Flux gate - practical aspects?

[Skogsgurra]

Hello again,

I have googled for Flux Gates and found a lot of information. Now, I have to decide; shall I roll my own and get it tailored to my needs (detecting ground fault in a DC system with minus grounded) or shall I buy a ready-made device with a lot more whistles and bells than I really need. They are also rather big and do not fit in the enclosure available.

I guess that I lean towards building one. Someone got any experience out there? Toroid or bar? Square wave or sine? Synchronous detector? Resonant? What have you used - and what is your experience?

Or is there any other reliable method of detecting 20 - 30 mA DC? (Shunts, Hall effect or magnetoresistive devices all have their problems that make them unsuitable). There is room for about three turns of wire, so hundreds of turns in a relay coil is not an option.

 

[nbucska]

How much resistance may the sensor have? How much is the
DC (V,W ?)



<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[VEBill]

You mentioned that the minus is grounded. Is the wiring such that you have a reliable return path (before it is grounded) ?

Assuming that this is so, then why not just measure PS output current and compare it to the motor return current (in that isolated return lead before it is grounded)? You could use simple voltage drop measurements (shunts) and simple op amp circuits.

It would seem that the 'shunt' approach would be much simpler than relying on any magnetic sensor.

 

[Skogsgurra]

Thanks to both of you,

I have looked at the shunt version (I think that nbucska is thinking of shunts when he asks about sensor resistance) and I do not seem to make it work reliably.

The system is a 350 V DC distribution system and the maximum current will be 200 A DC. I need to measure current in outgoing wire (plus) and return wire (minus). And, yes VE1BLL, the minus is solidly grounded.

The power dissipation in the box (less than a pint's volume) shall be as low as possible and it was decided that 5 W is an absolute maximum. So that leaves 25 mV for the two shunts (if we use shunts) - or 12.5 mV for the plus shunt and 12,5 mV for the minus shunt. We need to discriminate between 20 mA (shall not trip) and 30 mA (shall trip). The difference is 10 mA and that means a shunt voltage difference equal to 0.625 microvolts. I do not think that I would like to think of putting a circuit with that specification to work where people's lives are at stake. Yes, I know that there are beautifully stabilized opamps that probably could do it. But what about precision resistors for the differential network? How stable can they get in a box where anything between 0 and 5 W are dissipated and with a changing ambient? Not to mention power supply issues.

Sorry - lots of words - but I feel that I had to elaborate on why I do not want shunts.

So, the question remains: Do you have any practical experience with magnetic techniques? Probably Flux Gates. Or any other technology (No Hall effect or magnetoresistors, please. They also have severe temperature problems).

 

[IRstuff]

One thing you haven't indicated is what kind of sensitivity you're looking for. Is it compatible with the size you want?

Another thing is that the level of sensivity available, even in a cheap throwaway compass, such as the one built into the i-Top toy, is that the flux gate is sensitive to ANY fields strong enough to register. My i-Top goes bonkers anywhere within 2 ft of a cell phone, of course, that may be strictly an E-field problem. So, if you have ANY circuitry in the vicinity that's generating a field, you'll likely have some level of interference. Likewise, any moving magnets or fields in the vicinity can potentially cause interference.

Also, since the flux gate senses all fields, any metal objects, etc., will distort the local field, which means that the sensor will read different values depending on it physical positioning and placement.

TTFN

 

[IRstuff]

In case you haven't run across it yet, the i-Top is pretty cool.

One of its modes is a Magic 8-ball function and you spin the top to see the LED-scanned display. In order to keep the display consistently oriented, it has a built-in digital compass.

TTFN

 

[VEBill]

Looks like this function (Ground Fault protection of industrial 300VDC circuits) *might* be available off the shelf...

http://www.bender.org/prod.html

 

[Skogsgurra]

nbucska,

Do you have a working design? We need just one device right now. It is for a test and demonstration installation. And - of course - we need a spare part as well.

Please indicate specifications, size, approximate cost. 20 ms is actually a little too fast. Somewhere between 50 and 100 ms is better.


IR,

Never heard of the i-Top. Working with flux gates?

I plan to use a bifilar wound coil with three turns and enclose the thing in a ground connected soft iron tube to reduce influence from geomagnetic field. I hope that it will be a sufficient screen against mobiles and such sources too. And there will be a delay to avoid nuisance tripping. If you plough through my lengthy text you will find that I need to have no trip at 20 mA and trip at 30 mA differential current.

 

[Skogsgurra]

VE1BLL,

Yes. I had a look and it seems to be able to do what we need. Thanks.

I talked to the local Bender guy, but he knew nothing about this a couple of months ago. Perhaps better give him a ring tomorrow.

What I do not understand is that they clearly show measurement of a L+ and a L- wire bifilarly fed through the transformer. And that is fine. But the specification says 15 - 400 Hz? There must be some other current transducer available - with a flux gate sensor, perhaps? But why show DC wires through this AC transformer? Confusion reigns...

 

[nbucska]

it may not be a xformer -- they may sense the DC field in the gap.

<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[IRstuff]

Think it might have been magnetorestrictive, at least when I took it apart, I didn't see anything resembling a typical flux gate.

TTFN

 

[Skogsgurra]

We boys are all alike. Toys' MTBD (Mean Time Before Disintegration) always in the hours to days range.

Possibly magnetoresistive, but certainly not magnetostrictive. Can work in a device that doesn't need to bother about accuracy but needs to keep track of directions and relative field-strength.

Seems to be impossible to buy an i-Top over here in Europe. Will ask my friends in Albuquerque to send me one.

 

[Skogsgurra]

End of discussion - I think.

I managed to get in contact with the right guy at Benders. He said that the diagram was the right one in the wrong place. He also found the right unit and the right current transducer. It can detect 30 mA and still let a couple of 100 A through. And it can do it in 70 ms, which happens to be what we need. The device is an RCMA47xLY-xx plus an external current transducer that works from DC up to 150 Hz.

Thanks for an interesting excercise and good tips.

 

[IRstuff]

duh.... I knew that! really!!

TTFN