This is important stuff that people often learn the hard way. If you have a pretty beefy mosfet, gate resistance too small; mosfet goes bang. Gate resistance too large; mosfet goes bang. In both cases, often something else goes bang, like the gate driver and possibly as far back as killing the controller before that. Things really go bang when the mosfet fails as a short and its in a H-bridge or something because then you can short your extra beefy power supply and all the magic smoke comes out.
As I understand, to prevent oscillations the minimum resistance is: Rg min(non-osc) = 2×√ Lg/Cg. (Lg = parasitic inductance, not every datasheet have it) Rg in the datasheet is pretty close to it, I think. The maximum - is a compromise inbetween the desired switching losses and VDS max of the MOSFET vs. dI\dT (the voltage surge when the FET is off). Note that the driver resistance and the MOSFET internal gate resistance are also counted. For example: IRLR120N Rg min = 2×√ 7,5nH/0,44nF = 8.2 Ohm. But these minimal 8.2 Ohm = the actual resistor + the driver resistance + the connections and PCB traces resistance + the internal MOSFET resistance.
Definitely makes a difference having that resistor. I managed to find the value using a potentiometer and oscilloscope. Correct value drastically reduces "ringing" and makes amplified signal cleaner and stronger. I used rpi pico as a 7mhz square wave source, act244 as a mosfet driver and two parallel 12v powered bs170s as amplifiers. The resistor value varied between 220 and 370 ohms depending on mosfet load.
@Vojislav Great comment. Gate resistance also depends on Vds and gate lead inductance. It's also very layout specific as a result - when testing on oscilloscope, did you use PCB or breadboard?
@@subhobroto I used a breadboad, and it is not a nice setup :) just proving a concept though. I expect the value to change once on a PCB, so some more experimentation and calculation will be needed.
A resistor to ground or Vdd is also wise if you want a fail-safe design. If the input to the FET gate is floating (which it is until the mosfet or driver is turned on (and running the appropriate code section), then it easily can float to turn on, partially on or off. If using a micro, it can be several ms or 100s of ms before the output pin is driven with a known state. Also, low resistor does not increase heat, fast switching does (since the average current does) and also a slowly rising and falling gate voltage will cause the fet to be in the linear (ie resistive) region too much of the time compared to the signal it is driving, that translates to efficiency losses and heat.
> A resistor to ground or Vdd is also wise if you want a fail-safe design. If the input to the FET gate is floating (which it is until the mosfet or driver is turned on (and running the appropriate code section), then it easily can float to turn on, partially on or off Indeed - one way this manifests itself is when the gate resistor blows, and the designer did not put in the pull down. The gate's now floating and cascading failure ensues. I have no clue why designers miss this failsafe detail
There is another reason. To prevent a low power driver output to be blown up. The gate capacity can draw huge peak currents to charge the gate capacity when the FET is switched on. For sure with microprocessor outputs connecting to the gate it is advisable to use a resistor.
I don't think, who connects the MOSFET gate directly to the MCU pin? the G should connect to the MOSFET driver. And also with the drive, the resistor should be used, not for huge current but for noise and efficiency too.
It took me far too long to figure out what you mean by "Dyabadeety" 😀 Also, shouldn't there be another resistor to ground to decrease depletion time in the transistor?
Thx for sharing, however the Miller effect has the most profound effect on switching speed. What size range of resistor are you talking about 1Ω, 10Ω, 100Ω 1KΩ. It takes amperes of current to push the output fast against the Miller capacitance. A sim or lab work would have been most helpful but you did get a click outta me.
I will also often add a small value of resistance as close as possible to the output pin of the driver IC as well as at the gate. That helps soak up reactance in the circuit board trace, or, lead wire between the two devices. It can be a very low value, ~2.2-4.7 Ohms. I agree, the bulk of the total resistance should be at the MOSFET gate. An individual resistance should be used on each gate, if running several FET's in parallel.
Do the mosfets usually blow if the resistor is blown. My mosfets look good on my inverter but my resistor was blown. I didn't replace yet but wondering what else I should check
With resistor value too small, some microcontrollers fail to convey the output logic state. The software drives a '1', sets the register latch but the output latch stays at '0'. Because 'pulled down' by the gate capacitance.
Sir, I have a mosfet: Qg=200nC and tr=100ns, it mean's in theory I just need 2A to drive it but even I use a 4A out gate driver, it still not enough....why? the frequency is too fast?
