Noninvasive current sensing.

[kellym]

I have a single cable with 24vdc pos/neg within that switches a load of approx 200ma. Is it possible to determine when that line is active (switched on and running the load) by any noninvasive means currently available? Whether Hall Effect or other transducer/transformer?

Ideally, I'd like to be able to monitor such a line for on/off status without having to open the line and separate the conductors. I've been told that this is impossible - that any (noninvasive) attempt to detect a load will be cancelled out by the balanced lines.

Is this true?

 

[MikeHalloran]

Is the root limitation that:
- You don't want to interrupt the current?
- You don't have physical access to the cable?
- You don't have legal access to the cable?
- You need to install the sensing in a remote location, where e.g. good power sources for the sensor are not available?
- You need to develop a technique and equipment to do this rapidly/ many installations/ with unskilled labor?
- You don't have any money?

I.e., I don't think it's impossible, but it may be difficult and/or expensive.



Mike Halloran
Pembroke Pines, FL, USA

 

[IRstuff]

Depends on what the load is. If it's circuitry, the it's possible to see more EMI noise when the circuitry is on, than when it's off.

If it's a purely resistive load, that's much harder. There might be some detectable temperature change, but's probably pretty miniscule. The magnetic field ought to be neutral, but at very short distances, there might be some differences.

TTFN

FAQ731-376

 

[Skogsgurra]

I would say it is doable. It is true that the forward and return currents cancel each other at some distance. But not very close to the cable.

A hall effect device probably will not work to detect the miniscule magnetic field. They also have temperature problems. A flux gate can be made very sensitive and has much less temperature problems. Use two of them and wire them so the cable's field adds but earth's field and other magnetic fields cancel.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[itsmoked]

Yes, if you have a very small sensor face and it's pressed against the cable it should see one direction more than the other. You would need to rotate the sensor incrementally until you maximize the detection. This would essentially get it perpendicular to one of the conductors.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Clyde38]

Is the cable long enough to form a coil?

 

[VEBill]

If you have access to both the supply and the return conductors (and you can separate them a bit), then some of the problems with offset, drift and ambient environment can be cancelled out using a differential approach.

 

[kellym]

I'm trying to come up with a reliable, cheap, and universal (any machine) method of monitoring the status of a tig welding unit. An auxilliary process needs to know simply when the arc is energized, and when it is deenergized. No measurement is required, and accuracy is unimportant.

For this to be ultimately useful, it would not require any alteration to the machine or it's accessories. My first approach was to insert an adapter with pigtail between the pedal/rheostat control connection on the welder. Unfortunately, there is no industry standard connection. Next I considered a clip on current probe for the control cable. So far this seems impractical given the requirement that it not require alteration to the unit or cables.

Next I consider a clip on current probe designed for the work or ground to torch lead. Here the probe must deal with 0-300amps AC and DC, pulsed current, and exotic wave forms. Probably a deal breaker by virtue of complexity.

One more current option would be to conceive an adapter at the power outlet to monitor current at idle/energize. This would be mechanically very expensive and again, there are a number of standard outlets, as well as two and three phase considerations.

Lastly, I could put a flow meter inline with the argon supply. Essentially this would monitor the internal gas solenoid of the welder and be relatively nonintrusive and actually quite reliable. Approx 22CFH at low pressure. Off the shelf switches of this type are anywhere from 300 to 800USD. Again, I'm frustrated with being unable to find an inexpensive (20-30USD) solution to this part of the project.

Any other ideas?

 

[itsmoked]

Use the ground return. You said you don't care about actual measurement. The large current flowing in the return would make this a cake walk. And you don't care about exotic waveforms. Use a non-temperature corrected hall effect sensor so you can see the AC or DC welding currents selected.

If you care about the fact that the argon could run on for up to 20 seconds after welding ceases look a little harder for cheap low pressure sensors. You should be able to find ones in the sub $10 range that would work just fine.

Keep in mind that often the argon circuit is NOT used in some TIG welding as flooded work jigs are used instead.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[analogkid2digitalman]

If you need to simply monitor when the arc is energized, use a photocell or phototransistor aimed towards the arc area. Use appropriate ambient light nulling circuity and close a relay or transistor to interface to the auxilliary equipment.

No need to deal with mains voltage or high current.

