Replacing MOSFET, what to look for

[Futterama]

Hello forum,

I'm replacing the MOSFETs in an Electronic Speed Control (ESC) for RC cars.

My problem is, I don't really know what to look for when choosing the new MOSFET. The current rating of my MOSFETs are all higher than the old ones in the ESC.
I'm more interested in the gate charge ratings, to allow the same fall and rise times.
What is most important to take into consideration?

The following is stated in the datasheets:

- CISS Input Capacitance
- COSS Output Capacitance
- CRSS Reverse Transfer Capacitance
- Qg(TOT) Total Gate Charge at 10V
- Qg(5) Total Gate Charge at 5V
- Qg(TH) Threshold Gate Charge
- Qgs Gate to Source Gate Charge
- Qgd Gate to Drain “Miller” Charge

The swiching frequency is 13.9kHz according to the manual.

Thanks

 

[Nigel Goodwin]

Hello forum,

I'm replacing the MOSFETs in an Electronic Speed Control (ESC) for RC cars.

My problem is, I don't really know what to look for when choosing the new MOSFET. The current rating of my MOSFETs are all higher than the old ones in the ESC.
I'm more interested in the gate charge ratings, to allow the same fall and rise times.
What is most important to take into consideration?

As far as I'm aware (but I'm not a big user of FET's), one of the most important specs for ESC's is the Rds(on) resistance - this needs to be as low as possible, and is the main reason for using multiple paralleled FET's, hence providing the maximum voltage to the motor, while keeping losses (and therefore heat) in the FET's to a minimum.

 

[Someone Electro]

If you drive in an high freq. PWM signal the rise and fall times are going to be inportant as well.

The lower they are the better the MSOFEt is going to preform.Becose in that time the MOSFET is is half conducting mode.This makes a LOT of heat and so the less time its in this mode the less heat is made at high freq. switching.

But you must drive the gate whith high vurents to achive those times.Thats why they use up to 9 A peak to drive an moasfet wich is aculy voltage sensitive.This is becose of the gate capacitence wich needs to be low in oder for the mosfet to switch fast.

 

[Futterama]

I know some things about MOSFETs, including RDSon, this will be low enough, I'm more worried about the rise and fall times of the gate voltage.

This is becose of the gate capacitence wich needs to be low in oder for the mosfet to switch fast.

Yes, but which of the different types of capacitance is most important for this application? This is where I need help. Should I go for the lowest CISS (input capacitance) or the lowest Qg (total gate charge) or some of the other stated types of MOSFET charges?

 

[Nigel Goodwin]

I know some things about MOSFETs, including RDSon, this will be low enough, I'm more worried about the rise and fall times of the gate voltage.

This is becose of the gate capacitence wich needs to be low in oder for the mosfet to switch fast.

Yes, but which of the different types of capacitance is most important for this application? This is where I need help. Should I go for the lowest CISS (input capacitance) or the lowest Qg (total gate charge) or some of the other stated types of MOSFET charges?

The switching frequency isn't terribly fast, but how about comparing the specs of the original FET's?.

 

[Futterama]

The switching frequency isn't terribly fast, but how about comparing the specs of the original FET's?.

That's exactly what I am doing, but my question is, which property is most important to match the original MOSFETs (of the above stated capacitance properties)?

For instance, the original MOSFETs had a RDS(ON) = 0.010Ω (Typ), VGS = 10V where one of my MOSFETs have a RDS(ON) = 0.0019Ω (Typ), VGS = 10V. But the CISS of the original MOSFET was around 1000pf where the CISS of the low RDS(ON) MOSFET is 11000pf.

 

[Someone Electro]

But if the raise/fall times are too big the MOSFET could blow at an much lower curent than its rated for.

 

[Nigel Goodwin]

But if the raise/fall times are too big the MOSFET could blow at an much lower curent than its rated for.

Not really the 'current' but the 'dissipation', slow switching will increase the dissipation in the period between fully ON and fully OFF, the slower the switching, the more power will be wasted as heat!.

