Hi, I wonder if you could help me with some advice. I am looking to improve my energy monitoring system. I have used the OWL intuition system before, and am currently running two Energenie whole house monitors - one in my Shed monitoring my PV output, and one at my main consumer unit monitoring the whole house usage. Unfortunately the OWL PV system cannot deal with the remote location of my PV system (connected into a sub dis board and remote from the consumer unit), and the Energenie system does not give the accuracy or CMS flexibility I would like. It looks like the Emon system will do what I need, however I cannot decide on the best topology for the system. As background I have two spare Raspi 3 B+ available and my options seem to be to have EITHER an EmonPI running at the consumer unit with an Emontx in my shed monitoring the PV output OR to have a Emonbase located centrally near my router/homeserver and two EmonTXs, one at the consumer unit and one at the PV system. I also cannot decide whether its better to byuy the fully assembled units or try and use my existing spare PIâs⊠Any advice would be appreciated.
Thanks,
Ian
Hi Ian, welcome.
Do you mean measuring house usage or measuring Grid Import? How does the PV get connected to the house?
Either will option do. Iâd probably go for 2xTX. One other option is to use the ESP unit on one TX and connect one of the Pis serially to the other TX. The Pi will receive the data over the LAN then. There is a thread here where I have done exactly that - all you really need to make up is a short cable.
I happily run an EmonTX off an old USB charger so you can save there but do buy the AC/AC adapter that for some inexplicable reason is not listed an an option for the TX and of course a CT for each.
[edit]
Is WiFi to the shed an issue or do you want to use RF?
Just checked the figures and I think 2 x (TX, AC PS, CT) + ESP, is pretty much the same as an emonPi Solar Bundle (with 2 X CT, AC, DC) + Tx + AC PS.
As long as you can get RF between shed and consumer unit emonPi + emonTX is probably the way to go.
Hi Brian, thanks for the quick response, to clarify my setup, I live in an old Victorian house, the consumer unit is in the entrance hall, behind the front door, I can monitor the total supply import/export here. This consumer unit feeds a sub distribution board in the kitchen, About 20m away, an mcb from this board feeds out to my shed (About 30m) which has a further distribution board, an mcb in this board feeds my solar system which is installed on the shed roof. This is where I can measure the solar generation output. Because the house is old with solid walls Wi-Fi and 433MHz signals are not great⊠therefore I will probably need to connect the devices via Ethernet (hard-wire). FYI the solar is only rated 1200W hence the convoluted installation. My next question is can I use the standard pv setup on the cms system with this configuration?
Cheers,
Ian
As a fellow retired electrical engineer, Iâd largely go along with Brianâs suggestion - but which a slight note of caution: are you likely to want to expand they system beyond measuring the PV output and house consumption? If thatâs likely, then while the Emon system is flexible (to the point of confusion, sometimes), nowâs the best time to think about that.
If the house is fairly long and solid, then it would be worth checking Wi-Fi coverage now, but it sounds as if a wired Ethernet, or part-wired, solution is going to be necessary.
The ISM band radio that the emonTx comes with would certainly reach the 30 m from the shed to the house if itâs a relatively clear path, but until you try, you wonât know if it will reach all the way to the front of the house. It might, it would have a better chance if you turned the power up, but even so it may not be reliable. Failing that, youâd need an emonBase (RPi plus receiver) in the kitchen area to forward the signal via a wired Ethernet to the RPi at the consumer unit.
Unless of course the emonTx there could reach to the kitchen as well, but from your description, I doubt that it would.
Powerline Ethernet between house and shed might be a possibility?
Assuming this or wireless then what Iâve installed at my sonâs place might work for you.
At the PV inverter: emonTx + RPi + AC/AC voltage sensor + Optical pulse counter. This measures the PV output from the PV smart meter.
At the Grid meter: emonTx + RPi + AC/AC voltage sensor + Optical pulse counter + CT. A CT is necessary because the grid meter will not record export. I find CTâs inaccurate and so with a pulse counter it is possible to calculate a correction factor â pulse when importing vs CT when importing. Pulse is exact.
Each emonTx is has a serial direct connection to its RPi â the simplest/most reliable, in my opinion.
A third RPi just runs the emon software and receives inputs from both emonTx/RPiâs. This might seem overkill but it means historic data is recorded in 2 places and each emonTx/RPi is just doing a simple job.
Having read too many stories about RPi power problems, I only use the official Raspberry Pi psuâs.
Hope this may help.
Normally, yes. But a friend who used to run a farmhouse B&B in rural Scotland, and shared a final distribution transformer with his active farming neighbour, had his blown up by spikes from his neighbourâs milking machine. And the other neighbour on the same transformer had their central heating blown up - not at the same time either.
That depends on how carefully you calibrate them. The problem with pulses is theyâre historic. The best route is to calibrate the inputs as best you can against a decent multimeter, then tweak to get best agreement against the meter pulses in the long term.
Robert âŠ
That almost beats the Covid/5G scare
Thx for all yr help to me in the past and for what you do for the Forum.
Regards
John
But this was as proven as things like that can be. I specified a surge suppressor for him, then drove up there and tidied up a forest of armoured cables, non-armoured cables, dodgy earths to fit it. No more trouble like it. (I did drill his 2ft thick walls and replaced the mains-borne with cable though. But no more TVs got blown up - I didnât mention that before )
As with anything, you get what you pay for.
