OpenEnergyMonitor Community

Just a few questions


I’ve decided I want to get into the realm of gas and electricity monitoring.

I’ve been looking at an all-in-one solution for some time and initially came across the Saveometer. Whilst the Saveometer fits the pre-requisite of being an all-in-one solution it isn’t really for me given I can’t get the data off of the device. There are cloud solutions out there but want to stay well clear of them because of the inevitable bricks that the devices will become and thus have been searching for something that will meet my needs and, well, here I am. I’ll be honest I didn’t quite understand the respective parts of the emonXX range at first but I’ve stuck with it and I’m fairly confident Open Energy Monitor will fit my needs.

A little back ground on my meters:

  • Gas - Elster BK-G4M
  • Electricity - Landis + Gyr E110

This is where I think I am at the second in terms of the items I’m lining myself up to buy:

  • 2x emonTH
  • RFM69Pi
  • Optical Utility Meter LED Pulse Sensor
  • Reed switch

But I have a few questions and concerns that I’ll talk through below.

  1. I’m still not too settled on whether or not to use the CT sensor over the LED Pulse Sensor for the electricity monitoring. There’s a few pros/cons for both that I have been able to glean from various sources here and on the web.

I understand that the emonTH can’t be used for CT sensing however I feel easier that a cheaper device is placed within an accessible area. It’s not that I think it’ll go walkies but £40 is better to lose than £60 in my opinion (£80 for two emonTh vs £100 one emonTx and one emonTh since ultimately there’ll be two monitors placed outside). Both meters are external.

Additionally on the emonTx shop page I noticed this:

The EmonTx can be optionally powered by 3x AA batteries. Select the 3x AA battery holder option above if you wish to use this option. In order to preserve battery life the EmonTx needs to go to sleep between readings, we upload our discrete sampling firmware to the EmonTx to achieve this rather than the standard continuous sampling firmware. The discrete sampling firmware is less accurate than the continuous sampling firmware so its worth only selecting this option if you need to power the EmonTx from battery power.

This is exactly how I’d be powering the monitors for my energy meters since there’s no accesible power source nearby. The electricity box is just the meter, master switch and tails into the distribution board inside.

  1. I run a home lab and have a rack mounted server thus can host my own virtual machines without the requirement of any additional hardware like a Raspberry Pi. My OS of choice is Ubuntu Server and RDBMS of choice is PostgreSQL.

These are my questions to the above thoughts/concerns:

1.1 If my gas monitoring is using a pulse sensor why should I care about using a CT sensor for electricity monitoring?

1.2 Given that the emonTx is less accurate in my use case, will using an emonTH with a pulse sensor be more accurate over using an emonTx with a CT sensor and powered by batteries?

2.1 I’d rather not have a separate Raspberry Pi running just to receive data from the monitors, as I have existing infrastructure I’d like to utilise, is there any way that I could consume the data without the need for a Raspberry Pi? I was potentially thinking by using a USB stick of some sort.

2.2 The github page outlines that MariaDB or MySQL is listed as a requirement however within the issue history there’s a post about making the database agnostic, can I use PostgreSQL as the database backend?

2.3 If emoncms itself cannot use PostgreSQL is there any other way to consume the data from the monitors that I could potentially log into the likes of Zabbix, InfluxDB etc?

I think that’s all of the questions I had, I’d appreciate any feedback and insight anyone can provide, thanks!

Welcome, @NotVeryOriginal to the OEM forum.

That’s correct. And the emonTH is normally battery powered, so should be OK given that you won’t have mains power in an external box, nor will it be easy to get even a low voltage cable into there.

I take it you don’t have any PV, or other source of generation? Because if you have, an optical pulse cannot measure exported energy.

You don’t have a choice with gas. With electricity, c.t’s give you power every 10 s (our default rate) from which it’s easy to accumulate the corresponding energy. If you count pulses, it’s very difficult to differentiate energy to get back to power in a meaningful way.

It’s horses for courses. I think I might have answered it immediately above. You could have the best of both worlds - calibrate the c.t’s using the pulse count.

In practical terms - no. As it stands, the emonTH has no on-board data storage at all, and no provision for anything like a USB stick or SD card. All it does is gather the data and transmit it by radio. You need an emonBase (a RPi with a radio receiver) or something equivalent to receive the data transmitted by the emonTHs and launch it on your LAN. You can then have an instance of emonCMS running on your other server. (Or if you did use a RPi as an emonBase, you can set up emoncms identically on both, so each backs the other up.)
As a hardware development project, a battery-powered data logger with an SD card should be practical - but then the data is historical. With the emonTH it’s as old or as young as the last pulse.

