how to measure the direction of a.c. power flow

[soldar]

I did a test. I took an AC current of about 2 A and measured it directly on the scope on one channel and sensing through a current transformer on the other and the original appears 0.7 mS later on the scope. The CT leads by that amount of time which is about 1/25 of a cycle. Not huge but significant enough to degrade true measurements.

It was a quick and dirty test. I did not do other measurements to see if the phase difference changed significantly with the current. If it is pretty constant then it should not be difficult to correct it either with a capacitor at the CT or maybe just correct the voltage phase so it is in phase with the current reading. Should not be difficult.

But if you want true power reading you need to do it because otherwise You can have a purely resistive load and it looks like it is somewhat reactive and returning some energy when it is not.

The advantage of a CT is the isolation but if you want to have really good phase reference a shunt is much better.

As a side note: I have discovered that the second channel of my 'scope is defective and needs repair. It has a big DC offset upwards and I cannot get 0V into the screen. Oh well. That'll be a different thread.

[Circlotron]

Was that CT phase shift test using the proper burden resistor? I expect that too high a resistance will give phase shift problems.

[coppice]

I did a test. I took an AC current of about 2 A and measured it directly on the scope on one channel and sensing through a current transformer on the other and the original appears 0.7 mS later on the scope. The CT leads by that amount of time which is about 1/25 of a cycle. Not huge but significant enough to degrade true measurements.

It was a quick and dirty test. I did not do other measurements to see if the phase difference changed significantly with the current. If it is pretty constant then it should not be difficult to correct it either with a capacitor at the CT or maybe just correct the voltage phase so it is in phase with the current reading. Should not be difficult.

But if you want true power reading you need to do it because otherwise You can have a purely resistive load and it looks like it is somewhat reactive and returning some energy when it is not.

The advantage of a CT is the isolation but if you want to have really good phase reference a shunt is much better.

As a side note: I have discovered that the second channel of my 'scope is defective and needs repair. It has a big DC offset upwards and I cannot get 0V into the screen. Oh well. That'll be a different thread.

The delay through a CT, and its character, varies enormously with the core type. For some cores its approximately a time delay. For other cores is approximately a phase delay. 0.7ms is a huge delay for a CT. Is it made by VAC? They have some models with huge delays. Most CTs give a delay far smaller, and some give a delay so small that can be ignored for most purposes. For example, most large multiplier CTs are designed to introduce such a small delay that they have little effect on the calibration of the measurement unit whose range they are multiplying for.

A serious problem using CTs with a big delay is that the delay will vary a lot with frequency. Compensating for the delay at the fundamental may be straight forward, but getting the delay right for all the harmonics can be a nightmare.

[soldar]

I tested with a lower resistor and the phase difference diminished just a bit. I could not measure it with precision because I am having problems with my scope.

Still, it should not be difficult to correct the voltage phase a bit so that it is in line with the current phase.

Frequency is not really a concern because it is for mains.

The purpose is not to precisely measure power as a meter might do but to determine in which direction power is flowing.

[David Hess]

Will a kill-o-watt indicate negative power? I have never tested that. It would be interesting to connect the load to the prongs and connect the power input to the receptacle and see what happens.

I am not sure either.  Next time I visit Home Depot, I will get the parts to make a set of cords to hook one up backwards.

I figured this would be a good simple experiment for this Memorial Day weekend.

I made a pair of reversed extension cords and hooked the Kill-O-Watt that I have up backwards.  It still read positive power flow.  Bummer.

[soldar]

With strictly resistive loads it should not be difficult to do a rough indication by taking a semicycle and seeing in what direction the current flows during that time.

The problem starts when you have reactive loads. Now you can have energy flowing back and forth so you need to do a total measurement and see in what direction is the net flow. Not straightforward or simple.

I wonder if the old electromechanical KWH energy meters would turn backwards. If I remember correctly they probably did but had a ratchet to prevent it.

In that case an old meter could be used to detect the direction of energy flow.

I have a crate full of them. Maybe One day I can do some tests.

[richard.cs]

The mechanical meters without ratchets turn backwards, the ones with usually rotate backwards up to one turn of the disc and then stop. Some types have flags that drop to indicate that this has happened.

[soldar]

Come to think about it, if volume is no problem, a quick and easy way of detecting power flowing out would be to have some switch in place of the ratchet so that when the disk is trying to turn back it is activating a switch. That would be very simple to do using an old meter.

[rgody]

An easy way is to use a known load ( for example "test load" of 220V 220ohms) in conjunction with an ammeter measuring total current to the house .
The test load should be connected for a few seconds only to "see" the ammeter.
If you connect test load and the current in ammeter arise, your house is in "load" mode. If you connect the test load and current decrease, your house is a source.
If you know whether your house is a load or a source, then you can calculate the power just measuring current with no test load connected.
I this idea you just need a contactor, a known resistance, a TC and a simple circuit that just need the current signal to calculate power as a source or load.