Electronic fuse / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Electronic fuse

Move Thread

LAN_403

Dr. Dark Current

Sat Nov 22 2008, 09:38AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Hi all, I'm in need for an electronic fuse for my mains-powered projects. It should be adjustable from 0 to xx amps at 350 volts DC, very fast and latch in the case of failure.
Sorry if this seems a bit like asking for schematic, but I did my "homework", I searched the net and spend a few hours thinking about/designing the schematic, and haven't come up with anything.

Any ideas? (I don't need schematics, just "how it could be done")
Thanks

Mathias

Sat Nov 22 2008, 10:32AM

Registered Member #1381 Joined: Fri Mar 07 2008, 05:24PM
Location: Hungary
Posts: 74

You need a "Current Switch" , basicaly it's a CT with some logic that controlls a relay

Marko

Sat Nov 22 2008, 10:34AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Hello Jan,

Simplest solution I can think off would be to use a small signal SCR, like ones that can be found in PC power supplies, to pull down a power mosfet which interrupts the negative rail. The SCR is activated by a comparator and a current sensing shunt.
Normally I would prefer to use a current transformer whereever feasible, but I know in many cases it isn't so.

Alternatively, you could use a SR latch instead of SCR (with proper amplification for the mosfet drive).

Now, what does your xx amps mean - if it means you need the circuit to operate up to 99 amperes, shunt is not going to satisfy the requirements anymore and you'll have to use a hall current transducer. (I hope you don't need that much current still?)

The problem with shunt is that it needs to dissipate more power at higher currents for a constant voltage drop, and this voltage always needs to be like 10x your comparator's maximum offset voltage.
So you end either with really large and lossy shunt or having to use special precision op-amps.

Marko

Dr. Dark Current

Sat Nov 22 2008, 12:13PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Hi Marko,
I was thinking of a MOSFET because it is fast, and also because I had the idea of measuring voltage drop directly across the FET, eliminating a shunt resistor. Surely the trip current would be a bit temperature dependent but I don't need exact setting. Just so it trips when fault (=short) currents start to flow and doesnt blow itself up

By xx amps I meant a peak (trip) current of max. 30-40A, with average current a lot lower (~10A).
But I haven't succeeded in building the latch mechanism, any one I design requires to reset the fuse first to start conducting. I would like it to turn on as soon as the main supply comes on.

uzzors2k

Sat Nov 22 2008, 01:06PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

My MOT bench supply does pretty much all of that. It's not very exact, relying on the forward base voltage of a signal transistor, but it works. The main mechanism is a simple SR-latch made with discretes.

Dr. Dark Current

Sat Nov 22 2008, 01:28PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Ideally I would like the delay of the fuse to be in the tens of ns range, so it can withstand a direct short on the output (ie. 2 wires touching together).
Is this even possible?

uzzors2k

Sat Nov 22 2008, 02:14PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

Both my MOT and SMPS can withstand that. I haven't measured the response times of over-current circuits, it's shorter than the pulse rating of most devices. (Rectifiers and IRFP450s at least) In fact I had to put some delay in the over-current circuits to account for charging of filter capacitors on the output. (Just putting a 470ÂµF cap on my SMPS output would trip it otherwise.) Even what appears to be a direct short will still have some resistance in contact leads and internal resistance of the filter capacitors, etc, and on top of that every device has a peak current rating over a few ms. So even slowish over-current protection circuits will work fine. It's not really current that kills, but how long that current lasts. (Consider DRSSTCs and coil guns for instance.)

Dr. Dark Current

Sat Nov 22 2008, 02:30PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Well, I was going to use a 0.1ohm MOSFET as the fuse switch, and of course there will be decoupling caps with very low ESR. Lets say the total resistance is 0.2 ohms, which would result in about 2 kiloamps of peak current. I'm not really sure if the FET can stand this for <100ns, but I'll probably have to find out

Steve Conner

Sat Nov 22 2008, 02:49PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

To ease the switching time requirement, use an inductor in series with the MOSFET, and the appropriate clamp diode to discharge it. Or use a short-circuit rated IGBT, and go easy on the gate drive voltage.

I made one circuit like this:

but it kept failing in the field. I redesigned it with a LM339 and TC4422, but never got round to documenting it. The switching time was very easy to meet, because my OLTC2 had a big inductor in the DC bus.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob