Another Mosfet Question

[tazntex]

Good Morning to All,
I am driving a solenoid with a 12V coil, the coil is 34Ohms, off the top of my head thats about 1/3 of an Amp, I am also powering a 12VDC lamp that pulls about 2.5 Amps. I hear that I can use an IRF3205 Hexfet to be driven directly from my output on a 16F628A. The next question I have, on most datasheet for the FET's I've looked shows Vcc then FET then the load one is wanting to control (connected to drain) and then the Ground (connected to the Source).

Because of the solenoid and the lamp housing are ground, does it make a difference if I use the IRF3205 connected as: gate directly to output pin from 16F628A, Vcc 12VDC to Drain,
Source connected to load, and through the housing the load will get its ground.

Thanks

[mackrackit]

With an NPN mosfet, the Source is connected to "ground" 0 volts, and the Drain is connected to the load. The other side of the load (lamp) will be connected the positive rail.

The gate connects to the PIC pin with a pull down resistor. I normally use a 100K.

Check the gate saturation voltage if you plan to run the PIC at a lower voltage than 5.

If you are needing switching speed, you may want to look at IGBTs. Some mosfets are pretty slow if PWM is the goal.

[sayzer]

PIC gives you 5V, but your load is 12V; you can not ignite the gate.

To drive NPN mosfet with the load going to GND, consider the following way.


<img src="http://www.picbasic.co.uk/forum/attachment.php?attachmentid=2860&d=1221917652" >

---------------

[rmteo]

Another solution to drive an N-channel MSOFET with the load going to GND is to use a hi/lo gate driver such as IRS2001, LM5101, FAN7842 (all come in an 8-pin packages). They provide sufficient gate drive current and will allow you to PWM at pretty high frequencies - up to 1MHz depending on the application. I typically use them at 5-50KHz. Also, because they have hi and lo side drivers, they are very useful in bridge type situations using N-channel MOSFETs only.

[Acetronics2]

Hi,

Or output 50% PWM signal to drive the fet ... via a simple voltage doubler - or more - ( diodes + capacitors )


Hi, Sayzer

not so good for high current load ...

you'll drop at least 3 v in your Mosfet ( Vgs on + Vbe BC 337 + V diode ! ) ...

Bootstrap only works at "firing" the Mosfet ... not while steady conduction ...

Alain

[sayzer]

Hi Alain,

This way of "driving" is an issue that I always encounter with.

I have the following idea but not fully developed it yet.

Any comment is appreciated.


<img src="http://www.picbasic.co.uk/forum/attachment.php?attachmentid=2865&d=1222017290" >

[rmteo]

I would just use a dedicated gate driver such as FAN7371 (less than US$1).

The FAN7371 is a monolithic high-side gate drive IC, which can drive high-speed MOSFETs and IGBTs that operate up to +600V. It has a buffered output stage with all NMOS transistors designed for high pulse current driving capability and minimum cross-conduction.

* Floating Channel Designed for Bootstrap Operation to +600V
* 4A/4A Sourcing/Sinking Current Driving Capability
* Common-Mode dv/dt Noise Canceling Circuit
* 3.3V and 5V Input Logic Compatible
* Output In-phase with Input Signal
* Under-Voltage Lockout for VBS
* 25V Shunt Regulator on VDD and VBS
* 8-Lead Small Outline Package (SOP)

[Acetronics2]

Hi Alain,

This way of "driving" is an issue that I always encounter with.

I have the following idea but not fully developed it yet.

Any comment is appreciated.


>

Hi, Sayzer

Same problem ... the gate must be 5 + volts MORE than V+ for the mosfet to be fully ON.

and no existing turnaround ... except voltage multipliers ( dedicated drivers also have their own voltage multipliers inside !!! )

Alain

[rmteo]

Ace, is correct. A driver IC, like the ones that I mention above typically apply 12V above V+. So if your circuit is running from say 20V, the voltage applied to the gate of the MOSFET is 32V - this will ensure that is turned on fully.

[skimask]

I've used this 'STN2E10L' logic level N channel MOSFET before, kicking the gate with a PIC at 5v, switching a 24V load with little if any Rds(on), at least none my meter could measure.
A bit spendier than a standard N channel MOSFET, but it eliminates the whole driver issue, for me anyways...

[rmteo]

Logic level MOSFETs are definitely the way to go. Most recently released MOSFET's are logic level and are spec'd (for Rds ON) at both 4.5V and 10V gate drive voltages.

There 2 issues though with an N-Channel MOSFET. The load cannot be ground referenced and has to be on the high side - may or may not be a problem depending on the application.

Secondly, they are OK to be driven directly from the MCU if you are just turning them on (and/or OFF). For PWM type applications, they MCU cannot supply the current (and voltage) to efficiently overcome the gate charge/capacitance. With many applications running in the many KHz to MHz region (for reasons of efficiency), gate drivers become a necessity.

[skimask]

With many applications running in the many KHz to MHz region (for reasons of efficiency), gate drivers become a necessity.
True...I guess I should've quantified those statements a bit.
The bulk of those MOSFETs are driving a load of LEDs at 76Hz, a few are driving motors at 4882 Hz in other projects (yes, I know, the whine from the motors rocks!!! but I don't have to listen to it!). I've never measured I(g)...probably should. Might be the exact reason why I've had 3 pins fail to 'open-drain wanna-be' in the past couple of years.

[sheepdog]

Shame I did not read this thread before I toasted some irl2203ns, low rds on, logic level, I thought I was good to go. Powered up my battery vehicle drove down the street, heat sink perfectly cool I thought I had cracked it. Return journey mosfet stuck on until the brakes overcame its stall current, pop went the fuse. Only needed the wheels to fall off to finish the circus event.
It probably went through the Miller effect,shoot through etc,the other irl2003n for braking ending up in the same condition.
I suppose I live and learn, upped the voltage and used some 2n3055`s from a collection I have for some reason. Worked first time plus I have a bum warmer for the coming winter.

[Macgman2000]

The irl2203ns handles 82 ~ 110A @ 10v drive. Running a 12v supply I have used a PNP/NPN driver. Basically pnp emitter to 12v and collector to gate of FET. A base bias keeps the PNP in the off region and finally a npn pulls the pnp base bias low to turn on the FET. Works like a charm every time I have used this. Largest load I have used thus far is 75A and no problem @ 2kHz.

Nick

[skimask]

The irl2203ns handles 82 ~ 110A @ 10v drive. Running a 12v supply I have used a PNP/NPN driver. Basically pnp emitter to 12v and collector to gate of FET. A base bias keeps the PNP in the off region and finally a npn pulls the pnp base bias low to turn on the FET. Works like a charm every time I have used this. Largest load I have used thus far is 75A and no problem @ 2kHz.
So, this is almost like a double-stacked Darlington pair (with extras) and a mix of BJT and FETs?