For clarification purposes, as we’ve already covered this via PM…
The SXT101 is 240/208/120 Volts to 24/12/2.5 Volts, so the connection I was suggesting was 208 to 12 V.
(with the 2.5 Volt winding bucking the 12 Volt winding to yield 9.5 Volts.)
It had slipped my mind that you’re using an emonPi. There isn’t a 3-phase sketch available for the emonPi, and without that, the phase relationship - assuming you have a balanced 3-phase load - will be wrong if you connect your voltage sample Line-Line, so you must connect Line-Neutral at 120 V to maintain the correct phase relationship.
I apologize for my lack of knowledge in this field, but how is connecting Line-Neutral at 120 V different from how I currently have it set up?
Remind me how everything is set up at present.
Current transformer is attached to one of the three phases, and the ac power supply provided by openenergymonitor is connected.
And you have an emonPi?
But not an emonTx as well?
About the ac power supply:
Do you mean the 5 V USB d.c. power supply
or the 9 V a.c. adapter?
You must have a 5 V d.c. power supply to power your emonPi.
You can also have an a.c. adapter. If you don’t have an a.c. adapter, your emonPi assumes the mains voltage is 120 V, and estimates the apparent power (VA) from that.
If you have both, the 9 V a.c. adapter measures the true voltage, and the relationship between the voltage and current waves, and tells you the real power (watts) being measured.
I currently have an emonPi, but would not be opposed to buying an emonTx if it is necessary.
I have the 5V USB DC power supply plugged in, as well as the 9V AC adapter.
vrms is giving me a value of about 215. As you instructed I edited the configuration to work with the new ct I have connected. Should the true power I am now getting be somewhat accurate?
I may have misunderstood, but I thought that you said connecting the AC adapter would saturate the transformer core?
Which a.c. adapter do you have? And where is it connected? I get the impression that you have connected it Line - Line at 208 V. Is this correct?
The one from the OEM shop http://openenergymonitor.org/files/datasheet/USA77DA-10-09.pdf
Is only rated at 120 V a.c. I have no data for the maximum permissible voltage, and I have not had the opportunity to test this variant (only the UK one) but based on that, I can be reasonably confident that the US adapter will be seriously overloaded if it connected to a voltage significantly greater than 126 V (120 V +5%), and will not give the required faithful representation of the primary voltage.
I have the a.c.from the OEM shop, the same that you just linked. It is connected to the 9V a.c. plug on the emonPi, and plugged in to a 120V outlet via the extension cord you see in the picture.
I understand now. What worried me was your line:
from which I thought that there was a possibility that you’d connected it to the 208 V supply.
As you have it, that’s OK - what I obviously didn’t know was you’d adjusted the calibration so that the emonPi showed the Line - Line voltage of 215 V when it was actually measuring 120 V.
So, as you have it connected Line-Neutral, you need to have the c.t. on the same Line. When you do that, it will correctly read the true average power being delivered to your machine (subject obviously to getting the calibration of both voltage and current right).
If the current and voltage belong to different phases, then the phase shift between voltage and current will make the calibration both wrong and variable, because the number that you get out for power will be wrong by different proportion according to the power factor of the load.
The only adjustments I have made are the ones shown in post 15, so I am not really sure why I am getting 215 V.
This is where I get confused, as I am not even sure if the ct is connected to a line or a load. The power travels through the factory into an electric panel, and then out the electric panel to the machine. I have the ct connected to one of the phase wires traveling form the electric panel to the machine, so would that not be considered a load wire? If so how would I get the c.t. on the same line?
I’m using “Line” to mean one of the three phases, i.e. not the neutral. In that sense, it’s “line” all the way into and around inside the machine. (This is probably where you and I are divided by a common language! 
)
As you’re measuring current between the panel and the machine, you’re clearly measuring the current drawn by the machine.
Does rephrasing help: You need to connect the a.c. adapter Phase - neutral, using the same phase that the c.t. is on.
I can’t at the moment figure out where your 215 V is coming from. It’s in the direction of you not having set “USA” true in the emonPi, but the numbers aren’t really close enough for me to say with certainty that’s the problem.
In any case, for your purposes I think you need to adjust the voltage calibration so that it reads 360 V when the actual phase-neutral voltage is 120. That way, the emonPi will read the correct value for the total power across three phases (of course I assume a balanced 3φ load again).
There is no phase - neutral because there is no net neutral current.
How would I do this?
But there is still a neutral wire - you already have one side of your a.c. adapter connected to it - I know that because you tell me it is seeing 120 V - one leg is connected to one of the phases, the other leg to neutral.
All we need is the phase that you’re using for the a.c. adapter should be the same one that has the c.t. on it.
The same place where you corrected the powers, in emonHub.conf (except of course you change the scales = ... for the voltage (they’re all in order left-right).
Changing the voltage there won’t affect the power values (because the sums to calculate power from current and voltage have already been done, before the numbers get as far as emonHub).
I see what you are saying now. I am working on getting a schematic of the factory to try and determine which phase the outlet I am using for the a.c. adapter is connected to. Do you know of a simpler way to make this determination?
Yes - measure the voltage between the “hot” (your terminology) where the a.c. adapter is connected and each of the three phases where your machine is connected. You’ll measure 208 V for the ones the a.c. adapter is not on, and close to zero for the one that the a.c. adapter is on. If you measure 240 V instead of zero, it means it’s on the other leg of a centre-tapped 120 - 0 - 120 V supply.
Unfortunately I am not going to be able to connect the a.c. adapter to the same phase of the ct, so apparent power will have to do. Is it necessary to change the value of vrms to get an accurate measurement?
Oh dear. By default, the emonPi calculates apparent power also, but only transmits real power into emonHub (and thence into emonCMS).
There’s a way out of this, but it involves reprogramming the Atmel AVR chip inside the ‘emon’ part of the emonPi, which you need to do via the “Raspberry Pi” part of the emonPi. Then when you’ve done that, you must remember never to update the emonPi as an emonPi (which will cause your modified sketch to be overwritten back to our standard version), but update it instead as an emonBase.
Before I spend a lot of time with the details, would you be prepared to edit the sketch in the Raspberry Pi using a text editor there, then compile it, then use another software tool to load it into the Atmel chip?
(I thought this procedure was documented in ‘Resources’, but I can’t find it.)
I gave details of the procedure here but obviously you want to do this with slightly different variables.
That does not seem like something I’d be ready to do. At the moment I am able to use the emonPi to find the run time of a machine, which is what I am most concerned with at the moment. I think it would be very useful to be able to calculate power usage as well, but I am probably not skilled enough to mess with the AVR chip. If I find I want to move forward with this at a latter date I will reach out again, but thank you for all the help thus far.
In that case, the best I can offer is to take the a.c. adapter (the 9 V a.c.) that measures the voltage off the emonPi, and let it calculate the apparent power based on your nominal voltage. As long as you understand the limitations of that, you could adjust the voltage calibration to give a “best guess” for the overall power of the machine, based on the ‘typical’ voltage and power factor.
If you leave the a.c. adapter connected, then it will be using a voltage reference that is shifted by 120° or 240° from the current, which will introduce an error of around a half, but which will vary significantly with the power factor. (That is: if the power factor is exactly unity, the power is exactly half what it should be (cos 120° = -0.5) . If the power factor is not 1, then the 120° gains an apparent shift - I’m dangerously assuming pure sine waves here - that makes the factor greater or less than 0.5)