I would be somewhat wary of replacing it with one that's ELEVEN times worse!. Don't forget, the drive to the FET's has to charge and discharge this capacitance, so you require eleven times the drive power, and you have multiple FET's in parallel, so you need to multiply the 11,000pF by the number of FET's used as well.

 

[Someone Electro]

Yes is the the slow switching times make the dispation bad and that means that it will blow at an lower curent that it should.

 

[Futterama]

I would be somewhat wary of replacing it with one that's ELEVEN times worse!. Don't forget, the drive to the FET's has to charge and discharge this capacitance, so you require eleven times the drive power, and you have multiple FET's in parallel, so you need to multiply the 11,000pF by the number of FET's used as well.

Nigel, the MOSFET with 11 times worse CISS was just an example to make you understand what I was asking about.
And who ever said that I have MOSFETs in parallel? There are 4 MOSFETs, forming a H-bridge.

 

[Nigel Goodwin]

Nigel, the MOSFET with 11 times worse CISS was just an example to make you understand what I was asking about.

I understood before what you were asking about, and as I said before, choose something with BETTER specs, or at least not VERY much worse ones.

And who ever said that I have MOSFETs in parallel? There are 4 MOSFETs, forming a H-bridge.

You said 'ESC for RC cars', these pretty well universally use multiple FET's in parallel to reduce loses, and tend not to use an H-Bridge, instead they use a relay for reversing. This results in a higher performance car, as you have less loss in the speed control - generally you can get different 'grades' of the same controller, the only difference been the number of FET's in parallel.

For example, the 0.01 ohm FET you mentioned earlier, with two of those in an H-Bridge you have 0.02 ohms in total - if the motor draws 100 amps (which is quite possible with an RC racing motor!) you're losing 2V out of your original 7.2V, giving only 5.2V to the motor. Using ten FET's in parallel you get only 0.001 ohms, and a loss of only 100mV, giving 7.1V to the motor.

Here is an example , notice the specs are 65A and 140A maximum for the two models!.

 

[Futterama]

... and tend not to use an H-Bridge, instead they use a relay for reversing.

A relay? Really? Where can you get a 140A relay? I've been loking for those, and couldn't find any. It should be small too, to fit in an ESC. Also, wouldn't this wear out with time? Or is the reversing only done at 0% throttle?

The ESC I'm reparing, is rated at 50A, and there is no relay, it uses a H-bridge and no MOSFETs in parallel.

 

[Nigel Goodwin]

... and tend not to use an H-Bridge, instead they use a relay for reversing.

A relay? Really? Where can you get a 140A relay? I've been loking for those, and couldn't find any. It should be small too, to fit in an ESC. Also, wouldn't this wear out with time? Or is the reversing only done at 0% throttle?

Presumably the relays are used above their ratings?, and YES they will wear out in time - but the car has probably been crashed beyond repair before then :lol:

The ESC I'm reparing, is rated at 50A, and there is no relay, it uses a H-bridge and no MOSFETs in parallel.

Only a small one then?, not a racing controller?.

 

[Someone Electro]

At such high curents you realy need some good mosfets.

Juts try to find one that has an low gate capacitance and has the same or lower on state resistance.

 

[Someone Electro]

Damit beated by an few seconds. :lol:

 

[Futterama]

Only a small one then?, not a racing controller?.

Only a small one, came with a Mini Quake electric car (not mine btw):

https://www.duratrax.com/pdf/dtxm1260-manual.pdf

 

[Oznog]

There are some really great low rds-on devices if you look. In an RC controller you need small devices, and few of them.
https://www.st.com/stonline/products/literature/ds/6654/stv160nf03l.pdf

Lots of neat "HexFETs" in SO-8 pkgs have very low rds-on, though no thermal tab.

How significant gate charge is depends on the current drive capability of the controller as well as the frequency. Without specs on this, it's really hard to say what's too big.

Doing reverse with an H-bridge isn't so bad because the 2 transistors used for Reverse probably don't need to be as big. Also you would not PWM the PMOS side so gate charge is not important.