Hi,

A while back @Carl came round and we did some experiments with all his instruments to have a look at my grid-tied (no pushback) setup.

I think these two pics are interesting:

Here's his scope connected so that channel A is measuring voltage on the supply from CoCT.  Channel B is measuring current.

In this test the inverter was disconnected from the grid side.  So all the main house load was being supplied by the city.

At the top you can see that I have 223V RMS and 11.9A RMS - 2654VA.  

You can see that the voltage and current are in phase so the power factor will be close to 1.

(There is a current non-linearity near 0v which is probably a result of all the many many switcher power supplies in my house).

 

We then reconnected the inverter.  The panels were probably making 2kW or so.

So now looking at the numbers - 222V RMS but now only 2.22A RMS - so now drawing only 493 VA from the city.

But its interesting to see how the current load looks now - you can see that the inverter is pushing 90 degrees out of phase:

 

Interested in input from the clever people about this.   For me it was cool to see the waveforms - makes the idea of power factor and all that visible.

 

I don't trust anyone that leaves the clear remove on first use sticker on their devices. Shame! Shame! Shame!

The "angular" current waveform is due to large 3rd harmonic distortion - which is actually in-phase with the voltage waveform. There are also some lower-level harmonics that are probably 9th order.

This harmonic distortion is so extreme since you probably only have non-linear loads (switching power supplies etc.).  As soon as one turns on a linear load (incandescent bulb, kettle, toaster etc), the current waveform will then look a lot more like a normal sine wave.  This distortion is probably not an issue for you, due to it only happening at low power levels, but does become a significant problem at high power levels for industrial customers.

 

Edited July 22, 20196 yr by NigelL

31 minutes ago, NigelL said:

The "angular" current waveform is due to large 3rd harmonic distortion - which is actually in-phase with the voltage waveform. There are also some lower-level harmonics that are probably 9th order.

This harmonic distortion is so extreme since you probably only have non-linear loads (switching power supplies etc.).  As soon as one turns on a linear load (incandescent bulb, kettle, toaster etc), the current waveform will then look a lot more like a normal sine wave.  This distortion is probably not an issue for you, due to it only happening at low power levels, but does become a significant problem at high power levels for industrial customers.

 

HI,

The load was roughly the same as the first test.  It was primarily a resistive load IE the kettle.  The load from my switching PSU stuff accounts for a couple of hundred watts of that total load which was about 2500W.

 

Hi Elbow

I was referring to the second picture - where you said that you only had 2.22A load (493VA).  The larger linear loads mask out the effects of the smaller non-linear loads (as per picture 1).

2 hours ago, NigelL said:

the current waveform will then look a lot more like a normal sine wave

I would expect the current waveform to be shifted pi radians (or 180 degrees for the rest of you plebeians) from the voltage, and to be a nice smooth sine wave.

3 hours ago, root said:

I don't trust anyone that leaves the clear remove on first use sticker on their devices. Shame! Shame! Shame!

Man... that thing probably cost 30k+. Leave that sticker on for as long as you want as far as I'm concerned!

Edited July 22, 20196 yr by plonkster

4 hours ago, root said:

I don't trust anyone that leaves the clear remove on first use sticker on their devices. Shame! Shame! Shame!

? I do that until you can't see through it anymore!

25 minutes ago, plonkster said:

I would expect the current waveform to be shifted pi radians (or 180 degrees for the rest of you plebeians) from the voltage, and to be a nice smooth sine wave.

The voltage and current are normally perfectly in-phase for a linear load, but this totally depends on how the current sensor was clipped onto the wire 😀

On a side note, one needs to be careful of high-power 3rd harmonic distortion in 3-phase networks since this can result in a situation where the Neutral wire has to carry up to 3 times the normal current (i.e. the combined-total of the currents in each of the 3 phases)!  This is because the 3rd harmonics of each of the 3 phases all combine together, in phase, as a very large 150Hz current waveform.

1 hour ago, NigelL said:

The voltage and current are normally perfectly in-phase for a linear load, but this totally depends on how the current sensor was clipped onto the wire 😀

Indeed. My thinking is that if you have a linear load with a power factor of one (and let's assume the CT is clamped that way), then the current and voltage will be exactly in phase. If you then tell your inverter to push back, and assuming the inverter maintains a power factor of 1, I'd expect the current waveform to be exactly opposite to the voltage waveform.

And if you are balancing the power to zero... well then some really weird stuff usually shows up and the current appears to be around pi/2 out of phase (as is seen here).

51 minutes ago, plonkster said:

... well then some really weird stuff usually shows up and the current appears to be around pi/2 out of phase (as is seen here).

I have a question: If there are 100+ homes on one transformer, does that have a impact on the network?

Edited July 22, 20196 yr by Guest

14 hours ago, The Terrible Triplett said:

I have a question: If there are 100+ homes on one transformer, does that have a impact on the network?

I don't know. I've read articles that explain how grid-tied inverters make the overall power factor worse. It you keep in mind that apparent power (S = VI) is the vector sum of the real power (P) and the reactive power (Q), then a grid-tied inverter that runs at near unity (usually regulations require 0.9) will cancel out more of P than it does of Q, and this makes your power factor worse.

(Cos theta approaches zero as P approaches zero).

Image borrowed from Quora:

I don't yet know what happens if the entire neighbourhood is doing this.

Edited July 23, 20196 yr by plonkster

@Rautenk - do you have a thought to share re. the above two posts?

20 hours ago, NigelL said:

Hi Elbow

I was referring to the second picture - where you said that you only had 2.22A load (493VA).  The larger linear loads mask out the effects of the smaller non-linear loads (as per picture 1).

No sorry - I obviously wasn't clear.  The actual house load was the same, but the inverter was supplying most of it.  There was a NETT draw of 493VA from the council.

 

16 hours ago, plonkster said:

Indeed. My thinking is that if you have a linear load with a power factor of one (and let's assume the CT is clamped that way), then the current and voltage will be exactly in phase. If you then tell your inverter to push back, and assuming the inverter maintains a power factor of 1, I'd expect the current waveform to be exactly opposite to the voltage waveform.

And if you are balancing the power to zero... well then some really weird stuff usually shows up and the current appears to be around pi/2 out of phase (as is seen here).

Hi @Plonkster,

So my inverter is a high frequency inverter unlike the blue ones - which likely has an impact.

In this test the load wasn't balanced back to zero - there was a nett of 490-ish VA still coming from the council where we were measuring.

The test setup was the same for both tests - as you can see the voltage and current was in phase with the inverter disconnected - so the CT was indeed clamped that way.

Maybe if @Carl feels like it and we get another sunny weekend day we can do another test where we are able the balance the power to zero and I think we will see the current move to 90 degrees out of phase.  This seems consistent with the very low power factor my ET112 reads when the power is balanced to zero.