Author Archives: Greening Me

A bit of a diversion

I’ve recently replaced the diverter which manages diversion of my surplus solar electricity to immersion heater for hot water and/or electric car charger. The new unit is a Myenergi Eddi. I also needed a Harvi to transmit the output of remote current sensors to the Eddi and an accessory digital input/output board to control my older car charger from the Eddi.

The panel below sits inside my airing cupboard for close proximity to the immersion heater in operation but lifts down as an assembly for maintenance.

Eddi installed in airing cupboard

From top to bottom:

  • Home Energy Management System (HEMS) – left
  • Junction box – right
  • Transmitter to activate my older car charge point
  • New Eddi
  • TalkTalk powerline internet device which connects HEMS and Eddi to the internet over the house wiring (more robust that WiFi). There’s a mains socket behind which includes a USB output to power the HEMS. The mains plug in front powers the transmitter (above) and the various relay-activated operations.

The Harvi is connected to two current clamps and transmits the signals from those clamps to the Eddi. Each Harvi can connect to three sensors but I only use two: (i) the import/export to/from the house and (ii) the output of the solar panels.

To the left on the Harvi is the isolator for the solar panels (which has a current sensor inside) and above is the generation meter for the solar panels.

Harvi transmitter
The Myenergi app

The Eddi connects to the app via the internet connection described above. The app has a range of features but the screen illustrated shows (clockwise from 3 o’clock):

  • Power to/from the grid (measured via Harvi)
  • Power from solar panels (measured via Harvi)
  • Power diverted to immersion heater (measured via Eddi)
  • Power to home (inferred by calculation from the above).
  • Percentage of green energy (i.e. from panels rather than grid).

I’m pretty pleased with my purchase which I found intuitive to install and works well. The system also provides capability to add more sensors so I may well add to that including a sensor for my battery on the existing Harvi and possibly one in the garage for my older car charger.

Powervault battery control states

One of the roles of my Home Energy Management System (HEMS) is to switch the control state of my Powervault storage system. Many users would probably leave the system in Normal where it always either charges or discharges proportionally to solar surplus or deficit respectively. Other users on time of use tariffs might also program charging on cheap night time electricity in a fixed time window. However my HEMS integration allows for automatically charging in varying cheap windows only when solar production won’t be enough to fully charge the battery.

Powervault G200 storage system

The Powervault has six Control States covering different combinations of charging and discharging behaviour. Only three of these are used in my implementation. The other states could be more relevant if I was paid to export to the grid, and the value of that export started to exceed bought electricity costs; but in my case I’m not paid for export and thus any export is an error state.

Control stateChargeDischargeWhen used
Force dischargeNoFull powerunused
Only dischargeNoProportionateunused
NormalProportionateProportionateHigh electricity cost
Only chargeProportionateNoMedium electricity cost
Force chargeFull powerNoLow electricity cost
DisabledNoNounused
Powervault Control States

While I could use the HEMS to load a full daily schedule into the Powervault, instead I store all my schedules in the HEMS itself and switch the state of the Powervault (and indeed my other devices) every half hour as dictated by the schedule. This approach also gives me some flexibility to nuance Control State as a function of state of charge which isn’t available via the Powervault’s own scheduler.

The HEMS thus has three roles:

  1. Creating the daily schedule from future electricity prices and the solar forecast.
  2. Switching the states of the devices every 30 minutes following the schedule.
  3. Real time data display and data upload to Solcast the solar forecasting service.

Reviving the battery II

Last December I wrote of reviving my Powervault home storage battery with new internal batteries. However subsequently I decided that my old batteries were worth more to me for their remaining storage capacity than they were to me as scrap, and instead I would try to use both my old and new batteries.

I knew that there wasn’t space inside the case for two sets of batteries so I planned to add the additional batteries in a separate rack outside the standard case.

Powervault G200 storage
Electronics side of G200

I decided to explore the right hand side of the Powervault where the controls reside. The picture shows a large blue box which is the inverter/charger and handles the conversion of AC to DC to charge the battery, and DC to AC to discharge the battery. The open green circuit board to the right controls the inverter/charger and connects it to the internet for monitoring and control. Down the middle is an umbilical that connects the batteries (on the reverse side of the panel) to the inverter/charger. Each pair of batteries on a shelf is connected separately to the inverter/charger.

I fairly rapidly determined that there wasn’t space to add additional battery cables to the inverter/charger so instead I added two terminal blocks (see right) in the free space to the right, moved the existing battery cables from the inverter/charger to the blocks, and then added new cables from the blocks to the inverter. At this point the system was checked to ensure that all was working correctly.

