Noncontact electric field sensing

[Oznog]

I would like to be able to calculate REAL power on our building's mains, without opening up the wiring of course.

I can get the clamp-on inductive current sensors. That's straightforward enough.

However, it is still important to know the phase relationship of the voltage to current. Inductive or capacitive lead/lag is important too.

It's not essential to accurately read the magnitude of voltage, but that would be nice. The first goal would be detecting the zero crossing in the voltage to compare with the zero crossing in the inductive current sensor.

That should be easy, right? Would you just put a metal strip beside the wire's insulation and measure the capacitive induction?

Is it practical to get an accurate voltage reading off this, like +/- 2%?

 

[Tony Stewart]

Can you access the power meter rotating disk with an imaging solution?

 

[Oznog]

The "spinning disks" got replaced years ago. Now they're "smart meters", which actually do have a IRLED pulse every time it distributes a certain count of watt-hrs.

But I don't actually want to use that.

 

[alec_t]

You could use a mains-to-low-voltage AC transformer (wall-wart) to monitor the mains voltage and phase safely. With a very light load on the tranny its output voltage and phase should have a constant relationship to those of the mains.

 

[Misterbenn]

https://openenergymonitor.org/emon/guide has a very good description of how to do this

 

[crutschow]

.....................
That should be easy, right? Would you just put a metal strip beside the wire's insulation and measure the capacitive induction?

Is it practical to get an accurate voltage reading off this, like +/- 2%?

You would need to calibrate it to get even ±10% accuracy.
You could use a multimeter to accurately directly measure the line voltage for this calibration.

 

[Reloadron]

You don't mention what your mains power is? Volts? Approximate current in Amps? You also don't mention if your mains power is single phase, split phase or poly phase (3 phase)?

You could likely use a split core current transformer at your mains in in conjunction with a voltage transducer, that will give you current and voltage. You can buy a small panel type phase angle meter for around $200 or less for a Chinese import on Ebay. There are several ways to do what you want to do but more information about your mains power is needed to really give you a good answer and some suggested equipment.

Ron

 

[MrAl]

Hello,

The best non intrusive way would be to use a Hall Effect current sensor. This doesnt make contact with the wire but will still provide phase data. They make clamp ons for scopes but you probably want something cheaper.

Alternately since it is AC, you could 'build' a transformer 'around' the wire. This would be an ordinary current transformer using EI laminations but you would interleave them yourself as you build the transformer core around the wire. Even easier would be to just build your own butt stack transformer core, where you wind the secondary before you put the E in place. After you wind the primary you place the E such that the wire goes in one opening and out the other, so that it looks like one turn around the center leg. Then install the butt "I" and you have a current transformer built around the wire. An alternate to this would be to use a "C" core, where the core is made from two sections in the shape of the letter "C". The two C's are then placed together face to face so that they form an "O", and that's your core. You could wind the secondary all on one "C" and then when done place the two C's around the wire and use something to hold them securely together face to face. This forms a real current transformer, and you never have to break the wire, and you dont have to rely on excess slack in the wire for installation either.
Using a burden resistor you'll get a voltage that matches the current proportionally and phase that is close to the current phase.

 

[Tony Stewart]

I worked several years in the Automated Meter Reading biz, ( Wireless 2 way AMR)

What is the purpose of this project? Accuracy? UX ?

 

[Oznog]

You could use a mains-to-low-voltage AC transformer (wall-wart) to monitor the mains voltage and phase safely. With a very light load on the tranny its output voltage and phase should have a constant relationship to those of the mains.

No, under no circumstances will the mains be electrically modified, and that includes tapping in an AC transformer or sensor wire.

You don't mention what your mains power is? Volts? Approximate current in Amps? You also don't mention if your mains power is single phase, split phase or poly phase (3 phase)?

You could likely use a split core current transformer at your mains in in conjunction with a voltage transducer, that will give you current and voltage. You can buy a small panel type phase angle meter for around $200 or less for a Chinese import on Ebay. There are several ways to do what you want to do but more information about your mains power is needed to really give you a good answer and some suggested equipment.

Doesn't change the nature of the project. 3ph 208v = 120v legs most likely. Not important.

