How do non-contact voltage probes actually work? 2

[LiteYear]

This question doesn't neatly fit into electromagnetism, antennas, circuit design or power engineering but probably requires knowledge from all fields. I keep coming back to this seemingly simple question and never leave quite satisfied...

A non-contact voltage probe, aka a voltage stick, when brought near a live AC conductor will flash a light or sound an alarm. Elementary circuit analysis will tell you there are two possible methods of detection:

1) Inductive coupling: effectively the probe is the secondary of an air-cored transformer, with the live conductor as the single turn primary. This is not the whole story because this mechanism only works when current in flowing in the conductor under test, which is not a requirement for these devices.

2) Capacitive coupling: effectively the probe is one plate of a capacitor and the live conductor is the other plate. The dielectric is the air and plastic in between the two. An alternating current flows from the source, through the capacitor, to ground via the operator and back to the source. This is also not the whole story because the operator doesn't need to be part of the circuit for the device to work - it still works when wearing rubber gloves or in fact (as our experiments with lots of sticky-tape show), when the device is not being held at all!

All concepts from radio/antenna/electromagnetism are flawed in a similar way to (1) - without current there is no magnetic field. Electric field solutions are also unsatisfactory because they generally require a common reference.

I'm sure it's some capacitive/electric field effect at work, but I have not being able to find the exact mechanism. Can anyone help?

 

[IRstuff]

Do you have a specific product as an example?

TTFN
faq731-376

 

[LiteYear]

Sure, the one on my desk at the moment is the Digitech QP2269.

http://www.jaycar.com.au/productView.asp?ID=QP2269

But I'm also curious about the higher voltage Modiewark style products.

http://glmcgavin.com.au/products/modiewark.html

 

[VEBill]

I understand that the use capacitive principles.

Let me see if I can find a reference.

 

[VEBill]

http://www.trekinc.com/pdf/3001_Vibrating_Probe.PDF

 

[jraef]

They are capacitive. Read this, it's a fairly good overview of the principal.

http://ecmweb.com/content/what-do-you-know-about-capacitive-voltage-sensors

In theory it should not have worked if you were not touching it, i.e. your "sticky tape" experiments (video would be fun), but it may have been some sort of unnoticed a flaw in your experiment, essentially you may have STILL been part of the series capacitance circuit without realizing it. These things are very sensitive.


"Will work for salami"

 

[Skogsgurra]

You can make a very simple test with a CMOS gate. Find a simple gate, like a 74ACT00 or HCT or anything CMOS. A 4011 will also work.

Connect all inputs of gate #1 (or any of the gates) together* and solder a piece of wire, a foot or two, to it. Connect a 3 or 4.5 V battery to the supply pins (careful with polarity) and an LED with a series resistor use one kilohm, it will usually be safe and adequate, to the gate output (polarity again!).

Presto! There's your non-contacting voltage probe. Grab the battery and near the wire (without touching it) to an electrical cord or outlet. If there's AC, you will see the LED light up - or at least change intensity (if it was already lighted).

We have used that to check for voltage in remote sites where one needs to check for voltage in a simple and safe way. The Stugvakten is one example of application.



*Connect all unused input pins to either V_ss or V_dd to keep them quiet and not drain the battery.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[LiteYear]

VE1BLL,

Thank you, but I don't think that's quite it. Voltage sticks don't vibrate, don't require relative movement, don't detect DC (static charge) and don't meter (only a yes/no indication). The mechanism in the paper requires that the potential difference between the probe and ground is known. A voltage stick on the other hand, could be at any potential to ground or even entirely floating.

Curiously enough, the volt stick will actually light up if I hold it against my clothes and drag it back and forward.

I suspect the mechanism is related to electrostatic capacitive charge, but I'm yet to see a convincing description of how the mechanism jumps to AC detection using a floating probe.

I'm starting to think something like this could be involved:

[tt]
| |
(Neutral)---(AC Voltage Source)---(hot conductor)---| air |-------(probe)-------
| | |
(GND)
[/tt]

No circuits, no referencing. With everything dead, the charge distribution on the conductors is even:

[tt]
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- +-+-+-+-+-+-+-+-+-+-+
[/tt]

When the voltage source is switched on, no current flows (except a minute amount of "charging current"), but the charge distribution changes. The voltage forces the charges in the conductor to separate. When the voltage source reaches its maximum voltage, the charge distribution might look like this:

[tt]
-------+----+---+--+-+-+-+-+-+-++-+++-+++++-++++++++ ------+---+-+++-+++++
[/tt]

The distribution in the probe also redistributes, due to magnetic forces in a conductive medium.

When the voltage source alternates to its minimum voltage, the charge distribution reverses.

