Electricity Measurement at Oakhouse 2025-01-23

Having installed a new control unit for the heat-pump, we needed to measure electricity consumption. As part of the heatpump project a CT reader interface was added to the TM4C project, this revealed that when idle the heatpump appeared to be consuming about 270W. There were was no apparent mechanical activity such as pumps, motors or solenoid valves that might explain this, consumption by the control unit should be much lower at more like 25W. As the CT reader involved a home made board and untested software, this naturally came under suspicion. A couple of CT meters were borrowed from TECs, one of these was installed in parallel with the existing CT, at first this read 140W, but when it was installed in place of the existing CT, it also read 300W. I then tried a Tenma mini clamp meter This read about 1.1A. Perhaps the measuring equipment was accurate?

Even if current measurement was reasonably accurate all of these measurement devices only read RMS current, then derived power from this. As the heat pump contains reactive components, these measurements didn't give a full picture, so I looked further into metering.

Whilst I had all this equipment and the cover removed from the consumer unit I decided to measure other circuits as well, the easiest way to get a reasonably accurate picture was to use the Tenma meter on each circuit. The top 5 circuits were:

These readings have demonstrated that there isn't a serious error in the CT reading software, but there may be other errors in trying to read low currents with current clamp type technology.

It is also apparent that the consumer unit could do with a bit of rearrangement to enable CT's to be routinely fitted to more circuits; lighting and associated 12V systems should all be separated from ring mains; other high consumer on ring mains should be measured separately on a permanent basis.

More detailed investigations of high use positions

Heatpump

Circuit 9 - Sitting, Kitchen, Stairwell, Study

All these items plug into sockets so can also be measured using measuring plugs. The circuit as a whole was measured using a clamp meter, and the consumption of each device-width was calculated from the difference of two configurations, the results at 0.01A difference are not that accurate. Measurement with a measuring plug would eliminate other circuits, and could produce a profile for each device in use, which would also be useful for the consumption forecast.

Circuit 21 - Lean to, Utility, Dining room, Cloakroom

This circuit includes the lights at position A. An improvement would be to remove the relay for battery charger A (no longer used), and replace it with a meter, so that the lighting circuit could be permanently metered. As the control systems can all run off 5V, the ATX psu at A could be turned on by software only when 12V is required for lighting.

Other appliances on this circuit have plugs, so can be measured using a wifi plug.

Establishing consumption of things with batteries such as laptops and chargers for battery powered appliances needs to be done with care, to ensure the true standby is separated from charging.

AV cabinet - Circuit 8

Unlike the previous circuits, the AV cabinet was measured using the Efergy CT meter. So what is measured is the power to the AV cabinet a a whole, consumption of specific devices then calculated from the difference in readings.

The readings start by taking a power reading for the cabinet in its normal configuration, and then unplugging everything. As the cabinet is powered through a UPS, the UPS can be unplugged, and the router will remain connected, but when it is plugged in again time must be allowed for it to recharge.

Should go back and remeasure the server and 300W PSU separately. The PSU is now 12 years old consideration should be given to using an ATX supply which could power almost everything in the cabinet. The server is quite a complex unit, as it also includes several disk drives, including a 1.8Tb RAID array - does it require a 12V supply? Could it be replaced by something running off 5V? Could a 1Tb M2 SSD suffice in this server and the 1.8Tb array be moved to the desktop?

Microwave

This is a case of a choice of a heirarchy of actions:

1. Change behaviour to turn the microwave off at the wall switch except whilst it is in use
2. Determine if it is faulty, and if so attempt a repair
3. Buy a replacement which is compliant with EU 1275/2008
4. The whole kitchen is quite old get a new kitchen

The least cost both in carbon and cost terms is to switch it off, but lets suppose that this behaviour change isn't acceptable. Then there is a choice between repair and replace, which could be made rationally using a payback calculation both in carbon and financial terms. Incidentally this also applies to replacing the whole kitchen, but the carbon and financial cost will almost certainly be higher.

Carbon footprint of a replacement microwave

This could come from several places in aascending order of accuracy:

1. ACT CFT general factor for goods applied to new cost of microwave
2. TECs E-Pack factor for category???? applied to new cost of microwave
3. Determine COICOP code for a microwave from ONS Expenditure codes (look for microwave in Part 1 tab) then look up COICOP code in ghg_coicop tab of UK's Carbon Footprint - Full Dataset including SIC code.ods multiply latest factor by cost of microwave.
4. Find weight of microwave from specification and apply appropriate factor from appropriate table in Factors for GHG reporting
5. Get product life cycle assessment from manufacturer.

Payback calculation

Placeholder for payback calculation.

Needs to be like TECs one but provide several options.

Email to NEFF

I wrote to Neff, initially to enquire if a current of 0.23A whilst idle was expected, here is the correspondence ... need a tool to convert email to html page

Repair and Replace links:

• Microwaves
• How a Microwave works
• Ransom Spares - supports of DIY repairs?
• Article about Vampire Appliances
• How bad are microwaves for the environment

Reorganisation of the study computer supply

Different devices here have different requirements:

• Desktop PC and screens are switched on during the day and off at night
• The Chromebook, second Ethernet hub, power supplies for single board development are used occasionally, also the Chromebook needs charging from time to time, ideally at night. (conflicting with the desktop)
• The RPi 4 datalogger needs to be running 24/7 and has a low power requirement.

Some fairly simple rearrangement has been done:

• The desktop, screens and one 5 port ethernet hub are connected via a TPLink Kasa switchable energy monitoring plug, which has been reliable for a number of years (unlike some other plugs). This is a significant consumer, which could be part solar powered (in summer).
• The RPi 4 data logger uses relatively low power and runs 24/7. This has been moved to its own dedicated plug and ethernet socket, so that it is now completely independent of other computer equipment.
• The chromebook, and provision for ad-hoc ethernet connections has been moved to a 4 socket extension lead, so that the chromebook can be charged at night, and any single boards under development can be plugged in.
• The printer and TV are connected to different sockets, and are normally left switched off.

Reorganisation of Consumer unit and Lighting Circuits

• There are 3 blanks in positions 1 to 3 of the top row of the consumer unit
• Some of the highest consuming circuits (7, 8, 9) are clustered together
• Each light circuit has a separate circuit, with a relay feeding a 13A socket, originally used by a battery charger, but now batteries have been replaced by ATX power supplies which give a permanent 5V supply for electronics, with a 12V supply switched on when needed, so relays are now longer needed. Metering of use by ATX supplies would be useful though.

Direct metering of the Heat Pump

The biggest consumer by some way is the heat pump, which I am trying to tune for better performance. On a typical january day 25% of its energy use cannot be accounted for by heating, this will get worse as the heating requirement drops into spring.

It has previously been noted that direct metering could measure the power use in greater detail and more accurately. As a MODBUS interface has already been developed

Possible meters

• Eastron SDM120 Modbus registers: Eastron SDM120M Modbus Registers
• Eastron SDM630M Modbus Registers: Eastron 630M Modbus registers

Either of these will need the slave ID, and communication parameters setting SDM120 goes to 9600 baud, whereas SDM630 goes to 38400 baud

Metering other positions