OpenEnergyMonitor Community

Unreliable Power Supply

I’m looking at setting up a home power monitor to monitor two separate single phase circuits supplying my house. The most practical location is the power meter box on the property boundary, 40 m from the house. I have wifi available there, and a spare double GPO.
I understand that an emonTX + base system is not suitable due to the limitations of the TX radio over the 40m distance. An emonTX + wifi would be good, but seems to be unavailable at the moment, unfortunately.
I can use an emonPi in the power meter box, connecting to my network via wifi. However, I can count on a few unscheduled blackouts every year, and I understand the emonPi doesn’t like unscheduled loss of power. The blackouts can last from a few seconds to many hours, sometimes days. Although I use a UPS to support house comms and server, a UPS at the box is not feasible. What would be a good strategy for recovering the emonPi after a blackout? Having a backup SD card available? Loss of a little data would not be an issue.
Many thanks in advance.

Can you not connect the CTs to where the cable from the power meter box enters the property and meets the distribution box?

If not, I’d use a Pi ZeroW, PiJuice UPS hat and an EmonTX. Connect the PiZero via the serial/UART to the EmonTX.

TBH, I have not suffered any SD Card issues due to power outage. I believe that any associated power spikes are the cause of corrupt SD Cards - all my Pis are behind surge protectors.

If you have a clear line of sight between the two aerials, I’d question that. The claimed range is 100 m, and we (i.e. JeeLib, the library we use) run the radios at reduced power. If the power were to be turned up to maximum, I suspect the emonTx would reach the house - but whether it could get in would depend on the material and thickness of the walls. It’s a much lower frequency than Wi-Fi, so it should stand a better chance. If mains power is available, then you’re not worried about battery life.

Thanks for your helpful answers. Your comments are much appreciated. I’ll post my responses for the sake of anyone reading later. Firstly:

Physically not practical. Power/meter box is on the property boundary and has easily accessible power cables for CTs plus spare GPOs for power supplies. Ideal, except it’s a metal box. The house distribution board is 40m away, on a wall of the house. Space is cramped for CTs and there are no GPOs, so not really worth messing about with.

Nice idea, but the kit has to be installed inside the metal power/meter box for protection from the elements. We live in the tropics, and torrential tropical rain is something to behold! I would be using a remote wifi aerial, mounted on the outside of the metal power/meter box to ensure comms. I can waterproof that simple penetration OK. Although it’s been done by others, adding an external/remote aerial to the the PiZeroW isn’t a practical option for me. After successfully repairing my Carver M400 stereo amplifier last year, I’ve lost my enthusiasm for fiddly soldering jobs! A weatherproof case to allow the kit to be mounted outside the metal box would not really be a practical option either, with all the penetrations required for this system.


Nice thought. It’s 40 m line of sight but there is a tree and a single blockwork wall with a window in it between transmitter and receiver. It might work, but what if it doesn’t? What concerns me is that I can’t see a simple practical Plan B with this option. With wifi, if I have problems with the signal, I can easily add a remote access point under the eaves of the house. I know that will work, if required.

However, a very useful comment was

That gives me confidence that the emonPi will be ok when the power drops out. I guess, if the worst happens, I can have a backup SD card handy to rebuild the system.

So I’ve decided to go with an emonPi mounted inside the metal box, with a remote aerial connecting via wifi to my house network. With surge protectors on the power supplies!

Many thanks again for your comments.

Why? The EmonPi is just a Pi. The PiZero is no different from a Wi-Fi perspective.

You could always use a powerline adapter.

:+1: When I read the new details of the installation, I thought that too. I think that would be my preferred option.

How are you going to get the external WiFi antenna connected to your emonPi? The external antenna on the emonPi is for the 433 MHz ISM band radio used to communicate between it and its emonTx’s and emonTHs. The Wi-Fi antenna is ON the RPi itself - that’s the reason for the acrylic end-plate on the emonPi, it lets the Wi-Fi out.