I guess if this cycling happened continuously, that would constitute a "current" (even though there is no circuit) that could be amplified and detected. Obviously I'm still a little hazy on the details.


jraef, good reference - that has been my general, vague understanding of their operation since I first came across them. It wasn't until one Friday afternoon beers where we ran around the office standing on plastic chairs, jumping into the air, and making triggers out of sticky tape that I started to doubt I completely understood the mechanism. I'm willing to believe it's still the correct mechanism, and that the high sensitivity helps to outweigh my attempts to thwart it, but there is still something I don't understand. If it's just a capacitive divider, then it would not tolerate the huge variations in capacitance to ground caused by chairs, jumping and sticky tape, while still being highly sensitive to the small change in capacitance created by drawing the probe tip a few inches away from the conductor. Perhaps this calls for more Friday afternoon experimentation!


Skogsgurra, that's an interesting insight. So what's the mechanism?

 

[LiteYear]

Crap, ASCII art failure. The 3 figures should be:

.
                                                    |       |
(Neutral)---(AC Voltage Source)---(hot conductor)---|  air  |-------(probe)-------
          |                                         |       |                       
        (GND)

+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-         +-+-+-+-+-+-+-+-+-+-+

-------+----+---+--+-+-+-+-+-+-++-+++-+++++-++++++++         ------+---+-+++-+++++

 

[Skogsgurra]

The mechanism is capacitive coupling and extremely high input impedance of a CMOS gate. It is in the Gigaohm to Teraohm range.

Typical capacitances are a few pikofarad. Couple that with a Gigohm resistance and you get a time constant that equals milliseconds. With 1 Tohm you get seconds. The old CR circuit again, see

http://www.gke.org/presentationer/files/Smoked probe compensation and other topics the sequel.pdf

With a CR time constant equal to milliseconds to seconds, the attenuation is very low. So the mains voltage reaches the gate. It doesn't hurt it because the energy transferred through the few pF is miniscule.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[LiteYear]

Skogsgurra, but where's the return path? Capacitive coupling works when both circuits are referenced to the same point to create a circuit. Your circuit compares the voltage on the terminals to a floating battery voltage level. Without a return path the whole battery referenced circuit would just float up and down.

The usual explanation given is the one in jraef's link - that the return path is provided by capacitive coupling through the body. But then the original question remains - why is the phenomena so insensitive to enormous various in the impedance of that return path? Besides, if there's capacitance from the reference point back to ground, then there'll also be capacitance from the reference point back to the wire under test. Why is the impedance from the reference point back to ground always sufficiently low to create a stable indication?

Take a look at the attached photo I just took - I've taped the "on" button down on a voltage stick, placed it on a paper/plastic booklet and placed that on an energised, but unused 240V/50Hz power board. As you can see the indicator light on the tip of the probe is lit. If I turn the power board off at the wall, the light extinguishes. But if the power board is on, there are not too many orientations where the probe doesn't light, provided it is close enough. Curiously enough, yesterday I had the probe sitting on a 50mm cardboard box and it still lit up - today I can't replicate that.

During these experiments I was able to get the probe in an orientation where the presence of my hand did influence the indication - the probe alone would not light but if I brought my hand very near it or touched it, it would. So I'm certainly willing to believe that the body can be a current path. But just going by the numbers, it doesn't seem to add up to the whole story.

For example, the standard explanation still fails to the behaviour I mentioned before: the volt stick will actually light if I hold it against my clothes and drag it back and forward. This trick is pretty well known amongst sparkies. Now where's the voltage source and return path?

 

[Skogsgurra]

The return path is capacitive. As is the "entry" path.

The explanation to the insensitivity to the "enormous" variation in capacitance (the insulating tape etcetera) is that the probe is oversensitive and works over a wide range of conditions.

I cannot really understand what your problem to accept such a simple fact is.

Do you think that there's some magic involved?

As for the dragging against your clothes. You get a series of charges/discharges and that triggers the electronics in the probe. Most of those Voltage Sticks have a monoflop that keeps the LED lit for a little time after each triggering. So, the dragging against the clothes doesn't prove that the mechanism is other than I say. It rather proves it.

Do you have any idea of how capacitive paths work at all? Or do you need to see a visible wire to believe in electric charge transfer, aka "current"?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[LiteYear]

Wow, got harsh all of a sudden. Sorry to frustrate you, maybe we're misunderstanding each other - the tape was not intended to be an insulator. It just keeps the button depressed.

You say the dragging example is due to charges/discharges (I assume you mean triboelectric charge), not capacitive coupling of a voltage source. I could be persuaded that both have the same net effect, I just am unconvinced that it amounts to a proof.

I think it's clear I have some appreciation for current through capacitive paths, but perhaps you could fill me in on the finer points. In the attachment I've attempted to model the scenario in my earlier photo.

I was impressed to find that by playing around with the model, and in particular using a very large sensor resistance, I could make the sensor relatively insensitive to the magnitude of the capacitive coupling to earth, and even insensitive to the relative magnitudes of the coupling to earth and to hot. But then in that configuration, the sensor is also highly insensitive to the magnitude of the capacitive coupling at the probe tip. Yet, if I draw the probe away from the conductors by only a centimeter or two, the light extinguishes. That extra separation represents say, a double or tripling of the capacitive separation, so capacitance should only fall by a factor of about 2 or 3. This makes bugger all difference to the voltage across the probe sensor.

