Our current transformers detecting pv generation and grid flow are in a busy consumer unit - nowhere else to go!
I believe crosstalk from adjacent conductors is leading to innacurate reporting
eg GE report says we used .26 units yesterday (including .04 to battery??) vs .93 according to smart meter monitor
I'm not losing sleep over this but is there any posibility of sheilding current transfomers
mgjohn There have been various posts regarding this issue, which I think is really the discrepancy between your smart meter used to bill you and the Givenergy meter used to control the output of the inverter. The EM115 is rated for consumer billing but its primary purpose in the installation is to make sure the inverter output matches your household demand as closely as possible. Obviously, once PV exceeds battery and house demand, the surplus can be exported.
All meters have bands outside of which their accuracy declines and the EM115 is no different, being less accurate at the low values you are experiencing (Hooray!). The smart meter will be as well but that's the one that your supplier is using to bill you and unless you've good reason to suspect its accuracy, is taken as the benchmark.
I accept that over time, the readings on the two devices diverge. I look to my smart meter for "actual" import and export over time and use the Givenergy inverter "Meter" values for information only; usually the time at which export started or ended, or roughly the value. Trying to square the circle with these two devices is wattless energy. Oh maybe that's the cause of the error?
I have wondered the same thing, myself.
In theory the CT should only register the magnetic field from the conductor within the sensing coil (external magnetic fields should be cancelled out), but I suspect that conductors which are very close (and carrying current) could cause interference,
This site (https://ctlsys.com/support/current_transformer_crosstalk/) seems to confirm this theory, but I would be interested to hear from anyone who has moved the CTs, and seen an improvement in the reported system data.
You could if you so wish purchase a solid CT core for it.
Replace the clipped one. Much more accurate, but required removal of the live wire your monitoring to thread through the hole.
At those low readings the error will be more obvious.
Personally I wouldnt other.
Solid core current transformers by the way they are constructed don't respond to currents on the outside, unless the magnetic field produced is sufficient to saturate the core (in which case you'd probably be calling the fire brigade). The split core type commonly installed with the PV meters is somewhat less immune to external fields, because the windings are only on one half of the core and because the joint between the two halves isn't as good as continuous material.
You can get solid core CTs for the meters and they have half the error. You could also re-site the CTs so they are on the conductors outside the consumer unit, where there is more room.
The other thing is that, they aren't perfect and it might be that those differences are as good as it's going to get. Mine don't match up perfectly either.
oops, didn't see the other posts. 😀
My mains CT clamp is in a pretty good position (away from interference) and the GE reports consistently show grid import figures around one tenth of the smart meter readings. Like you my grid import is very low and it is just a fact of life that such small figures are not reported accurately. Don't worry about it and just use the smart meter reading as the one to go by.
Thanks for all the comments, the example was indeed very low usage on that day.but Woodstock's observations re GE reports match my experience.
Monitoring earlier in the year showed ~ 100% inaccuracy which I put down to CT positioning.
date eon ge deviation
11-Feb 6.56 5.81 113%
12-Feb 13.87 13.09 106%
13-Feb 11.13 10.40 107%
14-Feb 9.85 9.09 108%
15-Feb 10.33 9.50 109%
16-Feb 5.85 5.03 116%
17-Feb 5.27 4.56 116%
18-Feb 7.49 6.70 112%
19-Feb 20.89 20.09 104%
20-Feb 7.68 6.90 111%
21-Feb 11.58 10.82 107%
22-Feb 4.44 3.71 120%
23-Feb 5.16 4.39 118%
24-Feb 10.32 9.49 109%
25-Feb 3.67 2.89 127%
26-Feb 11.99 11.22 107%
27-Feb 2.78 2.00 139%
mgjohn Yes, the lower the figures the greater the inaccuracy. A typical weekly total for me would be 2.0 0.2 1000%.
These are my readings for the last week or so. I'm not sure whether the issue is the CT position or something else.
Date S/M / GE/M S/M=Smart meter grid import
GE/M = Givenergy meter grid import
29/07/2022 2.19 / 2.03
30/07/2022 4.85 / 4.48
31/07/2022 3.32 / 3.38
01/08/2022 0.27 / 0.13
02/08/2022 0.36 / 0.16
03/08/2022 0.46 / 0.22
04/08/2022 0.3 / 0.15
05/08/2022 0.54 / 0.23
06/08/2022 0.18 / 0.05
07/08/2022 0.17 / 0.07
I agree that the lower import values have a greater difference than when the import is higher. Higher values with less differences are seen 29th to 31st July. After that the import was reduced by a factor of 10 and the difference increased by 50 to 80% at a guess. But as pointed out these grid import values are not very high- so not too much to worry about.
Another issue I have asked my installers is whether placing 2 CT near to each other may cause interference.
I had an Apollo Gem immersion heater controller which used a CT to detect grid export - turning on the immersion heater when exporting. Now I have a battery this device is less important but still can be useful to even out the immersion tank thermal turbulence produced by my solar thermal system. When this is active, cooler water can find itself being pushed upwards by the solar heated water at the bottom of the tank. Before I had the battery the Apollo gem CT would detect grid export and turn on the immersion heater - thus heating this cooler water.
Since the battery installation there are now 2 CTs and the Apollo Gem seems to get confused and starts to turn on in the evening (no PV) and gets energy from the battery (this is OK at this time of year but not when PV generation is reduced in the shorter days). My remedy is to remove the Apollo Gem CT from its conductor in the evening.
Apparently Givenergy are working on there their own water heating device.
Any comments would be welcome to find a better solution to one I use at the moment
AlanStevens Apollo Gem seems to get confused and starts to turn on in the evening (no PV)
I'm guessing that when your kettle switches off, resulting in a temporary grid export while the inverter ramps down, the Apollo detects this and starts heating water. If it's possible to either increase the export value at which it switches on, or set a time delay, or both, it might prevent this behaviour. Once it grabs the energy, the inverter may see it as house load and just feed it.
I assume your solar thermal system is a coil near the bottom of the tank and your immersion is a short one at the top? If so and you have a fair amount of export, you could replace the short immersion with a longer one which would heat water from the bottom to the top of the tank. You'd get exactly the same amount of energy put into the tank from the diverted export but it would result in a larger volume of water being heated to a lower temperature.