Terminal block with applications
DC fuse

I then added two DC cables from the blocks, through a suitable fuse, and through new cable glands to the external batteries.

All the additional parts including batteries, blocks, cables, fuse and glands are intended for similar battery storage applications in boats or caravans.

The results of this are that the battery continues to function correctly supporting the needs of the house. The above graphs, from Powervault and smart meter respectively, show:

  • 00:00 to 00:30 – home still running on stored energy from yesterday having already run through yesterday evening on stored energy.
  • 00:30 to 01:00 – battery charges for half an hour on cheap rate power (7.5 p/kWh) as HEMS predicts that today may not be sunny enough to fully charge the battery.
  • 01:00 to 04:30 – battery idle while other household loads charge from cheap power.
  • 04:30 to 07:00 – house runs from stored energy until sun comes up.
  • 07:00 to 18:00 – house running from solar with battery charging, car charging and water heating at times from solar. Some export from 15:00 as all storage devices (battery, car and water cylinder are fully charged).
  • 18:00 to 00:00 – home runs from battery.

Through the day the battery supported the house to the tune of 5.18 kWh and then the next day for 30 minutes until the cheap rate started. The Powervault had notionally 4 kWh usable capacity as supplied, although theoretical capacity was higher (3 x 120 Ah x 24 Volts = 8.64 kWh). Theoretical capacity is higher now, but I suspect that the control board continues to limit the usable capacity. A low ratio of usable capacity to theoretical capacity should be good for battery longevity as depth of discharge is limited. The older batteries have now been in use for over six years, although they have only been supported by the newer batteries for a couple of months.

Transportation update

It’s approaching sixteen years since I bought my first electric vehicle.

Things have moved on enormously from my first electric vehicle – a REVA G-Wiz. The G-Wiz is a small vehicle similar in size to a Smart car, but strictly was not even a car at all but rather a quadricycle. As a quadricycle it is limited by law in both weight and power resulting in a tiny underpowered vehicle compared to even a small car, however that was all that was available at the time. It was good for 30 to 40 miles according to time of year and speed and maxed out at around 50 miles per hour although 30-40 miles per hour was more achievable.

2007 REVA G-Wiz
2014 Vauxhall Ampera

In the summer of 2014 I bought a new Vauxhall Ampera. I’d initially seen the Ampera at a motor show and dismissed it as neither a full electric vehicle or a plug-in hybrid. However a few years later and now married with a young child I was looking for something bigger. Now I saw it as being more capable than the G-Wiz in size, range, and safety with the bonus of having an engine for the occasional longer trip. I’ve now owned it for over eight years.

The latest addition to the household is a 2020 Ford Kuga PHEV. This is similar to the Ampera in many respects with some variations.

2020 Ford Kuga PHEV
AttRiutereva g-wizVAUXHALL Ampera Ford kuga PHEV
Usable battery size10 kWh10 kWh10 kWh
Vehicle typeBEVPHEV/E-REVPHEV
Range30-40 miles30-40 miles30-40 miles
Seats2+2 seats4 seats5 seats
ConnectedNoNo (at least in Europe)Yes
Electrical capabilityLow (heavily compro-mised versus ICE)High (Same as ICE)Medium (Poorer than ICE but keeps up with traffic)
Electrical inlet connector3-pin caravan5-pin BS EN 62196 Type 1 / J1772 7-pin BS EN 62196 Type 2
ICE capabilityn/a (BEV)Conventional 1.4L Atkinson 2.5L (for economy)
Transmission Simple reduction between electric motor and wheels. No ICE or variable transmission ratio.Similar to Kuga but has additional clutches and brakes allowing transmission to either work as powersplit or separate into a series hybrid with a generator set connected only electrically to the motor and wheels.Powersplit – combines two electrical machines with torque trans-mission from ICE to wheels via both electrical and mechanical means. Lacks the ability to split the transmission into a separate generator and motor set.
Comparing my plug-in vehicles
Powersplit transmission

Both Ampera and Kuga use epicyclic or planetary transmissions to create a variable ratio between engine and wheels. Both gear together three shafts such that the speed and torque on any two shafts determines the speed and torque on the third shaft. Both have the wheels connected to one shaft, the engine to the second, and an electrical machine to the third. However Ampera and Kuga differ in that Kuga places the second electrical machine on the same shaft as the wheels, while the Ampera places the second electrical machine on the same shaft as the engine (which helps when splitting the transmission to create a series hybrid).