The law of big numbers says that our intuition tends to break down when dealing with them, and I think that's part of why I'm confused. I just can't imagine a configuration that provides such dramatic insensitivity to the coupling in the return path, yet is quite sensitive to the coupling at the probe tip. The problem of course, with just ramping up the probe's sensitivity, is that in our electromagnetically saturated world, the probe light would never extinguish!

I suspect the non-linear radial distribution of electric field around a conductor plays a large part here. But the simplification expressed by the electrical treatment of capacitance doesn't really lend itself to incorporating that effect.

Are there any improvements on the model you can suggest that will more accurately reflect the probe's behaviour?

 

[Skogsgurra]

Do you also fail to see that the repeated charge/discharge when "dragging" represents a rudimentary AC voltage source. The discharges contain very high frequency components and since there's a timer (pulse stretcher) involved, you only need one discharge to light the LED for a while.

Yes, you are right about the radial field. And that the E-field falls quickly when removing the probe tip from a live wire.

Harsh? Yes. This site is for engineering questions. Hobbyists can find playgrounds other places. Harsh? Yes - someone that seems to have a hidden agenda and never can accept an explanation, or find that explanation himself even when it is pointed out, makes me harsh.

Question: Are you trying to prove an alternative explanation? Or are you just thinking that there is no explanation?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[LiteYear]

No, I don't fail to see the rudimentary voltage source. I fail to see it as proof.

Yes, you are right about the radial field. And that the E-field falls quickly when removing the probe tip from a live wire.

But if the coupling is capacitive, the fall in capacitance is only 1/d. If the coupling to earth is also capacitive, then its fall ought to be the same. Why is the probe so sensitive to the proximity of the conductor and so insensitive to the proximity to earth?

I've read lots of your posts and I respect your knowledge and expertise and contributions, but your professional interests are different to mine. I'm a research and development engineer and we design things that don't exist yet, like voltage sticks once did. Knowing "it's capacitive" is not sufficient for the professional work I do, but it likely is plenty sufficient for you. To suggest my line of work as hobbyist is insulting.

But I don't want to turn this into a mud-slinging match as I'm sure you don't, so let me be clear: what is the electrical model (with all the relevant parasitic elements) that demonstrates the behaviour exhibited by these devices? These things exist, they work, I don't believe in magic, I have a few rough ideas (as fleshed out in this thread), but I'm yet to find a model that is sufficiently accurate. Perhaps the model I offered could be critiqued or improved. If I never do find such a model, then so be it, I'll just accept the best we have and go away.

 

[Skogsgurra]

I have described the model as good as I can. I repeat:

1. A voltage with respect to ground. That voltage can be an AC voltage or a transient voltage.
2. A coupling capacitance that varies from tiny pF to, maybe, tens of pF.
3. A high-resistance detector that is coupled to the input capacitance (this is the tip of the stick). Resistance is in the gigohm to teraohm range. The detector includes battery and a pulse-stretching circuit so that one single nanoseconds edge results in a visible blink.
4. A coupling capacitance from body of stick to ground. Again, a few pF if stick left hanging in the air. And perhaps around 100 pF if stick kept in hand.

Sorry, I cannot be any clearer than this. And if you think that such an explanation "is not sufficient for the professional work I (that is you) do, but it likely is plenty sufficient for you" (talk about insulting) then I can offer you our design, analysis and simulation services. The price list is available here:

http://www.gke.org/pub/files/GKE Elektronik AB fees and compensation 2012.pdf

(Note for site administrator: this is not selling or recruiting - it is argumentation)



Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[VEBill]

General comment follows.

I fail to see it as proof.

On the Internet, anything can be proven.
On the Internet, nothing can be proven.
Both of the above statements are perfectly and exactly true.

Providing proof is not the role of Eng-Tips. We provide tips, not 'proof'.

 

[Skogsgurra]

Very true, Ve1BLL!

I am somewhat surprised to see that our friend BrightLight does not understand the difference between the field pattern from a thin wire and that from an (almost) flat surface. The Earth is almost flat, you know.

Or is that something I need to prove also?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[LiteYear]

VE1BLL haha nice one, will keep that in mind.

Skogsgurra, wow, name calling of strangers over the Internet. Please don't be offended if I don't end up taking up your offer for your professional services.

Thanks for your contributions everyone. Even if the exact mechanism is not clear to me a few of the details are certainly clearer.

 

[Skogsgurra]

OK LiteYear. We didn't have a good start. Let's hope that we get on better when we get used to each other.

I never looked at your "voltage_stick" LT Spice diagram. Simply because I didn't bother any more. I have looked at it now. It is correct. Except for the body-ground 10 fF capacitance. It is a lot higher. Like 10 - 100 pF.

The difference between the capacitance between probe tip and wire is that the probe tip is a point and the wire surrounds itself with a radial/cylindrical field. So capacitance falls off quickly when distance is increased while the body (still somewhat 'pointish') is in the homogenous field from Ground and surroundings so that the capacitance doesn't change much when the position in that field is changed.



Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.