I’ll duck out of the database questions - it’s not my area of competence.

Thank you for the welcome, @Robert.Wall.

I don’t have PV or any other external source of generation, no.

I don’t think I explained myself too well on reflection.

Whilst I have a Raspberry Pi to hand and I understand that I need the RFM69Pi hat what I am trying to achieve is removing the requirement of needing a RPi given that I have existing infrastructure that I want to use. Is there any way of removing the RPi element of it and instead having the receiver in USB format? That way I could plug the USB receiver directly into the server and expose it to the emonCMS instance without the requirement of needing an external Raspberry Pi.

I suppose my question is - are there any USB receivers that can replace a RFM69Pi or is there a USB version of the RFM69 module available for purchase and will work with the emonTH/emonTx?

Ah, not what I understood.

Not that I’m aware of.

There’s a degree of processing taking place inside the processor on the RFM69Pi module that turns the data stream into text (bytes in decimal format: e.g. OK 23 121 0 0 0 208 2 25 0 244 31 0 0 (-45) ) so you’d need that plus the USB conversion. You should be able to connect the RFM69Pi to a serial-to-USB converter to achieve what you want, but I don’t immediately recollect anybody mentioning they’ve done that.

IIRC, JeeLabs makes one that will do the job. (JeeLink)

See this thread for more info:

Where to buy:

1 Like

Thank you @Robert.Wall, @Bill.Thomson for your responses.

@Robert.Wall, your suggestion for a serial to usb converter had me scratching my head. All I could think of was the old RS232 usb to serial converters you would use to configure a Cisco network switch. Was wondering how I’d connect one to the other! I started to take a look around amazon and found these, which in my totally inexperienced judgement seem like they’d fit the bill when compared to the wiring diagram for the RFM69Pi:

I did stumble across this post and a member had created this which was also a consideration.

Whilst as feasilbe as they seem, I was looking a more off the shelf solution and @Bill.Thomson seems to have met that. Unfortunately the link Bill provided would have incurred import charges as I’m in the UK. I did however find a European seller and for the item plus delivery you’re talking roughly £30 all in, the difference in price is nothing to write home about between it and the usb to serial converter + the RFM69Pi hat. Taking a more indepth look it seems that the one I just linked to was referenced in the post you linked @Bill.Thomson

I’m still on the fence about using an emonTx with a CT & Pulse sensor over an emonTH with a Pulse sensor and with that I have one final question as I can’t seem to find it, I see than an emonTH reports over 4 years. What sort of battery life can I expect, within reason, to see with an emonTx with two sensors?

Thank you for your help thus far.

I have no idea, but it will be very much shorter than an emonTH, for the reason that the emonTH has been heavily optimised to reduce the energy use to the absolute minimum. Off the top of my head, I think you are talking weeks, possibly months, rather than years.

I take it that you can’t get access to the meter tails as they enter your consumer unit?

Nope, the two tails enter the property from an external meter box where they share the same conduit. The conduit then is affixed in to the consumer unit via a compression gland. Without opening the consumer unit there’s no way to affix the clamp to any of the tails.

I was thinking about this last night though. Please correct me if I’m wrong. My understanding so far is that the CT clamp is to gather the current usage in watts, and then calculating total usage from that, whereas the pulse sensor is relying on the meter on providing this metric via a delayed means as the former is real time or delayed by 10s due to running off batteries whereas the latter is finger in the sky figures because usage could be so low that the meter doesn’t pulse for quite some time.

To save on battery power even further couldn’t I update the firmware to poll the CT clamp once every 30 seconds to gather the current usage at that particular time and then rely on the pulse to gather running usage?

Still trying to weigh up everything, i.e. is it worth the extra work/money for an emonTx to be able to gather real time, or near to real time, uage.

One other thing that hasn’t been mentioned so far (unless I missed it) is the fact that battery operation
won’t yield a real power, measurement. The results will be an approximation of apparent power in Volt Amps
vice real power in Watts.

As Robert mentioned, deriving a power measurement from the pulse count is not easy,
but pulse counting will yield a correct energy measurement. (kWh)
The meter is “correct” even if it is incorrect, as that’s what the energy supplier uses to calculate your bill.

… because it has to make a guess at the voltage - which varies all the time, often by quite a lot.

The last 24 hours for me:


1 Like

Definitely not steady!

Mine looks like this:

If you’re lucky enough to live a quarter of a mile from PoCo’s voltage regulators, it looks like this: