Current Level Detection

Don't get catty with us.

That comparator is only good for 5 volts, but that's actually good. A resistor/zener thing will let one hang it on the 12v rail, and provide transient protection. And a regulated threshold.

With 25 uV offset, the shunt can be pretty small, like a fat PCB trace or a few 1206 resistors in parallel.

A less exotic amp or comparator might be OK with a trimpot to set the trip point.

They get much less warm at 100x operating current?!

Yes, it does. Anyway, you're one of the Top Blokes here nowadays, what in your opinion is the best approach to use?

It doesn't have to work over that current range. Cursitor Doom just wants to detect when the current falls below 200mA.

A seventy-five time higher magnetic field isn't going to damage a Hall effect sensor. It will probably drive it into saturation, but that isn't going to be a problem.

The magnetic field at the centre of single turn current loop is

B=μ0.I/2R(at center of loop) Telsa

where u0 is 4.pi.0^-7 newton per ampere^-2 ( Henry/metre).

For a 1cm radius single turn loop carrying 200mA that's about 12.6uT, or about 0.126 Gauss (which isn't much).

The A1324 has a 10 Gauss worst case shift across it's temperature range, so you'd need more turns and a soft-ferromagnet core to get it up into the range where it would be easy to get a useful output.

The chance that a one-off engineered by Cursitor Doom is ever going to work isn't high.

Off-the-shelf relay coils aren't going to take 15 amps, that'd be silly. You can make one, though, and adjust it with permanent magnets to trip at any current. There will be some hysteresis.

Stable? An air-core coil and Hall sensor aren't likely to suffer mechanical failure, or oscillate.

I'm a fan of NVE's GMR sensors.

has various application notes including a 0-10A current sensor using a PCB trace. A coil of say

10 turns could make that sensitive enough with the right geometry. They do work at high temperatures, too.

hystesis shouldn't be a problem as by my reading of the original post it starts at 15A for a few seconds which will set the reed then reduces to long-term 240mA, it just needs to release at around 200mA

They do look interesting. This part is about $5 per unit, and Digikey has 57 in stock.

A saturating inductor can be a super-sensitive current sensor that will tolerate a lot of current. But the circuit would be moderately complex.

Hey: build basically a 2-transistor ferrite-core dc/dc converter. Drive a loudspeaker. It will natively scream. But DC current through or near the core will stop it.

Not very practical, but fun.

There are some DC capable devices about as clamp on meters although ones intended for AC are much more common. The sensor intended for an Owl mains power monitoring system might be a suitable choice off the shelf.

You only want the clip on sensor and DIY electronics to monitor it. eg.

Afterwards you can use it to get realtime domestic power usage.

Possibly if they arc weld closed forever the conductivity will be better.

I though Jason's scheme had the current flowing in a fat solenoid wrapped around the reed and only the piezo buzzer current through it.

Extremely practical, and remarkably accurate (if done right). Back when I was working at George Kent in Luton (1973-76) our senior engineer - Colin Hunter - had side projects. One of them was for measuring the thousand-odd amp currents on the bus-bars on the London Underground.

If I remember rightly, his current sensor was a pair of C-cores that clamped onto the bus-bar. He'd wound a centre-tapped coil for a two-transistor Royer inverter onto one of the C-cores, and measured the difference between the two saturation currents. His claim was that it was remarkably precise measure of the current in the bus-bar. He had a habit of being right about that kind of stuff - he'd effectively been trained as an apprentice to Peter Baxandall at the UK Royal Radar Establishment, and he'd emerged from the process as a remarkably ingenious engineer.

Look to ebay for cheap breakout kits:

I think I'd try to put it on the return line, if at all possible. The CM range goes down to 200mV below Vss.

A 1 Ohm resistor in parallel with a big schottky diode and a comparator for 200 mV.

Bye

Huh? There's no reason that a winding around a relay has to have such resistance that it will take a 12V constant bias (though there IS a market for such). Just use big enough conductors. The reed only sees a bit of B field, not a current, so it has no problem... as long as the alloy doesn't permanently magnetize.

Reed relays that don't magnetize exist, and 200 mA will pass fine through wiring that tolerates 15A, so... I don't see any problem in the physics.

You need to wind a lot of turns of wire around the reed to make the threshold 200 mA.

Getting a lot of turns around the reed needs thin wire, but standing

15 amps needs fat wire, #14 maybe. So the problem is to wind some hundreds of turns of #14 around a tiny reed.

Reeds are awful anyhow. Use electronics.

My estimate is about 40 turns if you can get hold of a Comus RI-03AAA reed capsule (which isn't anything exceptional).

That's a bit generous

For chassis wiring 17 AWG - about 1mm OD would do. 14 gauge is closer to 2mm OD.

The Comus reed capsule is 2.54 mm and about 14.5 mm long. Forty turns of 1mm OD wire could be a three layer winding 14 turns long, which doesn't sound too bad.

They aren't all that awful. The electronic solution I suggested would need much the same kind of coil, probably wound onto something like an RM10 core former, with the RM10 core halves gapped by the thickness of the Hall effect field sensor and held together with the loaded coil former with a Ty-Rap or two or the like.

John Larkin doesn't get as specific as he should - though Cursitor Doom doesn't really deserve the effort.

You can tune the B field with a permanent magnet, biasing the reed to any necessary extent for a given wire winding.

No, reeds are WONDERFUL; you can listen for the click instead of wiring up an idiot light.

To tune it, you engage the reed with a magnet, put 200 mA through the coil, back off the magnet until it sounds off, and THAT's where you glue (or clamp) it. Every alarm-your-windows kit has the parts you need, ready to stick down wherever you please.

Electronics content: the magnets and reeds will stick to a standard bread-slicing board. It's called 'breadboarding'.

Reeds have enormous hysteresis.

They are big, expensive, erratic, and unreliable.

Try it.