I love your short content and the summarize. I would like to see real oscilloscope images Or simulation? I would assume this could be simulated if you add the other parasitic components
Please please please read the datasheet that comes with your mosfet driver before following this advice. The datasheet will nearly always tell you the prefered set up, drivers like Microchips MCP1415 specifically ask for it to share the same pad as the mosfet its driving with a diagram showing that the pins of both share the same copper fill area.
Power loss is the power (V x I) lost during switching and during steady state. The power is lost due to dynamics of Vds, ID and Rds(ON). Current flowing through the Rds(ON) value of the MOSFET causes I^2R power losses. Entire power lost is in terms of heat. For saving the MOSFET device from failing due to overheating, the power needs to be dissipated (thrown away). Power dissipation is regarding how effectively and how much, the MOSFET package is able to convey out the heat generated within the junction while the Id current is flowing. The power dissipation is entirely an ability of the MOSFET package. User can effectively dissipate (to the cooling medium - Air) the entire amount of power losses if one has effective heat sink in place. Hope this helps.
sir can you please explain whats voltage over shoot mean in a mosfet and how to calculate the gate resistor value ,how does a high gate resistor value will increase power dissipation of a mosfet please explain
The current flow can be huge. Because at turn on the gate capacitor is discharged, it looks like a short circuit to the gate driver. The ONLY current limit is the gate resistance.
There's always a possibility that the MOSFET could short circuit and blow the computer being used to drive the device if there's no resistor, especially if the parts are chinesium. It's not uncommon for people to directly connect the gate directly to the output of a computer or microcontroller. A .25 cent optical isolator could be money well spent.
Hi, will be videos of others topologies? For example, half bridge LLC Resonant? If what i don't know English, so if what - sorry for my literacy. I remember already leaving this question and you and you put a like on it.
1M is too high, usually it's recommended between 10K and 100K. The gate source resistor is required when the MOSFET gate is driven by a microcontroller that can be in sleep state that will put all GPIO pins to high impedance state
@@dragoscucu3128making a circuit to drive a solenoid valve with up to 1 amp at 12 V (followed by pwm to keep the valve open with very little power), I discovered the logic level mosfet could remain high for minutes after disconnecting when messing with it, so I added a 100k resistor to ground in the circuit. I didn't know you're supposed to put a resistor between the controller and the gate as well.
@brainwater I think it depends on the driver's topology. If it is push-pull, you should turn the MOSFET off, but if it is open-collector the MOSFET gate capacitor will not discharge, then you need the resistor to GND. IMHO, it is better to put it from gate to GND instead of driver output to GND
It is simply the driver is driving a high impedance MOSFET input. A lot of ringing happens. A small resistor is needed to dampen this. We saw it with our circuit in the lab. Don't rememebr what it was, probably in the range of 22 to 56 or so Ohms.
You don't need to calculate. The value depends on the driver. You can check the application circuit of any driver. you might get the resistor value there
When we change low current fet with high Current one we may. Decrease the value of gate. R. ...large current fet have large Gate capacitance withThe existing r the gate. Vdc raise Time is too large ...small r restore. The balance
I thought the overshoot was because of stray inductance. In an inductor the current doesn't like to change suddenly so it will try to carry on pushing current, your pushing that current into a capacitor and when you do that the voltage rises. So thats why when you try to stop the current instantly it overshoots. The ringing is because the inductor will then steal that energy back from the capacitor; the reverse happens and then it flows back again, and they argue it out between them.
@@dr_jaymz Err ok, not sure what that reply has to do with my problem understanding the speech itself, if you make a video in English then it should be understandable by the majority of English speakers or it renders the content mute.
No algorithm was given to indicate the best compromise value. In fact, I don't think there was any starting value suggested for the series gate resistance. Also a shunt gate resistance can help mitigate some performances....
I am sorry if this video disappointed you, but there is no proper calculations to calculate gate resistance, at least I didn’t find it, if you find any reference please let me know
@@FoolishEngineer Many youtubers make videos on mosfet gate resistance but do not tell about the calculation for gate resistor. I have destroyed more than 50 mosfet due to wrong resistance.
I learned that gate ringing can be a problem, though I don't believe it, but that's only because I need a second opinion, or even better, oscilloscope traces. Of course, I learned very little on how to mitigate the problem.
