Current Limiting and Electronic Fuses

[ludzinc]

Hi,

One thing I want to include in my designs is a current limiting circuit for my on board power supplies.   This is because I want to string a bunch of modules together (think home automation) and I don't want a short circuit starting a fire.

In the past I've used polyswitches for this job but I find them lacking - time to tip can be measured in seconds for a 200% over load for example.

Anyway, I've attached an image of the circuit I'm playing with.

The MOSFET I'm using as a pass element is an NTD2955PT4G (datasheet here: http://www.mouser.com/ds/2/308/NTD2955-D-90147.pdf)

Rated at 12A and 60V, I guessed that it would be fine at 500mA from a 15V rail.  (Guessed, yes.  No maths was harmed in this experiment.  This was probably my first mistake).

I built my circuit, and hooked it up to my copy of Dave's constant current load.  This I limited to 800mA, and confirmed by running the power supply under test (through my ammeter) into the load.  Adding my current limit circuit in series, the current indeed was limited to around 500mA, and I thought I was successful.

Until I noticed that the solder around my FET was re-flowing nicely.  Very, very hot.

I was assuming that when current limiting, the BC807 would turn the FET off, but that's a bad assumption.  It *limits* the FET - I measured the gate voltage to be around 9V, not near 15V and thus the FET (I assume) is in a linear mode of operation, and is dissipating around 7.5W (15V * 500mA) of heat.

1.  Would you agree that this assumption is true?
2.  Could some one suggest an improvement so that my circuit indeed switches the FET off, rather than limit current through it?In an ideal world this switching action would turn the ET back on when the fault condition / current over load is removed.

Not that if I pull the gate up to 15V the FET does turn off.

I could 'solve' this by using a large heat sink, but that adds both cost and size to my project. 

At least this experiment has allowed me to define my requirements for this circuit - I don't need a current limiting circuit, rather I need a resetting electronic fuse.  I think...

Thanks for any comments.

[ovnr]

Yeah, you're going to need a bit more stuff than that if you want it to work well.

If it did cut off completely when your 500mA limit was reached, it'd turn right on again afterwards, because the PNP is no longer driving the MOSFET gate high. You need it to latch off if you want it to work as a fuse and not a limiter.

I've been working on an electronic fuse design myself (but haven't had the time or :effort: to complete it) using the LT6109, which is ideally suited to this kind of application. You should take a look at it - if you just want a simple "fuse", it'll do what you want with little extra fuss. IIRC, the -2 model latches - read the PDF.

[dannyf]

rather I need a resetting electronic fuse.

Lots of reset-able fuse. Tons of electronic fuses - the simplest would be the craw-bar type - it is however self-destructive.

I would imagine that it wouldn't be that difficult to design a reset-able electronic fuse, with a 555 or a mcu - never tried that myself.

[minime72706]

rather I need a resetting electronic fuse.

Lots of reset-able fuse. Tons of electronic fuses - the simplest would be the craw-bar type - it is however self-destructive.

I would imagine that it wouldn't be that difficult to design a reset-able electronic fuse, with a 555 or a mcu - never tried that myself.

It's only self-destructive if the designer failed to choose an appropriate SECONDARY FUSE, I imagine. No matter what you need a real fusing device or you're still asking for your !@#$ to blow up.

[wasyoungonce]

Have you thought of using an IRF programmable current shutdown FET?  Like IRF3315.  They are reasonably new (not really) but ideal for this application, just need one resistor to fold current back.

Oh and power dissipated by the device will be the voltage drop across the device (basically Rds on times I passed) times the current passed.