The current sensor is not an issue. That's straightforward and there's several options. Sensing the wire's electric field, not magnetic field (current) without electrically clipping onto the mains is the problem.

The best non intrusive way would be to use a Hall Effect current sensor. This doesnt make contact with the wire but will still provide phase data. They make clamp ons for scopes but you probably want something cheaper.

Hall Effects only sense magnetic field from current. They cannot sense the voltage at all, AC or DC. It is impossible to determine voltage, phase angle, or power factor, or power from a Hall Effect.

As you can tell this is not a simple question, if it were readily solvable I'd have already done it.

There are noncontact voltage sensors used for determining if an AC wire is "hot" without finding copper to probe. They are not remotely accurate for measurement, but they are not applied in a consistent manner. AFAIK they are still reliable for detecting zero crossing. I'm wondering if the sensing is applied in a fixed installation (no distance or geometry variations) if it can't simply be calibrated and become an accurate voltage measurement.

 

[Reloadron]

No, under no circumstances will the mains be electrically modified, and that includes tapping in an AC transformer or sensor wire.

Doesn't change the nature of the project. 3ph 208v = 120v legs most likely. Not important.

The current sensor is not an issue. That's straightforward and there's several options. Sensing the wire's electric field, not magnetic field (current) without electrically clipping onto the mains is the problem.

Hall Effects only sense magnetic field from current. They cannot sense the voltage at all, AC or DC. It is impossible to determine voltage, phase angle, or power factor, or power from a Hall Effect.

As you can tell this is not a simple question, if it were readily solvable I'd have already done it.

There are noncontact voltage sensors used for determining if an AC wire is "hot" without finding copper to probe. They are not remotely accurate for measurement, but they are not applied in a consistent manner. AFAIK they are still reliable for detecting zero crossing. I'm wondering if the sensing is applied in a fixed installation (no distance or geometry variations) if it can't simply be calibrated and become an accurate voltage measurement.

Well then since you seem to have all the answers just go ahead. I based my reply on your original post and apparently we each feel different things are important. If you want to know "Real" power for example then you should know that you need to know if the power is single or poly phase which I asked but you see as unimportant. I wish you the best with your project.

Ron

 

[Oznog]

Well then since you seem to have all the answers just go ahead. I based my reply on your original post and apparently we each feel different things are important. If you want to know "Real" power for example then you should know that you need to know if the power is single or poly phase which I asked but you see as unimportant. I wish you the best with your project.

Ron

We need to know how many phases, of course. But that's not in any way the question. You'd just add a voltage and current sensor to each leg. That's a trivial detail and I know how to do that. Nor is the current or voltage important to the question at hand, as that's just selecting the appropriate clamp-on current sensor, and if there's a noncontact sensing method for AC that works, it should work equally well on 120v or 460v with minor changes. The question at hand is how you'd get a voltage measurement (amplitude and phase) in the first place without electrically tapping into the wire.

As simple as tapping the wire sounds, the connections at the power entry point and breaker panel are not physically suited for running a signal-wire tap off of. Physically the connectors are not designed to clamp a second wire, nor is there any place to screw a ring terminal. Code wouldn't allow that, nor would the facilities manager. Once you go further out from the panel, the voltage is going to be somewhat less than the mains, and still there's no good way to just tap the wiring. I could plug in a power strip with a voltage-measuring device, but it could get moved and/or lost, it needs to be permanent, which no plug-in device can be.

A clamp-on Hall meter which doesn't alter the wiring, I can get away with that. An antenna that goes alongside or around the wire, I believe that would also be acceptable to do as it does not alter the wiring. But it needs to be consistent and avoid picking up current-induced voltages.

 

[Tony Stewart]

What is the purpose of this project? So I may see if your expectations or requirements are best suited.

"how you'd get a voltage measurement (amplitude and phase)"
.... in a wireless fashion

Answer:
You may an antenna with a known length of Stub wire and measure the Electric Field strength, optimally coupled with a radiating antenna not close to its return path.

Conclusion: Since antenna coupling factor for E-Field is strictly depending on radiator and receiver, an ad hoc antenna is not accurate.

