Simple Power Mosfet - (use as a reliable switch)

Archive of useful information. This is not a discussion area.
Post Reply
User avatar
methods   10 GW

10 GW
Posts: 4385
Joined: Aug 08 2008 12:08pm
Location: Santa Cruz CA
Contact:

Simple Power Mosfet - (use as a reliable switch)

Post by methods » Jun 07 2018 12:10pm

Here is a place to start if you want to use N-Channel power mosfets for reliable switching.

Do not get caught up in the second order nuances of switching speed... that is controller design stuff. Here we are talking about using the power mosfet as a simple low side switch.

1) Failure Mode

The failure mode is reverse bias voltage (from Drain to Source) in excess of rating.
The failure results in a shorted mosfet, damaged mosfet, overheating, etc
The solution is to not allow the mosfet to be damaged (duh)

We are not concerned with a steady DC over-voltage. This is an unlikely case.
We are concerned with "noise spikes"which occur regularly in systems which have a lot of inductive switching going on

Start by reading this 1 page:
https://www.semtech.com/uploads/documen ... osfets.pdf


2) Solutions

In the link above you see recommendations for strapping the Drain to Source as well as the Gate to source with a Transient Voltage Suppressor.

"TVS Diode"

This is an ultra fast zener diode basically
They come with incredibly high impulse power handling (1.5kw, 3kw) and this happens in the picosecond time table

These TVS diodes do not have a sharp leg... so selection is important.
They are rated by "suggested use" or "nominal voltage"

Here is how I use them:

We want to turn 120V on and off
We spec a mosfet with sufficient margin - say 200V, maybe 250V
Our goal is to never let the mosfet see greater than 200V (or 250...)

Our job is to control up to 120VDC
We select a "120V TVS" - this will NEVER clamp below 120V - likely... at something like... 135V

This soft leg is the reason why we must spec a mosfet with voltage margin, to allow the TVS to clamp somewhere in the middle between what we are attempting to control and our upper voltage limit.

...

For the gate - since you have less control over voltage range - you basically have 2 options
1) Logic level control (typically less reliable for our use case)
2) Standard +/-20V (what I would use in a "12V system")

In this case we look at the three important TVS ratings

1) Minimum clamp
2) Rated duty
3) Maximum clamp

We select a gate TVS that has a maximum clamp below the maximum Gate to Source voltage
1.5kw will more than suffice here as we expect to see an order or two magnitude less noise on the gate than the drain*

Consider that your 12V system will likely not be lead...
Maybe 4S life or even 4S lipo
A 4S lipo can run upward of 16.8V so that leaves a very tight window... to not bleed off DC voltage... but to clamp in time

Solution is to drop 4S lipo as a 12V replacement option
4S in a 3.3V chemistry is MUCH BETTER SUITED for many reasons - so go that route

That being the case - maybe 15 or 16V is your max (call it 16) and you now have 4V to play with to clamp the gate voltage

If you are interested in learning more about mosfet failure modes - check these two articles out

https://www.vishay.com/docs/90160/an1005.pdf

[url]https://www.infineon.com/dgdl/Infi ... 0210977cde[/url]

"Avalanche" is the term used to describe how a mosfet behaves when overcome by voltage. This is the achilles heel of solid state switching.

The take home is this:

* There is impulse and AC noise
* This noise is significantly greater than your maximum DC Voltage
* This noise needs to be bypassed or shunted off
* Shunting means we are going to crowbar with a TVS didode
* Advanced circuits bypass additional noise thru use of low pass filters, bypass caps, and other means to "let the voltage get by us"

Thats my "for dummies" explanation of using Power Mosfets as low side switches.

Next Episode
A) The single switch - directional - low side only
B) The bi-directional high side switch made up of N-Channel mosfets

B can be accomplished with only 3 unique parts (4 total)
2 N-Channel Power fets tied together at the Sources
1 DC-DC rated for Gate Voltage
1 Pull Down gate resistor

The naive solution uses about 1W to control many KW
There are any number of ways to improve efficiency by orders of magnitude
A power mosfet requites (almost) no power to control - therefore - the only efficiency you should worry about is Isoaltion

Isolated DC-DC's allow bi-directional switches (SSR's effectively) to be completely agnostic to direction or switching voltage.
Efficiency can be improved by fixing the direction and enforcing minimum voltages - and the range of voltage - such that this voltage can be utilized to control the gates.

There are any number of complex solutions out there (chips) but what we like to focus on here is SIMPLE STUFF... that you can understand...

...

We want you to have a parts bin on your desk
In this parts bin will be ... 20... maybe 30 key pieces that you can build anything with.

1) N-Channel Power Mosfet - surface mount or thru hole - TO-220 or Dpak - low internal resistance - maybe 100V and 200V
2) Standard thru hole and 0805 resistors - 1K, 5K, 10K... that sort of range - for use biasing
3) Small, simple, puck, isolated DC-DC converters. 1W variety.
4) TVS Diodes. 1.5KV thru hole. In 5V, 12V, 100V, 150V, etc... (your use range - buy an array and keep them around)
5) Small signal latching relay - 5V or 12V coil - VERY USEFUL for low power designs
6) ...

Sort of like that.
Or... thats what I have... and I can knock together just about anything in a pinch.

Happy sailing
-methods
Increasing battery voltage and controller current limit will result in a non linear experience

Post Reply