H-Bridge Driver Design and Shoot-Through

RPLaJeunesse

Joined Jul 29, 2018
252
500K is exceedingly high, and could make the circuit sensitive to noise and stray path leakage. I consider 47K a high number, and would prefer to see more like 10K. Given a 12V LED load and a DMC3021 safe max of 7.1A steady state at turnoff you possibly would hit maybe 3V at 3.5A max, from the DMC3021 curve for SOA (fig. 11) the part will survive an over 10ms turnoff. But that is one pulse at 25C case and 125C chip temp rise. Your case temp will be higher, the allowable rise lower, and repetitive pulsing. I'd apply a safety factor of 10x and stay below 1ms turnoff. Using 47K meets this comfortably.

Side note: If you are PWM dimming (and you didn't say you were) I've learned 15Hz is very flickery, 48Hz is tolerable in a dark, stationary environment, and 100Hz+ is called for if there is motion between LED and viewer. AT 100Hz a 1ms turnoff is too long, more like 100us maximum. Using 10K resistors should be fine for this speed.
 

Thread Starter

bassel@gnc.net

Joined Sep 20, 2011
12
500K is exceedingly high, and could make the circuit sensitive to noise and stray path leakage. I consider 47K a high number, and would prefer to see more like 10K. Given a 12V LED load and a DMC3021 safe max of 7.1A steady state at turnoff you possibly would hit maybe 3V at 3.5A max, from the DMC3021 curve for SOA (fig. 11) the part will survive an over 10ms turnoff. But that is one pulse at 25C case and 125C chip temp rise. Your case temp will be higher, the allowable rise lower, and repetitive pulsing. I'd apply a safety factor of 10x and stay below 1ms turnoff. Using 47K meets this comfortably.

Side note: If you are PWM dimming (and you didn't say you were) I've learned 15Hz is very flickery, 48Hz is tolerable in a dark, stationary environment, and 100Hz+ is called for if there is motion between LED and viewer. AT 100Hz a 1ms turnoff is too long, more like 100us maximum. Using 10K resistors should be fine for this speed.
I'm sorry but I misspoke (and you are correct, 500k would be way too high).. I'm using 500 ohms for the GS resistors, not 500k. That gives a fast enough turn-off to avoid shoot-through. I'm not using PWM dimming; The LEDs need to flicker by design. They will likely be pulsed at a rate of about 5Hz.
 

RPLaJeunesse

Joined Jul 29, 2018
252
Over temp the guaranteed VOL617A-1 current transfer ratio can drop down below 32% (datasheet 25C - 40% min * 0.8 per fig. 11). With 5mA LED drive that is only 1.6mA transistor current. You want 6V (ideally) from gate to source at that 1.6mA. That means the resistors should be at least 3.75K Ohms for consistent operation. So 4K3 or 4K7 as G-S resistors would be a good choice, and will still result in a good, fast turnoff. Using 500 may work right now because you have a high CTR optocoupler and/or are slamming the LEDs. But with 500 Ohms don't be surprised if the FETS burn up way sooner that you expect.
 
Top