Like the âhotrodâ shop owner asked his customer, âspeed costs money young man.â
âHow fast do you want to go?â
Viz:
-
Magnelabâs SCT-075H is a revenue grade current sensor meeting ANSI class C57.13 0.6
-
The SCT-075H has a phase angle <= 0.25° and accuracy +/- 0.5% from 10% to 120% of the rated input current
ACTL-0750 Split-Core CT - Continental Control Systems, LLC with thte C0.6 option
With Option C0.6 , the Accu-CT is calibrated and verified to meet IEEE/ANSI C57.13-2008 class 0.6 accuracy and IEC 60044-1 class 0.5 S accuracy and each CT is shipped with a certificate of calibration. When used on 50 Hz services, Option C0.6 must be ordered with Option 50Hz.
- Accuracy: ±0.50% from 1% to 120% of rated primary current
- Phase angle: ±0.25 degrees (15 minutes) from 1% to 120% of rated current; ±0.50 degrees (30 minutes) below 0°C from 1% to 10% of rated current
- IEEE C57.13 accuracy: class 0.6 from 1% to 120% of rated current
- IEC 60044-1 accuracy: class 0.5 and 0.5 S from 1% to 120% of rated current
At 55 and 61 USD respectively, theyâre not cheap. But not outrageous like some of Sentranâs CTs.
Theyâre a hundred bucks a pop.
In that case, you can do either setup (emonPi + TX, or 2xTX).
To hardwire the system, use a Pi(s) serially connected to the (or both) emonTX(s).
Iâd be inclined to use the emonPi Solar Bundle with AC PS (you can choose to use your own 5V DC supply). That bundle comes with 2CTs, you can use the second CT on the emonTX. Also get an emonTX (with AC PS - DC same choice) serially connected to one of your RPis (all you need is to make a short cable or use breadboard leads).
Thanks guys for your advice, being a skinflint, could I just use two emonPi - Raspberry Pi Energy Monitoring Shield Add-on KIT(s) on my existing RASPIâs?
My shopping list would be:
2 x emonPi - Raspberry Pi Energy Monitoring Shield Add-on KIT
2 x CTâs
2 x Plug in VTâs
2 x emonSD - Pre-loaded Raspberry Pi SD card
I already have suitable 5V 2.5A RASPI PSUâs, and RASPI Model 3 B+ (s)
If I install one system in the Shed on the solar AC output, and one at the consumer unit, then connect the two RASPI ethernet ports back to my server/router I assume I can then use Emon CMS to provide the web based GUI - All for around ÂŁ100 - Thoughts?
Yes is the short answer - note there is no case for that shield.
You will need to set it up so that one is a âmasterâ for storing and visualising (though actually you could do it both ways for data replication). Come back when you have them to walk it through.
Do you mean 2x AC/AC sensor to confirm power flow direction?
I often call those v.tâs - voltage transformers - to differentiate from current transformers. Never to be confused with a c.t. that has an internal burden and outputs a voltage, that is STILL a c.t.
The Monitoring Shields look like a good & economical solution.
It appears this is a new product offering â maybe not a new development but the âgutsâ of the old EmonPi unit?? â so a proven component. And no doubt thereâll be a subsequent add to this forum thread which will clarify.
You will have noted the Optical sensor vs CT discussion above. The Monitoring Shield accepts an Optical sensor. Using this at the PV end would cost ÂŁ19 and save the ÂŁ21 for a CT and AC/AC voltage sensor.
Pulse counting is accurate. As dawn breaks and the first few generated watts start to flow, the CUM pulse count recorded by emon every few seconds does not change at every sampling. It changes every nth sample but that quickly changes becoming continuous as the sun rises. And if you are visualising data at say 5 or 10 min intervals then you wonât even see this. But the pulse count will have accurately recorded that tiny first flush of generated power.
Whereas an inaccurate CT which measures instantaneous current (from which instantaneous power is calculated) will be even more inaccurate at the low generation levels at dawn & dusk. All CTâs are inaccurate â itâs just a matter of degree and cost. Furthermore power flow is calculated from instantaneous current & voltage values sampled every circa 10 secs â this of itself is an approximation. And, to cap it all, the AC/AC voltage sensor is inaccurate to some extent.
As an aside â I recently became aware that smart meter clocks drift (my sonâs is 6 mins slow) and so that can complicate precise comparisons between downloaded smart meter 30 min data and emon recorded data. If connected to the internet, emon time is accurate internet time. Also smart meters use GMT throughout the year whereas emon visually presented data switches to BST when appropriate â but this is selectable.
Yes it is the emonPi board on itâs own so no cost of assembly - they are doing a soft launch.
Strictly, GMT no longer exists - replaced by UTC.
Actually by default emonCMS displays data in the Browser Timezone, but yes selectable.
Or go the whole hog and get an optical sensor and hook it up to your Pi and read the flashes with python or Node-Red.
@borpin
Does that mean that at the PV end, you could dispense with the Monitoring Shield and just connect the Optical sensor directly to the RPi?
How do you do that?
Are there any Forum links?
Thx
RPi direct pulse input: Directly connecting to Optical Pulse Counter with RPi? - #2 by pb66
Did that not show up with a search?
Iâd have never found that!
Iâd probably deploy Node-RED, easy to do and then send on the MQTT messages. I have a spare sensor here so Iâll order up a couple of RJ45 mounts and breakout pcbs and see how it goes.
@IanHampton - you might want to consider this - even cheaper!
Thanks again guys, order now placed, I also splurged for the rather nice cases, one with a display for the consumer unit end and a blank Emon pi case at the pv endâŠ
Will update you once I receive the bits and commence installation.
Cheers,
Ian