Not that it’s actually doable with a Pi, but I can see why he said that.
An RF transmitter/receiver isn’t going to work too well if it’s located inside a metal box.

When Adrian wrote:

I think he was thinking the antenna on the emonPi is Wi-Fi: It isn’t.

Quite right, but in the end, it doesn’t really matter WRT to the frequency the xmtr/rcvr is operating on,
it’s still not going to work too well from the inside of a metal enclosure.

That said, I edited my post…

This is the line that I picked up on.

That gave me the impression that Adrian would uncouple the “Wi-Fi antenna” (as he thought, but it isn’t) and use a weatherproof Wi-Fi antenna on the outside of the box, drilled through and sealed, and cabled into the r.f. connector on the emonPi.


I changed my post to indicate that any RF source/receiver isn’t going to work in a metal enclosure.

Thanks for all the helpful tips and comments. You’ve understood my proposed installation and you spotted my incorrect assumption!

I had assumed that the aerial in the pictures was a wifi aerial. Now I know it’s not. Thanks for saving me. I can see me scratching my head, saying “why isn’t this working”! Open Energy have suggested using a wifi dongle, and the suggestion …

has merit as well. So it looks like there’s a a number of ways forward. I’ll do some more due diligence, then get the project moving again, hopefully today.

Many thanks again.
Kind regards everybody

Presumably you’d need to put it in a sealed plastic box and deal with cable entries to both boxes.

How does rain affect the Wi-Fi path between the meter box and the house?

Not really. There’s plenty of room inside the meter box. I’ll just have to find an aerial with a detachable antenna, and put the antenna on the outside of the box, as per the previous plan. The only entry point to the box will be where the aerial wire comes in. Now to find the right part!

Don’t know really, but we have wireless broadband where we live and it can be rain affected. It varies, from no effect to completely blocking the signal (rarely). Still, rain permanently damaging physical components is of much more concern. Temporary data transmission losses are of no concern. It only rains heavily in the wet season Dec to March. Last year we once had 430mm of rain overnight, and 700 mm over 3 consecutive days!

Powerline adaptors are a good idea that I hadn’t thought of. I’ll keep them as a backup plan though, for a couple of reasons. The power box/pole being out on the street makes it more prone to lightning strikes, and a physical connection to active gear inside the house is a non-preferred option. I also don’t know how the powerline adaptors would like surge diverters. Then there’s the stacking of equipment at the GPO; surge protection, pass-through powerline adaptor, Euro-to-AUS travel adaptor, AC-AC adaptor. Big stack! Also my network switch is at capacity in the office, so I’d have to buy a new one. Little costs, but they add up. So I’ll try wifi first.

You definitely need a power surge protector on the Pi then (AC & DC). Ethernet surge protectors are pretty cheap.

I wonder if there is a path from the CT into the EmonPi? @Robert.Wall? @Bill.Thomson?

There’s next to no input protection on the emonTx or emonPi - your main hope is the c.t. will saturate on the current spike. But even the fast edge is likely to cause damage. And if the c.t. flashes over from primary to secondary, then it’s very unlikely that anything would survive.

“the fast edge” - lightning strikes have incredibly fast rise times, even a 90° bend in a lightning conductor has been known to have enough inductance to stop the current and cause it to jump out of the copper onto the adjacent wet masonry. And of course, transformers work by rate-of-change of current.

1 Like

In general, they don’t.
Connecting one to a surge protected power point usually results in poor or no operation.


I was thinking of adding a surge protector to the RJ45/ethernet connection, possibly at both ends in this case.

That may or may not cause the same effect as connecting protection to the power side of the link.
The anti-surge devices often have filters in them that kill off the high frequencies in the signal.
(i.e. a low-pass filter) The higher the frequency, the more it gets attenuated.
Whether or not that would be enough to affect performance, I can’t say.