Fuses: A circuit to share and a problem to discuss.

[Zanderist]

This circuit will tell you if your fuse is blown or not.

Here's a video of it in action



I talk about the problem in the video but I will state again the MOSFET gets destroyed when short circuit amps is applied. The issue with that is that it's a 100 Amp MOSFET. What I suspect  is that it needs a heat sink to perform properly. I think all the heat in the MOSFET breaks down the glass insulator inside it and fuses the gate pin the substrate.

Datasheet of the mosfet
http://www.mouser.com/Search/ProductDetail.aspx?R=PSMN2R7-30PL%2c127virtualkey66800000virtualkey771-PSMN2R730PL127

[CraigHB]

I fried a MOSFET once by applying reverse polarity only for a fraction of a second.  I accidentally touched the battery backwards for a split second and that was the end of it.  It was a 70A MOSFET.  Problem is with the batteries used in high power mods, they can deliver several hundred Amps on a short circuit.  Two VTC5s in parallel will deliver 600 plus Amps instantaneously on a short. 

PTC fuses have a trip time that varies depending on the fuse rating.  Typically makers post trip time charts in their data sheets.  You notice it does take some time for the fuse to shut down the circuit.  That being the case, the MOSFET has to endure the short circuit current until the fuse trips. 

Another issue with MOSFETs is they can generate huge amounts of heat when operated in the "active" region.  This is the region in between fully off and fully on.  So for example, if the MOSFET is dropping one volt because it's not turned on all the way and there is a flow of 20A, that's 20W on the MOSFET which is well past the typical 5W tolerance of the package.  If you are using a MOSFET with a high gate source threshold and the fuse drops voltage slowly, then the MOSFET is operating in the active region with short circuit current.  That's likely why it's burning out on you.

You didn't post the specs on the fuse, but you will probably have to select an array of smaller PTC fuses with a fast enough trip time to protect the circuit.  Additionally, you may also need to use a MOSFET with a lower gate-source threshold that ensures it remains fully on until the fuse trips.  If worse comes to worse, you can use a fast-blow single use fuse that will trip very quickly.  Has to be the fast-blow type, not the slow-blow type in the glass capsule.

[Zanderist]

I looked up the spec for the fuse and it states 20 seconds to trip at a current of 40amps.
http://www.mouser.com/ds/2/54/mfr-72330.pdf

My question is now, could a heat sink correct the MOSFET burning out problem?

Is 5W a standard value for TO-220 packages?


I am also thinking about just using fast blow fuses like the blade type in a car.
http://www.mouser.com/Circuit-Protection/Fuses/Automotive-Fuses/_/N-ba8b5?P=1z0z33z

[Visus]

My thoughts on a cool circuit..
If you parallel the mosfet it will increase it's current handling at 4.2v. 
Have you spec's the fet at 4.2v amp limitation?
Like Craig says 95% fets,  they are not fully on at 4.2v and all have a mushed  lower amp rating @ 4.2v.
What fet are you using?

A heatsink wont stop it from blowing if its amp limitation is overdrawn in a short maybe for run/on resistance ya..
Your hitting it with over 100A with that sony and unless you dial in the short I do not think that circuit will work because of the power of the sony cell without a relay..  But if ya parallel some fets it may/can..

[CraigHB]

I looked up the spec for the fuse and it states 20 seconds to trip at a current of 40amps.  http://www.mouser.com/ds/2/54/mfr-72330.pdf

Look at the chart and determine the trip time at the actual fault current.  With the resistance in the circuit and the resistance of the battery, it's probably 100 to 200 Amps.  Ideally it should be under 1 second, 2 seconds might be okay.

My question is now, could a heat sink correct the MOSFET burning out problem?

No.  The problem is those fault currents are overheating the channel faster than heat dissipates through the package.  The root of your problem is you're getting big voltage drops during a short circuit condition that lower gate-source voltage.  This causes the MOSFET to drop into the active region creating large amounts of heat in the channel.  Beef up wiring so fault currents do not cause as much loss in gate-source voltage or use a MOSFET with a lower gate-source threshold to ensure it's fully on during a fault condition.

Is 5W a standard value for TO-220 packages?

5W is typical for high power packages without the aid of a heat sink, but it's higher if the part is mounted to a heat sink.  It's usually specified in a part's data sheet.

I am also thinking about just using fast blow fuses like the blade type in a car.  http://www.mouser.com/Circuit-Protection/Fuses/Automotive-Fuses/_/N-ba8b5?P=1z0z33z

Those are slow-blow fuses and have trip times no shorter than a PTC fuse.  The fast blow type are round in a white ceramic package.  They are specified as fast-blow fuses.  You may have better luck with a slow-blow fuse if you use a lower rating with a faster trip time.  They post trip time charts for single use fuses in data sheets as well.  However, the problem you run into is a lower rated fuse may blow under normal use.

[Zanderist]

Things are starting to make sense now in terms of mosfet box fuse selection.

I will post something sometime later explaining what I have learned.

Just a quick question to confirm if this is correct,

In a textbook of mine it has a formula of power delivered over a period of time.

It is W=VI(t)

Assuming 5W of power disappation at the mosfet my trip time should be if a short circuit of 50 amps at a perfect voltage drop of 4.2 volts

5W=4.2V(50A)t
t=0.023 seconds

So I would then choose a fuse that can trip at 50 amps with a trip time close to .023 in order to protect the mosfet.

[CraigHB]

Keep in mind those are continuous ratings.  It all comes down to heating which is a matter of energy.  Watts being Joules per second means you'll get the same energy at 10W for one second as 20W for half second.  If a MOSFET can handle say 10W for 2 seconds than it can handle 20W for one second.  So that being the case, power tolerance increases as the duration of the fault decreases.

It's going to be hard to nail down an exact number in terms of power tolerance for a given time frame.  If you can keep power dissipation within a factor of 10 of the continuous rating for one second you should be good.  In other words, if you can keep power dissipation under 50W for 1 second or 25W for 2 seconds it should be fine.  That sounds like a big number, but if you're looking at a fault current of 100A, it's actually pretty small.  Consider that if the on-state resistance is 5 mOhms then power dissipation is .005 times 100A squared which is 50W.

[Zanderist]

I've switched to a lower rated fuse and it works much faster now.

I will make a video though how I concluded the fuse choices for comments, because I don't think it was the best way.