Connecting multiple MOSFETs for extreme current switch

gofastman

Enlightened
Joined
Oct 6, 2014
Messages
376
Location
New Hope, MN
I'm trying to setup a solid-state switching device for my coworkers truck winch.
The manual says it can draw a maximum of 335A at full load.
I was planning to use 8 TO-220 mosfets that are rated at 49A each. They will be connected to the A/C evaporator line, so heat shouldn't be an issue.

I'm trying to figure out the best way to connect all of these mosfets little pins to a 2 AWG cable :confused:
Perhaps solder them to a copper bar and bolt on the cable with a ring terminal?
 

gofastman

Enlightened
Joined
Oct 6, 2014
Messages
376
Location
New Hope, MN
have you considered something with screw terminals?

Here's a N channel mosfet in a power package, rated for 100V and 360A....
http://www.digikey.com/product-detail/en/IXFN360N10T/IXFN360N10T-ND/2116933

What are your plans for handling the inductive spike when turning the mosfet off? Maybe a snubber network, or a big TVS diode across the drain to source?

Ohhh, that could work! I have 20 of the TO-220 mosfets already that were for a different project, but maybe it's not worth the hassle.
The inductive kick was the next issue I was needing to address, but I wanted to see if anyone could help with mounting first.

EDIT: I see it says max power 830W. Is that supposed to be max power dissipation?
 
Last edited:

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
regarding the 830W... the datasheet says it can handle 830W when the case is kept at 25C. I interpret this as an indication of the package's thermal resistance, saying that the junction temperature will be at the max allowed value when dissipating 830W with a 25C case.

Of course, the odds of the case staying at 25C for any length of time are pretty small. It's probably a useful number when looking at how much power can be dissipated when turning the mosfet on and off. There will a brief time in between "on" and "off" when the mosfet is passing a lot of current and has a large voltage across it, and that's when it will dissipate a lot of power.
 

gofastman

Enlightened
Joined
Oct 6, 2014
Messages
376
Location
New Hope, MN
Ok, thanks.
Like I said, it will have some pretty significant cooling capacity.

And about the inductive kick, the issue I have is the motor needs to be able to swap polarity in order to real the cable in or out. Is a ~15V bidirectional TVS diode what I need?
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
How are you planning to be able to switch polarity to the motor? My first thought would be to use an H bridge, but that's going to require 4 mosfets.

Sizing the TVS diodes requires two bits of info.. what is the voltage rating of the mosfets, and how much current will be flowing when you try to stop the current in the motor? Well.... it really will be more complicated, because it's more a question of the energy that you have to direct around the mosfet, and that is related to how long the pulse will last. That is related to how much energy is stored in the motor winding.

Another factor is fast you are turning off the mosfet. A slower turn-off will generate a small voltage transient from the motor, but will mean that the mosfet spends more time in the linear region and will heat up more.
Everything in engineering is a trade-off, including this. :)

When selecting a TVS, you need to look at what voltage it starts to conduct and what the voltage across the TVS will be at its max current. There can be a large difference.

I would suggest trying out any TVS being considered by using a mechanical switch instead of the mosfet(s). If you end up blowing up the TVS, you won't also kill the pricey mosfet at the same time. Once you find a TVS that will survive, then start working on getting the mosfet(s) to work.

How is the rest of the design looking? Any need for an isolated mosfet driver? There are some neat photo-voltaic mosfet drivers available now.
 

gofastman

Enlightened
Joined
Oct 6, 2014
Messages
376
Location
New Hope, MN
If an H-bridge was used, could I use two of those mosfets that you linked in you previous post, and two of the smaller ones that I already have? I assume reeling the cable in requires FAR more power than letting it out.
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
If an H-bridge was used, could I use two of those mosfets that you linked in you previous post, and two of the smaller ones that I already have? I assume reeling the cable in requires FAR more power than letting it out.

That might work. It seems like a reasonable assumption.

