode is paralled with the resistor, normally this has the effect of discharging the gate charge quicker, assuming the driver can sink the increased current spikes without blowing up. A small fast signal switching diode is normally adequate, but when using totem pole drivers its not really required, the Bridge driver should have a sufficient dead time period.
If a gruntier driver is used eg TC4452, with a 13amp peak, then our 6 mosfets could be supplied each with 2 amps, requiring a gate resistor of approx 5.6 ohms, the turn on time 180/2 = 90 nS, much better. We dont want too fast a turn on of course as this can lead to device dv\dt avalanche breakdown or severe ringing waveforms, causing voltage spikes to exceed drain source limit.
When higher power fets are used, currently I'm using IRFP4468, these each have a gate charge of 400nC so its essential something like a TC4452 is used when 4 or so are paralled together.
Cheers
Mike