Good Monring,

I'm going to use a Full Brigde Inverter for my application. The problem is that the current wich flows through the MOSFETs during when they are on can achieve 70A rms and the power dissipation is too high. In fact using a IPW65R019C7 one obtain

Pd=Rdson*Ieff^2=0.019*50^2=47.5 W.

I would like to reduce this power losses making two MOSFETs in parallel. In this way the current is divided by 2 and one obtain:

Pd=Rdson*Ieff^2=0.019*25^2=11.8 W.

The question is: are these calculation correct? Do you know the rules and precautions to ensure that the current flows half in one MOSFET and half in the other one?

I thought to use an L 6491 as MOSFET Driver. Can I use just one MOSFET driver to drive both of them I need one Driver for each MOSFET?

Thank You,

Fabio.

 

Read the following app notes for authoritative information of paralleling FETs.

Using power MOSFETs in parallel
Paralleling Of Power MOSFETs For Higher Power Output

The current won't be divided by 2 strictly. The MOSFET with the lowest RDS(on) value will take the highest proportion of the current. It therefore has the highest power dissipation. Using two equivalent MOSFETs will allow the current to divide by 2, but one of them will inevitably draw a little more than the other.

As for the drivers, just one is necessary to drive both FETs.

 

Not so sure about the driver, its rated at 1000 pF, its AC specs.

But your MOSFET ratings (3X) are -



I would spend some time into looking into this.


Regards, Dana.

 

Actually any regime in which mosfet works with less than half max power ought allow (no-one yet may swear it) to work brutally. With demand that paralelled are from one and the same stock. Then even if some 20% differences will arise, it bill not be devastating.
In all other cases the source balancing ohm tenth must be applied (say so let the full power voltage on this resistor be about 1V +/= 100%. However, the all speeds of switching will decrease unavoidably. The gates are connected, of course only far side of gate resistor, 3 ohm...5 ohm or 10 ohm. Just explore by oscillo what is the best value. This resistor must be non-inductive, therefore parallel tank of SMD resistors is the best choice for any gate. If too much tended for oscillations, may apply the small ferrite ring (3...5 mm) over the gate wire.

 

if this approach is for product, you would need to calculate the thermal dissipation in a worst case.