Just an idea, don't know how feasible it may be for your setup.

-AK2DM

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
"It's the questions that drive us"
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

 

[IRstuff]

must deal with 0-300amps AC and DC, pulsed current, and exotic wave forms. Probably a deal breaker by virtue of complexity

Why is this a dealbreaker? It's just a simple thresholding requirement to detect current va. no current.

However, this seems to be a rather strange device. If the welder is running, I would think that only a legally-blind operator would fail to see the smoke and whatnot coming from the workpiece.

What exactly is the INTENT of this exercise?

TTFN

FAQ731-376

 

[VEBill]

A simple loop next to the output cable should be able to generate enough noisy AC to create a signal that could be easily detected.

 

[OperaHouse]

I actually built about 300 of these devices to put on TIG welders that were used robotically. These welders had a shunt resistor for the current and the device monitored the voltage on the shunt. You don't want the circuit to be fooled by HF TIG initiation voltage. Never had one complaint on how they operated.

 

[danw2]

Fluke makes a handheld, clamp-on meter for measuring 4-20mA loop control circuits without breaking the loop, the model 771.

It's a $400 tool, with its relatively high accuracy (0.2%), not what the OP is asking for, but what technology does it use? Something that can be adapted for on-off detection at 10x the magnitude of the range of the Fluke?

 

[Skogsgurra]

Wires shall not be separated. You have to detect the residual field from the two cancelling currents. A flux gate, as I said earlier, is the way to go. Done that. It works. Have a look at

http://www.speakesensors.com/

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[VEBill]

Even with a DC arc welder, isn't the waveform of the arc current sufficiently noisy that it could be (easily) detected using AC techniques (as opposed to assuming it is just pure DC)?

Similar arcs used to be used for radio communications...

Using AC techniques would eliminate many of the issues mentioned above.

 

[itsmoked]

Excellent point!

Keith Cress
kcress -

http://www.flaminsystems.com

 

[OperaHouse]

There is a high frequency/high voltage placed on the line to initiate the arc. You do not want to confuse that with the arc current that actually tells you when the arc is established. Rule out sensing any high frequency to determine if an arc is established. Stick with DC current sensing.

 

[kellym]

update to this..

I need to tell (trigger with a signal level relay) a separate process control when there is arc established, and when it is stopped. This needs to function with any tig welding machine and be non contact (not require any modification of the machine/process). I only need to recognize/sense the first couple of amps through the cable. When that threshold is reached and lost.

The HF start would not be an issue. The trigger does not require strict accuracy in this regard. It can activate upon application of HF for arc starting since the arc will follow anyway in short order.

Ideally I could clamp on the pedal/rheostat control cable (pedal on.. pedal off) but here I don't have wires separated out as with the torch current supply, and I'm dealing with only milliamps to boot. Perhaps the fluxgate as was mentioned can be investigated. I'll read that link.

1 - 300 amps range is mentioned only to point out what the process is capable of. I am only concerned with the first few amps. In need to sense arc starting and arc stopping. I need to identify it at AC and DC, but as has been pointed out, in this application I may be able to use AC sensing and still get by.

I've spoken with a few manufacturers now and generally have been told that one or two amps is going to be extremely difficult to sense. The current supply to torch is approx 12ga copper with a water jacket, the whole being approx the diameter of a bic pen. No telling what the water jacket may effect on sensing. The return is commonly a 2ga copper ground wire. Non-contact sensing two amps through this is, I'm told, pretty impossible.

MR sensing technology has been suggested a few times, but noone seems to be very up to speed on it. Finally, it was suggested that I simply get this

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=620-1213-ND

and just start experimenting!

I had no idea this was going to be such a can of worms.

 

[Skogsgurra]

I must be the lousiest communicator!

I have mentioned two times that a flux gate can detect the flux unbalance from the 200 mA DC current going in two wires with opposite current direction. A flux gate does not have temperature issues (like Hall elements and MR sensors). It just works. You may have to shield it from the Earth's magnetic field, but not always.

Again: have a look at

http://www.speakesensors.com/

I have used their products in custom RCDs for DC. Trip level between 20 and 30 mA achievable with no sweat.

Is my english so bad that it can't be understood? Or is it the flux gate technology that can't be understood?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...