Total dead-time for MOSFETs.

General topics and discussions about the VESC and its development.
ThierryGTLTS
Posts: 86
Joined: 09 Aug 2017, 11:10

Total dead-time for MOSFETs.

Postby ThierryGTLTS » 12 Sep 2017, 07:32

Good Morning everybody,

I noticed that the dead-time of the MOSFETs seems to be composed of 3 parts.

1. HW part DRV8302 (or 01), 10K sets 80nS dead-time, 50µS the min possible + 3nS/1K of resistance.
2. In mc_interface.h #define mcpwm_dead_time_cycles 60, but I ignore the lenght of one cycle, help, please :!:
3. In BLDC Tool or VESC Tool, you can find dead-time compensation, preset to 0.08µS.

So is the total dead-time equivalent to the sum of those 3 parts, and how long (in nS) is the lenght of a cycle :?: :!:

Have a Nice Day.

Thierry

ThierryGTLTS
Posts: 86
Joined: 09 Aug 2017, 11:10

Re: Total dead-time for MOSFETs.

Postby ThierryGTLTS » 14 Sep 2017, 07:31

I've discovered tests done by Benjamin on his "Vedder" website.

The measured risetime seems to be 560nS, so we can imagine that the deadtime is longer than that.

high-side-gate.png
high-side-gate.png (4.6 KiB) Viewed 1246 times


Anyone who knows what is the total deadtime or how long is a cycle :?: :!:

Have a Nice Day.

Thierry

Prevas_erra
Posts: 24
Joined: 11 Apr 2016, 16:03
Location: Stockholm

Re: Total dead-time for MOSFETs.

Postby Prevas_erra » 14 Sep 2017, 08:06

Hi

1. 50 nS minimum added to the handshaking. There is an internal hand shake between the high side and low side MOSFETs
during switching transitions to prevent current shoot-through.
2. I think the timers are clocked with 168 MHz so one tick is 5.95 nS (when the value is 0 to 127).
3. This is used by SW to compensate for the HW dead-time, not added.

ThierryGTLTS
Posts: 86
Joined: 09 Aug 2017, 11:10

Re: Total dead-time for MOSFETs.

Postby ThierryGTLTS » 14 Sep 2017, 08:50

Thanks for your answer.

So you have a 50 x 5.95nS = 357nS from #define mcpwm_dead_time_cycles 60

A 80nS defined by the Pin 7 10K resistor of the DRV8302 and the deadtime compensation allows the user to decrease it.

After better reading of the datasheet, I see: The DRV8302 uses automatic hand shaking when the high side or low side MOSFET is switching to prevent current shoot-through.

I need a good pair of glasses :lol: ;)

Have a Nice Day.

Thierry

rew
Posts: 943
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.

Re: Total dead-time for MOSFETs.

Postby rew » 14 Sep 2017, 17:24

The biggest problems come from the FALLING edge. When the gate is driven low, the mosfet starts to turn off. When that turning off takes too long, the other side may already be turning on. That's when shoot-through happens. I had this happening when I used different mosfets with a much lower threshold voltage. The "1A max output current" of a gate driver means they are using a 12 Ohm output mosfet on the gate driver. Then when the threshold voltage is say 2V, then at 3V there is only 0.25A running....

ThierryGTLTS
Posts: 86
Joined: 09 Aug 2017, 11:10

Re: Total dead-time for MOSFETs.

Postby ThierryGTLTS » 15 Sep 2017, 07:12

Thanks for your comment Rew.

I wanna use CSD19536ktt from TI with much lower Qt.

Also planned a diode and a smaller resistance in parallel with the 4R7 of the gate to accelerate the discharge of the cap.

Hope it'll improve the switching.

Have a Nice Day.

Thierry


Return to “General”

Who is online

Users browsing this forum: No registered users and 2 guests