Need help on n-fet pwm

[aq_mishu]

Guys,
I need to pass current to a battery from a PV. Now, I'm intended to use n-FET. the pic tells my connection, where the battery is missing...


now, I have connected the drain to PV and source to battery (infact an inductance first). Is that okey??? or I have to flip the drain to source??? But FET have to be at high side, and n-FET, no alternative.


I'll apply pwm at gate.



Mishu~

 

[naseerak]

You can switch the ground with the nfet but measuring battery will become impossible.

 

[FvM]

As you'll see in all NMOS switcher circuits, gate voltage must be applied between gate and source.

 

[NOT2MUCH]

You could use floating gate driver like IR2110.

 

[naseerak]

Another option is to use SEPIC topology with two Inductors and two FETS indeed both are N-channel.

 

[aq_mishu]

what all I wanted to know is should i swap the gate and drain?? gate will be fired from a clock source... but I'm a bit confused for Drain and source, as there is a bodydiode effect...

 

[FvM]

You know that the body diode conducts for Vds < -0.5 V. Assuming Vin > Vbat, you must not flip source and drain, just keep the NFET drain and source connection as is.

But the problem of applying correct Vgs voltage holds nevertheless. We can't see the applied gate voltage level in your simplified schematic, just guess you might ignore the requirements.

 

[aq_mishu]

okey... I'm gonna use IRF Z44N. The diode shows it will block from drain to source (as a blocking diode) and will conduct source to drain... [if I consider it as a general diode].

Now, I wish to tie the PV to Drain, (Pv is Vin) and Batt (Vbat)[aka load] to source... Gate will have pwm and thus it will work as a switch for a buck converter. Should I then flip?? according to the drawing... should not it be opposite??

BTW, gate pulse will be 5VDC from a uC for PWM

 

[FvM]

Should I then flip?? according to the drawing... should not it be opposite??

Why? According to your description, the circuit is O.K. as is.

 

[aq_mishu]

the gate pulse will be 5vdc from a uC

- - - Updated - - -

Why? According to your description, the circuit is O.K. as is.

Sir,
here is a link that is exactly what I want... it's on nFET right?? then the Vin is my PV and Vout is my load and Vin > Vout controlled by PWM. So no need to make any change at all... am I right??????


http://powerelectronics.com/site-files/powerelectronics.com/files/archive/powerelectronics.com/images/mosfet-fig1.jpg

 

[FvM]

You need gate drivers as in the link.

 

[Electro nS]

use P channel , if you want a high side switch , otherwise if you choose the N mos ( and that is offcourse a better choice) , you will need some kind of gate driver , charge pump , transformer based gate , opto isolated MOSdriver ....etc . because you need Vgs to be higher than 10v in normal mosfets and 5v in trench mosfets and in addition i doubt the micro can supply current directly to the fet .

you know what : my opinion use a DC-DC solid state relay ( crydom for example ) and save yourself the hassle .

hope that helps

 

[aq_mishu]

i do agree about p FET... but no way... i hv to stick with n-FET... now, i was planning use a inverter for synchronous.... thus one single pwm from uC will one give a HI for first n_FET and LO for other... and vice versa... now, as said... are you saying that 5V at gate will not work??? if then I have to move back to some other form like transistor based of better mosfet driver ofcourse...

now can you tell me which one should I take?? I found IRF2110 available here n my local market, but i'm a bit confused, can i use single pwm source to both input pins of the IRF2110 or I need 2 pwm, complement of each other???? the reason is one way another, I need to run 2 n-fet, alternatively (i mean complement of each other) and I have a single source of PWM... (one output from uC)... any ideas??????

Also, as I think it needs two PWM sig, complement of each other, I better use a inverter before it, as as said, I HAVE ONLY ONE PWM CHANNELS TO BE USED...

And rest I'm using the same way... though the datasheet of z44n says 5v is enough for upto 10A (I need 5A max). [though Vgs = 5-12 = -7] am i right?? I'm confused suddenly on this Vgs issue... :-(

 

[Electro nS]

as I think it needs two PWM sig, complement of each other, I better use a inverter before it, as as said, I HAVE ONLY ONE PWM CHANNELS TO BE USED... :-(

It will work one 1 PWM signal . just use an inverter to generate the HIN and LIN , because you will never want both transistors to be on the same time .

and regarding the VGs issue : yes you are right if you use 5v to a standard mosfet , it will not be fully ON (so Resistance of channel is still high) so this means more heat and less current carrying capability . but there is a new type of mosfets called Trench mosfets that have very low resistance on 4.5v Vgs . check : http://www.irf.com/whats-new/nr090428.html

but the problem with controlling a mosfet directly from microcontroller is current ! averge micros supply up to 20mA and that is very low . so even if you need to use a logic level mosfet and without a mosfet driver , you will need so kind of switch between micro and mosfet for example a totem pole. example for a 12v mosfet driver using totem pole , can be changed supply to 5v to suite your application with trench mosfets:

but you didnot say whats your opinion on solid state relay ? why not ?

hope that helps

 

[FvM]

The problem of gate drive for a NFET buck converter isn't only too low Vgs or providing gate current but requirement of a level shifter. Even a logic level FET won't work without a gate driver that raises the gate voltage above Vin.