Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

High side or low side?

Status
Not open for further replies.
Hi there,

When driving af MOSFET gate with a driver IC like the TC426CPA through a 4k7 resistor, how important is the high side/low side thing anyway? The high side MOSFET will open with almost the same amount of current as the low side MOSFET, so why bother buying a high sided or a low sided driver IC?

By the way isn't it possible to just flip a N-channel MOSFET vertically and power it on the source pin and it'll work like a P-channel, so you only need to buy one type?
 
Last edited:
A 4.7k resistor value is way too high to be in series with the very high capacitance of the gate of a Mosfet. A 10 ohm resistor should be used to keep the Mosfet from oscillating and can allow the driver IC to supply 1.5A to quickly charge the high gate capacitance. The current is used to quickly charge and discharge the gate capacitance. The DC gate current is zero.

Mosfets don't work when you connect them upside down. If the circuit needs to have a P-channel Mosfet then use one.
 
So I'll need a 1.5A*18V (maximum) = 27W resistor and thats a lot. I can't even get one under 30$. Is my calculation wrong?

Compromising: (18V/47R)*18V=6,89W - I can get that
 
Electronics4you said:
So I'll need a 1.5A*18V (maximum) = 27W resistor and thats a lot. I can't even get one under 30$. Is my calculation wrong?

Compromising: (18V/47R)*18V=6,89W - I can get that

yes completly wrong!
echoing Nigel.. where was 18V derived from

but then it comes back to the 1.5A requirement
MOSFET's are Voltage-controlled devices!, but they do require charging current (I lost a job interview because some ass-hat would not accept you need to pump current into the gate of a FET :rolleyes:)

Anyway At turn-on/off you might get 1.5A at the start (the actual inductance and resistance will determine the actual current... my gate-drives try to push 12A up an IGBT gate, but due to the inductance in the gate-leads they rarely get over 6A)

But as the gate-region charges the current flowing into the gate (due to the potential difference between the gate and the drive decreases) decreases as well - becauase effectively a capacitor is getting charged.

ONCE the FET is hard-ON/OFF no significant current will flow to/from the gate and thus the gate-resistor will disipate no power (it is only in switching events that it will)


Now if you just want to turn the FET on and then leave it on a 1/4W or 1/8W resistor should do. BUT if you want to PWM the FET you have to work with RMS current

The actual charging curve of a FET's gate capacitance isn't exactly the exponent of a normal capacitor (there is a nice millar effect at a certain voltage when the gate capacitance is effectily infinite in size) but a normal cap curve will survice

take that waveform and for yr given switching freq work out the RMS current and then go I^2R to work out the power that will be dissipated in the gate-resistor


I have a very nice Excel spreadsheet that does all this but it is on my work network (the python script that *should* do this is also a bit buggy)
 
Try the circuit with 10 ohm/1/4W resistors in series from the driver to the gates of the Mosfets. If the resistors get too hot then try 1/2W resistors instead.
 
Styx said:
I lost a job interview because some ass-hat would not accept you need to pump current into the gate of a FET
Probably a good thing, I certainly wouldn't want to work with morons. I probably would've called him a moron, sworn at him and stormed out. I don't have time for people like that.

You don't actually need a resistor, it's only used to prevent high frequency oscillation, a small ferrite bead will do the job.
 
Hero999 said:
Probably a good thing, I certainly wouldn't want to work with morons. I probably would've called him a moron, sworn at him and stormed out. I don't have time for people like that.
You have no idea! I was fuming when I left the interview (was looking for a graduate job). Went to the interview got asked some tech questions and then they asked me abt my final year project (a 1KW switch-reluctance converter with a current controller) we went through the PID section (with coaxial-shunt for current sensing) all around the analogue ccts with a few questions abt bandwidth calc and the need for resistors on the +VE terminal of an inverting amp.
Got the the gate-drive (a simple opto-driver with a 1W Newport DC:DC converter) and I said the opto can source 1A to the gate of an IGBT... (explained an IGBT is like a FET to the drive and a BJT to the power side) and they went "FET are voltage devices, you don't need amps to control them... that is BJT's

Fuming on the way home!


Hero999 said:
You don't actually need a resistor, it's only used to prevent high frequency oscillation, a small ferrite bead will do the job.

Weeeelll.... a gate resistor on a FET or an IGBT (from the drive side an IGBT looks like a FET) is needed for one real reason BUT it help in a number of places (all inter-connected)

1) the resistor stop's a pierce oscilator being setup
2) limits the current flowing into/out of the Gate region (so you don't burn it out, IGBT's can handle alot higher gate-current tho)
3) to control the turn-ON/OFF speed's. Too fast and the overshoot at Toff will increase losses (or even overvolt the silicon... took out a 1200V IGBT from a 270V link due to this) and too fast at turn-ON can cause the reverse-recovery of the diode to be higher then it usually is.

likewise too slow and you have higher switching losses (since hovering around the active region too long)

For an 8W (H-bridge output stage) Gate-drive (to control a 1000A, 1200V IGBT) I tend to use 47R for the gate resistors to the P-types and 100R to the N-type in 0805 package (with 6Watts worth of leaded resistors of 2R3 for the gate resistor to the IGBT)
 
Last edited:
The 18V is the maximum voltage supply to the TC426 MOSFET driver.

I'm going to do PWM to the gate through the TC426 (logic on the one side, high voltage/high current on the other side to drive MOSFET).

Styx do you have a sample on how to find the RMS current (haven't worked with that before)? The effect of the resistor must be lower because of the pulsing RMS current in stead of a steady state current...

What is the maximum gate current for a IRFZ34N, when it's not mentioned in the datasheet?

And again - what's the different of a high side and a low side MOSFET driver, when the current needed to turn on the HS and the LS MOSFET is likely the same?
 
Depends on the type of FET. Mosfet's require gate drive relative to their source voltage so P channel fets are usually used on the high side to prevent having to create a higher voltage that VCC to drive it. A Hiside N channel FET driver may have a built in charge pump to provide the higher voltage source to turn on an N channel fet that's on the high side, N channel fets are prefered because given the exacty same physical size an N channel fet has a higher current capability than a P channel because of the way they're doped.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top