Society of Robots - Robot Forum

Electronics => Electronics => Topic started by: Admin on July 24, 2007, 02:05:53 PM

Title: mosfet . . . source on ground is better?
Post by: Admin on July 24, 2007, 02:05:53 PM
So Im hooking up a MOSFET and my coworker insists Im crazy . . .

Anyway, I have it hooked up this way:

6V - mosfet - motors - ground

But my coworker cant figure out why it even works this way . . . he says it should be:

6V - motors - mosfet - ground

The gate on the MOSFET is attached to a I/O pin on my uC.

So . . . the question is which way is better and why? Im sure it has something to do with the internal physics of the MOSFET which I never fully understood . . .
Title: Re: mosfet . . . source on ground is better?
Post by: sturnfie on July 24, 2007, 04:57:27 PM
Well, the point of a MOSFET is to allow you to use an applied voltage to control a current.

I can see benefits of hooking co-workers way (Power -> component -> MOSFET -> Ground) simply due to having a direct sink on the other side of the MOSFET (since this is really what you are trying to control, the closed path to ground). 

If you do it your way (Power -> MOSFET -> component -> ground), it acts on the same principle (applied voltage allows the current to transverse the MOSFET to a sink), you are "grounding" into a component.  If the component "breaks" and that arm develops an impedance higher than the Gate connection (where you apply your voltage to control it), you *will* fry the MOSFET.

Title: Re: mosfet . . . source on ground is better?
Post by: paulstreats on July 24, 2007, 05:14:34 PM
It could be due to electron / conventional charge flow.

Electrons travel from the negative to the positive. (electrons are negative and are attracted to their opposite).

Conventional current flows from the positive to the negative.

When the gate of the mosfet is open, the current flows from the drain to the source.

Conversley, the electrons actually travel from the source to the gate.

By putting the mosfet last in the chain, i.e. closer to the negative, the mosfet is the first in line for electron flow(volts) from the source to the drain (making the mosfet always available for current flow). But it is last in line for the charge flow, and therefore current would be used up more from the actuators etc.. which are first in line from the flow of the current, so if theres less current available to go through the mosfet, it wont get so hot so quickly.

see below:-

              >>>>>>>>>>Volts (electrons)>>>>>>>>>>>>>>

       Ground -> Mosfet -> Component -> Power
             <<<<<<<<<<conventional charge<<<<<<<<<<<<<

Also, If you use any voltage reducing components between the negative(ground) and the mosfet, it immediately reduces flow through the mosfet, and if this is reduced to an extent, the mosfet will not allow the flow of charge.

Title: Re: mosfet . . . source on ground is better?
Post by: Admin on July 25, 2007, 07:24:50 AM
On further thought, this all depends on if its a PNP or NPN MOSFET . . .

My way is better for PNP and my coworkers way is better for an NPN. In NPN, the source must be connected to ground to keep a voltage difference between the gate and ground. With PNP, the source must be connected to power so that when the gate goes to low it turns on with the voltage difference.

I think I understand it now . . .
Title: Re: mosfet . . . source on ground is better?
Post by: zamboniman60 on July 27, 2007, 11:41:53 PM
It really doesn't matter which way you connect it. It depends if you want to make the MOSFET a current source or a current sink. Just do what you need to do -- the only place where it matters much is in switching regulators (in a buck, you put the MOSFET before the inductor; in a boost, you put the MOSFET after the inductor, momentarily shorting it to ground).
Title: Re: mosfet . . . source on ground is better?
Post by: zamboniman60 on August 15, 2007, 02:13:32 AM
After wasting too much of my time learning about MOSFET device physics, I think that I've figured out why your setup was working. When the MOSFET is off, it drops the voltage in the circuit "before" the gate -- this means that applying 5v from your uC acts as 5v between the gate and the source, turning the MOSFET on. Once on, the voltage dropped in the MOSFET disappears, so your 5v would be below the required drive current, right? Actually, the Gate-source capacitance keeps the charge there, and the uC won't allow the charge back into its pin, so when the MOSFET "floats" up to 6v,  because of the charge in its capacitance, the Gate floats up to 11v! Then, when you set the pin on your uC low, the MOSFET sees that as -6v Gate-Source, turning it off almost ideally.

You've inadvertently created a psuedo-bipolar power supply for your MOSFET, Admin!


P.S. Let me know if you find this description reasonable -- feedback would be appreciated.
Title: Re: mosfet . . . source on ground is better?
Post by: Soeren on August 15, 2007, 11:27:57 PM

What kind of nonsense is that?

N-ch devices should go to ground, to get the gate-source potential "in order".
P-ch devices should go to V+ to get the source-gate potential likewise.

V+ and ground counts equally as low impedance rails, and N or P devices should be chosen from where it goes, to always have a nice gate-source potential.

The above assuming switching - for analog work any thinkable combination/position could find a use.
Title: Re: mosfet . . . source on ground is better?
Post by: zamboniman60 on August 19, 2007, 02:55:12 AM
Of course, it would certainly be WISER to set the system up as you described, Soeren. I was merely trying to figure out why the system worked at all when set up (+V) - (N-channel MOSFET) - (load) - (Ground).

I'm not going to bother trying to explain the physics because this sort of thing goes very badly into words... I would recommend that you look up some MOSFET device physics explanations; the reason Admin's setup worked has to do with the electric field caused by high-density current flow.