But anyone knows you can measure voltage changes with your finger on a scope probe

 

[Reloadron]

When it comes to the mains voltage entering the facility I do not know of any way to measure it without physically tapping into the entry point. Yes, if you measure downstream as you mentioned, there will be loss. All of my experience with this sort of stuff involved being able to tap into the mains switch gears. Mains current was always measured using the old fashion current transformers until the new voltage and current transducers became popular. So while your current can be done using a clamp on sensor or split current transformer I simply can't think of an accurate means to measure the mains voltage accurately by proximity. Hopefully someone will have a thought on how to accurately measure the RMS voltage?

Ron

 

[MrAl]

No, under no circumstances will the mains be electrically modified, and that includes tapping in an AC transformer or sensor wire.



Doesn't change the nature of the project. 3ph 208v = 120v legs most likely. Not important.

The current sensor is not an issue. That's straightforward and there's several options. Sensing the wire's electric field, not magnetic field (current) without electrically clipping onto the mains is the problem.



Hall Effects only sense magnetic field from current. They cannot sense the voltage at all, AC or DC. It is impossible to determine voltage, phase angle, or power factor, or power from a Hall Effect.

As you can tell this is not a simple question, if it were readily solvable I'd have already done it.

There are noncontact voltage sensors used for determining if an AC wire is "hot" without finding copper to probe. They are not remotely accurate for measurement, but they are not applied in a consistent manner. AFAIK they are still reliable for detecting zero crossing. I'm wondering if the sensing is applied in a fixed installation (no distance or geometry variations) if it can't simply be calibrated and become an accurate voltage measurement.

Hi,

Oh so it is the voltage measurement you are after, as well as the current. Usually the voltage would be assumed and then just measure the current.

So it is starting to sound like you are asking too much from anyone except an engineering firm that is wiling to take the contract and study ways in which this could be done.

The measurement you talk about is just too riddled with constraints, which start to make it physically impossible. You can usually have one constraint and still get away with different kinds of tricks, but once you get too many constraints it starts to make it a very tough problem to solve, which may not have any good solution.

For example, you say that you need to measure the voltage right at the mains, yet you cant do anything to the wiring. Well if you have two single wires where one is hot and one is neutral then you might use a differential capacitive coupling setup, where you run a pickup along each wire. It must be differential because you need the voltage between the two wires not the voltage to physical ground. But because you cant alter the wiring and it must be close to the source, that means the sense pickup would probably be too small and more important the hot and neutral wires will be too close together to pick up the signal. The only possibility is to try to sense at the very outside edges of the two wires, the place where the surfaces are farthest apart. But the problem here is you cant get to the wires individually either it sounds like, so you'd have to wing it by building a sensor and placing it along the wire at different places and hope to find a good spot, and then epoxy it to the wire. The other problem is the insulation thickness will be more because of the outer cable, and the ground wire inside the cable might interfere too much too. Also even with a longer pickup, if the internal wires twist then the pickup signal will be diminished unless the pickup sense area can twist with the wire, and that will be hard to do because you can not see inside the wire. Maybe multiple pickups, where each one is adjusted separately along the wire, and each one is kept relatively short.

So with all these constraints the best you could do would be to build up a sensor and amplifier, use DSP to calculate the quantities needed, and move the sensor up and down the wire to see where you might pick up the signal best, then attach it permanently. It must be extremely physically stable or the calibration will change.

To increase the capacitance for the pickup try to make it as long as possible, probably in the shape of two long C shaped metal pieces that cup the wire on each side, one C on each side. It's hard to say how well this will couple and how much will be picked up because the two wires being close together will cancel out some of the field. If the wires could be separated it would work better.

By all means tell us how this works out if you try it. Remember you can always try it at home first too, just by using some lamp cord. Later move to some Romex cable and try that. If you next tell me that they are using BX cable then all bets are totally, totally, totally, and completely, off.

 

[Misterbenn]

Depending on how strong your grid connection is and the resolution you need, you might be able to get the information live from the utility operator. I remember seeing an interface with the national grid in the UK that gave the PF, frequency and voltage (from the closest monitored point to you) it updated every second or so. Whilst interesting I had no need of it at the time.
But i suspect that you'll also now come up with additional requirements for your project along the lines of: must be updated every cycle or must be remote i.e without wifi access.