Hi
I am a new user of Energy Monitor and have read most of the site. The closest thing I have found that discusses an answer to my problem is Martin Roberts plan of Solar PV Power Diversion with emonTx Using a PLL, emonGLCD and Temperature Measurement. (Learn | OpenEnergyMonitor)
However this will not work for me as the Solar Power generation connection to my local power network is 75 yards from the place where my power network is connected to the grid. Also we dont like wifi or radio waves we believe it is harmful to humans so we have an installed an extensive LAN to provide internet access.
Surely the long term general solution is a box (probably a Emon Pi or tx) near the grid connection, a second similar box where the power generation connects to the local power network and a third box near any diverted load to deal with excess power. This last one should also support a battery rather than just hot water.
This would support a general solution to the generation of Solar energy and reducing power consumption from the grid. Martin’s solution is good but does not allow for adequate local communication where the supply, PV generation and load (preferably battery) are physically distant and radio/wifi is ruled out.
Any solutions?
Thanks
Gordon Webb
Welcome, Gordon, to the OEM forum.
Your potential problem with a network-based solution may well be delays between measuring the power and acting on the data. Both MartinR’s and Robin’s diverters react within two cycles to a change in conditions, a delay could lead to the system ‘hunting’. The link between the “decision-maker” and the power switch could well be hard-wired using a telephone-type cable - ‘pulse’ (in the diagram) is the drive current for a LED so cable resistance could be compensated by reducing the value of the current-limiting resistor. That would eliminate one possible source of delay.
If you’re looking to charge batteries from the surplus generation, you probably need an analogue solution rather than burst mode (unless you can control the charger in burst mode, as it’s unlikely to be happy if it receives chopped up mains).
If you do go for a commercial battery solution, there are a couple of threads about using that with one of those diverters: there have been stability problems because it would appear that the installer didn’t know about or didn’t understand how the two systems might interact, and it ended up with each trying to optimise the other.
I have now installed a 3.6kw solar PV installation and it is working fine.
It is located about 100m from the building with the consumer unit.
This building is located another 35m from another building with the grid supply.
I would like to install a PV router like to one designed by Robin Emley.
This would be located in the building with the consumer unit and immersion heater but the CT would need to be in the building 35m away. Is there any discussion of how I could achieve this using hard wiring? I already have a telephone cable between the two buildings in the same duct as the power cable between these two buildings. Could someone point me to an article discussing this and in particular which CTs i could use that can accept this length of cable? Thanks
Take a look at ‘Learn’ → C.T. Sensors.
You should be OK with either a voltage output or a current-output c.t., but I think I’d be happier using a current output c.t. with the burden at the diverter end in this case, because (if it’s capable of generating enough voltage, i.e. not working right on the limit of its rating) the resistance of the wiring will make no difference to the calibration, in each case the impedance of the connection will be low and a good twisted pair cable should be OK. Until recently, we weren’t aware of the presence of a particular capacitor in switched mode power supplies, which causes a problem because it introduces interference if you use the input circuit of the emonTx V2 or V3, but not if you have the buffered bias circuit that Robin uses or the “grounded input” circuit we use in the emonTx4.
If your telephone cable is a 4-pair or something like that, you should be able to steal a pair for the c.t. As long as you don’t have a mechanical dial telephone, you’re unlikely to get interference from dialling out, but the ringing current on an incoming call might be a problem. The power cable is a little further away and could present a problem if there’s an interference-generating appliance being fed by it.
Thanks Robert - that is helpful.
I read the " learn - ct sensors "
It would seem that the resistance of any additional wiring ought to be small in relation to the value for the burden resistor in order not to introduce a significant error.
I have searched for one of the ct sensors used with this device - “Clip-on CT (blue), YHDC SCT-013-000, 100A:50mA without a burden resistor” and found it for sale - seller says burden resistor is 10 ohms.
If this is correct then any wiring i use ought to have a resistance of less than say 100th of this or .1 ohms to minimise any error - telephone cable has a resistance of 100 ohms a km so 1 ohm is only 10 meters and i need 35 meters. if i combine two wires that will half it so I will still have a resistance of say 1.7 ohms or will introduce an error of 17% - or double that for the two ways - does that sound correct? If so I will need to get a bigger wire to make it work accurately?
Sorry but to be clear the telephone wire is spare and unused so i can combine each of the two pairs.
The problem seems to be the small burden resistor - do they have current transformers which will work with larger burden resistors which will also be OK with this device?
Thanks again Gordon
You’re not thinking current sources. The c.t. is a current source and it will generate whatever voltage it needs (within its rating) to drive the current it generates into the burden.
(Hint: turn almost everything you know about voltage sources - the sort you’re most familiar with - on its head and you won’t be far wrong.)
I suspect the seller doesn’t know much electrical engineering. No, the burden value is whatever you need to generate the voltage you want - again within the c.t’s rating. We use 22 Ω in the emonTx V3 for that very c.t. (and the OEM shop still sells them and you can read my test reports in ‘Learn’), and at 100 A this gives 1.1 V rms - 11 mV per amp. You can use a higher value burden at the expense of accuracy, but as you’re balancing to a null, this is less important. It’s phase error that’s affected most as the burden value increases, yet Robin has found that it’s pointless trying to compensate for that because you gain nothing.
If you’re going to copy Robin’s circuit exactly, you should have no trouble with the SCT-013-000, because that’s the one he uses as the main “balance detection” c.t. on the incomer.
Hi Robert
Thanks for your reply. Sorry - yes you are right that as a current transformer the increased resistance should not affect the current and thus the voltage drop over the burden resistor.
However, your site says “The burden value needs to be low enough to prevent CT core saturation.”
Presumably this means the burden value plus the extra resistance of the wire.
I have looked at the documentation for the CT we are discussing for any guidance of the max resistance so as not to cause core saturation - linear operating range for the product - but I cant find any guidance on this in the technical specification for the product - can you help here?
I do understand that the circuit is trying to detect zero so the actual value is not too important but it would be nice to be able to calculate the max resistance i can add without significantly affecting the linearity by staying within the core saturation limits ?
This would give me the distance I could have between the CT and the immersion heater for any particular wire diameter.
Thanks again for your help
If you look at the emonTx V3, the 4th channel is a “high sensitivity” one intended for UK PV infeeds, with a max current of 16 A. The burden for this is 120 Ω. When you use this sort of value, the first casualty is the phase error.
It looks as if your ≈3.5 Ω (I take it your 100 Ω/km is the loop resistance - it seems a bit low?) won’t have a significant effect as long as you’re working reasonably below 100 A. It all comes down to the voltage you want at the input of your ADC/op.amp or whatever, because the closer to zero you can get, the happier the c.t. will be (remember it’s a current source).