Hi @tarunarya1780, thanks for the question. The value of the “local” gate resistor (the one close to each MOSFET) should be low so that the gates of the MOSFETs are well coupled together, around 2 ohms should be enough - there should be a bit of resistance in the gate-source loop to dampen eventual ringing. The value of the common resistor at the driver side depends on how fast you want the MOSFETs to switch, you could use a simple simulation to choose it. Furthermore, the equation at page 17 of our application note AN50005 “Paralleling power MOSFETs in high power applications” could be helpful in certain scenarios. assets.nexperia.com/documents/application-note/AN50005.pdf With power MOSFETs, the RDS(on) decreases with increasing voltage VGS following a law of diminishing returns - i.e. usually after a voltage of around 10V for a standard level MOSFET RDS(on) won’t decrease much further. Always be mindful of the VGS rating of the MOSFET over the whole temperature range. Regarding the driver; usually, designers might prefer using an integrated gate driver because of convenience, performance and space savings. However, a non-integrated solution (like a charge pump) is still possible and sometimes even preferable - due to cost for instance. Check out the design at page 14 of our technical note TN90002 “H-bridge motor controller design using Nexperia discrete semiconductors and logic ICs” - it shows a very simple gate driving circuit that can be designed with common BJTs, diodes, resistors, and capacitors. assets.nexperia.com/documents/technical-note/TN90002.pdf. We hope this helps. Have a great day =)

@@Nexperia Thanks. I really appreciate your replying and pointing me to documents that should help. "All" I wanted to do was make an RC lawnmower using wheelchair motors 24v 250w motors, and make a controller for a brushless motor whose controller failed. I thought the sabretooth 2x60 was expensive so I would make my own. I had no idea that there were so many considerations with nuances. I will persist thanks. I'm sorry to be cheecky and ask another question. Is there a reason why people wanting to dc pwm controlled brushed motors don't just use one mosfet or parallel of mosfets to control a motor with a relay to reverse the direction rather thanan h-bridge , as this would save energy and the components?

@@Nexperia regarding "Regarding the driver; usually, designers might prefer using an integrated gate driver because of convenience, performance and space savings. " I have not actually tried drivers yet though seen examples. I thought that they needed to have the voltage to be supplied to the mosfet gate separately, just as with an optocoupler. I did not realise that they boosted the voltage from the microcontroller pin. I have I misunderstood what you have said. That is why I thought you may need 3 voltage levels - processor logic, gate pin drive voltage and load voltage. This is why was thinking of charge-pumps, or possible buck-boosters to provide a raised voltage to drive the mosfet rather than an additional power supply.

Thank you, we love your movies ;)

Hi Jack! The capacitors represent the parasitic input capacitance of the paralleled MOSFETs. The circuit illustrates the principle of splitting the gate resistor and why it’s needed (due to the capacitors). More info about this can be found in our interactive app note IAN50005 “Paralleling power MOSFETs in high power applications” - www.nexperia.com/applications/interactive-app-notes/IAN50005_paralleling_MOSFETs_in_high_power_applications.html

Thanks for the question, @apoorvapatwardhan9394. We don’t recommend connecting a resistor in the source since it will significantly affect your overall efficiency. Usually, we are paralleling MOSFETs if the current is too high for a single MOSFET to handle, and if you have a resistor in the MOSFET’s source pin (Rsource), it will add loss to your system (Rsource*Id^2). In addition, another purpose of the individual RG is to minimize the impact of the difference in the threshold voltage (VGSth) of the MOSFETs. The Millar plateau voltage occurs after threshold voltage is reached, in this region the gate current will be flowing via the drain and not the source of the MOSFET.

Hello, thanks for your question! As a general rule of thumb, paralleling devices with different part names, either BJTs (as in your case) or MOSFETs, is not recommended. You want the paralleled devices to act as a “single” unit, i.e. as synchronized as possible, so that your design is robust and reliable.

@@Nexperia thanks for the reply!. I just wanted to power up a 4W LED light (20 LEDs, 0.2W each) from a 4.2V li-ion battery pack (1S 3P). Sorry!, I forgot to mention that. I already tried that, and it's working. They barely heat up. I was able to get 1.25A for them. I used separate base resistor for each transistor. It's not ideal, but works for the LED light.

@@Nexperia I saw a similar circuit somewhere else, but in that circuit they've used D882 BJTs, 4 of them in parallel with separate resistor to drive their base. I just wanted to try that with different types of BJTs. I'm using it about a week now, so far so good 👍.

They must be biased to saturation when 'on', so the collector to emitter resistance is as low as possible. Otherwise axsmall resistance will be present, which causes overheating. Also when switched off, they must be switched off completely.

Agree, which is similar to what happens to MOSFETs where RDSon will change with respect to VGS. Therefore, one would want to increase as much and as safely as possible the VGS to get the lowest RDSon and thus lowest dissipation.

Hello, thanks for your question! Keep the MOSFETs close together (better thermal coupling), reduce the distance between driver and MOSFETs (to keep the gate-source loop as small as possible and reduce layout parasitic), employ a larger gate resistor common to all the FETs and a smaller one for each - close to the gate (improve sharing during switching). Finally, use parts from the same order - it will increase the chances of using parts from a same batch/lot, i.e. with minimal parameters’ spread between them.

Hi @flyback_driver. Thanks for your question! If you are referring to oscillating circuit as in circuits where oscillations are an undesirable behaviour then: oscillations are always bad, especially when paralleling because MOSFETs should operate “in synch”. This is why it’s always worth including a small resistance in the gate-source loop to dampen any unwanted oscillation and keep the loop as small as possible, by placing the gate driver close to the MOSFETs gate. Good layout can also help to reduce parasitic inductance and/or make it so each MOSFET power loop has the same impedance. If by oscillating circuit, you refer instead to a 1-port device able to generate an electric signal then I wouldn’t expect an oscillator to require paralleled devices - as the “heavy lifting” may be done by an amplifier where paralleling is more common practice.