It does require having some idea of the actual current involved, though, or else being willing to destroy some electronics in the process. Depending on the ability to make some of the crucial measurements, it is sometimes easiest to just over-design the device and not waste time.
If you do experiment with smaller mosfets, make sure you have a video camera running when you power it up. :)
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
.....

Pretty much indestructable.
Infineon-BTS555-DS-v01_00-en.pdf

but the datasheet does show that external protection diodes are required for large inductive loads.

Picking a good mosfet(s) is the relatively easy part of the design. Clamping or snubbing the voltage spikes when turning off the motor will be the hard part.

Depending on how the mosfet(s) is/are controlled, some isolation at the gate may be required. Is this just going to use a mechanical switch, or is there some form of electronics involved?

oh... if a H-bridge is used, some care will be needed to ensure that there is an "off" state in between the "forward" and "reverse" state. You don't want to ever turn on the high and low side mosfet on the same side of the H. ... but you probably already knew that...
 

gofastman

Enlightened
Joined
Oct 6, 2014
Messages
376
Location
New Hope, MN
That might work. It seems like a reasonable assumption.

It does require having some idea of the actual current involved, though, or else being willing to destroy some electronics in the process. Depending on the ability to make some of the crucial measurements, it is sometimes easiest to just over-design the device and not waste time.
If you do experiment with smaller mosfets, make sure you have a video camera running when you power it up. :)

I emailed their technical support about the power required for letting cable out under load, I'll see if they provide anything helpful.
I do have access to a oscilloscope, I'll see what kind of inductive kick I'm actually dealing with and report back.
 

gofastman

Enlightened
Joined
Oct 6, 2014
Messages
376
Location
New Hope, MN
I emailed their technical support about the power required for letting cable out under load, I'll see if they provide anything helpful.
I do have access to a oscilloscope, I'll see what kind of inductive kick I'm actually dealing with and report back.

I just heard from them:
Yes it drops but theres no chart because the load would vary in weight.

______________________________________

Sent: 2/6/2016 12:15 PM

To: [email protected]

Subject: 8.5e winch electrical question

I'm trying to find out how much current your 8.5e winch draws when reeling out cable under max load.The manual states a maximum electrical draw of 335A (let's say 350A)I would assume this is when pulling the cable in, would the electrical draw be significantly less when letting cable out at high mechanical load?
I guess I'll try to parallel a couple of the smaller mosfets and see what happens. I can't imagine it would exceed 100A reeling cable out :shrug:
 

gofastman

Enlightened
Joined
Oct 6, 2014
Messages
376
Location
New Hope, MN
I'm not sure if I measured it correctly, but I think the motor generated 350V of kickback under no load.
From this I'd imagine I'd be around several kV when switching off a high load.
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
I'm not sure if I measured it correctly, but I think the motor generated 350V of kickback under no load.
From this I'd imagine I'd be around several kV when switching off a high load.

be careful that you don't exceed the voltage rating of your scope probe!

I'm assuming that you are using a mechanical switch when testing this. The switch does arc briefly (usually) when opening, so that is absorbing some energy. I'm not sure how to use measured voltage to predict what size TVS you'll need.

If you can measure current through the motor, you might be able to estimate the inductance by measuring the rise time of the current when the voltage is applied. With a value for inductance, that would let you estimate the stored energy and then power dissipation in a TVS. You'd also have to estimate the series resistance in the motor & battery, etc. It might be enough to throw together a quick simulation, which would then let you play around with how fast you can turn the mosfets off, how much energy gets dissipated in the mosfet and TVS, etc.

in any case... lots of opportunity to experiment! Just be sure to wear safety glasses... flying molten copper beads and shards of plastic mosfet and TVS packaging can cause a lot of damage!
 

gofastman

Enlightened
Joined
Oct 6, 2014
Messages
376
Location
New Hope, MN
I used an inductive ignition coil clamp. Again, I have no idea if I set it up correctly or if the clamp and scope are even compatible with each other (the clamp is Fluke, the scope is Snap-on)
And Roger that on the safety glasses!
 
Top