I’d differentiate the Ampera from the Kuga as describing the Ampera as a short range electric vehicle with an ICE for occasional longer trips, while the Kuga is more of an enhanced hybrid optimised more for those longer trips with a higher efficiency Atkinson engine. The Ampera on the other hand has greater electrical capability (uncompromised performance electric versus ICE) with greater transmission sophistication having the ability to switch between being a series hybrid or a powersplit.

Battery Economics

Having recently spend around £600 replacing the internal batteries in my Powervault G200, I thought I’d look at the economics of that decision.

Powervault G200 storage

Originally I bought the Powervault to store the surplus electricity from my solar panels and then use that electricity later instead of buying electricity. I don’t receive any revenue for exporting electricity so to fill my battery from my solar surplus costs me nothing. I currently pay 35 p/kWh for daytime electricity. So, if I fill my battery from my solar surplus then I save 4 kWh (the capacity of the battery) x 35 p/kWh = £1.40 per day.

More recently I also charge the battery on cheap overnight electricity when there won’t be enough solar to charge the system fully.

I only get four hours of cheap power which isn’t enough to fully charge the battery. The battery charges at 0.8 kW so in four hours would use 4 hours x 0.8 kW = 3.2 kWh which at 7.5 p/kWh costs 24 pence. The system has a cycle efficiency of 80% so 3.2 kWh in gives 3.2 kWh x 0.8 = 2.56 kWh out. 2.56 kWh energy out is worth 2.56 kWh x 35 p/kWh = 89.6 pence. Having spend 24 pence to save 89.6 pence then I’ve made net savings of 65.6 pence per day.

G200 User Portal

A day like that shown in the portal with 4.75 kWh in savings is a bonus in terms of annual savings although it may accelerate the deterioration of the battery as it’s been double-dipped during the day with both overnight and solar charging.

If I assume that half the time the system is filled by solar and half from the grid then my annual savings are 0.5 x (£1.40 + £0.656) x 365 = £375.22. Having spent £600 on the new batteries then my payback time is £600 / £375.22/year = 1.6 years. The previous batteries lasted for 6 years although the new batteries were relatively cheap and might not be expected to last for so long. However even if the new batteries only have half the life of the prior ones then that still seems like an attractive investment.

My total investment in the battery storage is now £2600 – £2000 to buy it six years ago plus the £600 just spent on new batteries. That’s a payback of about seven years with expectation that the new batteries will extend the life of the system to nine years.

Reviving the battery

Powervault G200

My Powervault battery is now six years old. Initially I used it only to store surplus electricity from my solar panels but more recently have also managed it from my HEMS to reduce my energy costs by charging it when electricity is cheapest to offset consumption when electricity is most expensive. However it’s become clear in recent weeks that the working capacity of the battery had significantly declined.

Powervault portal for G200 before battery replacement

My Powervault is designed to have a reasonable life – its 7.92 kWh internal batteries (6 batteries x 12 Volts x 110 Ah) are limited to 4 kWh to maximise the number of charging cycles. However after 6 years working capacity had dropped to less than half of that and I decided that the time had come for replacement. Declining battery capacity is illustrated above with the blue line showing battery charging between 00:30 and 04:30. If the battery had continued at constant power (about 750 Watts) then it would have charged 3 kWh (4 hours x 750 Watts / 1000 Watts/kW). However here the battery has ceased to draw 750 Watts from around 02:00 and thereafter power consumption has significantly declined – this may be about 1.75 kWh rather than 3 kWh.

These figures (1.75 and 3 kWh) are also energy into the battery – but you don’t get out everything that you put in. With a quoted round trip efficiency of 80% then 1.75 and 3 kWh in would deliver 1.4 and 2.4 kWh output versus the original capacity of 4 kWh.

I reached out to both my installer and Powervault themselves but neither got back to me with any proposal to investigate or replace the batteries so I resolved to do it myself.

Powervault G200 internals

Initially I did a little investigation by removing the top and side panel of the Powervault (after turning off the power). Inside I found six twelve Volt one hundred and ten Amp hour batteries arranged over three shelves on plug and socket connections – so fairly straightforward to replace. A search of eBay for deep cycle batteries of similar capacity indicated prices range from around one hundred to around three hundred pounds per battery (six required). I eventually decided to go for the cheapest as I wasn’t entirely sure of success and I don’t know for how long my six year old system will be supported even if battery replacement was successful.

6 x Numax CXV31MF Batteries 105Ah

The Numax batteries are sold as deep cycle batteries ideally suited for solar installations.

Powervault G200 installation.


The process for installation of new batteries consists of unplugging and removing all the batteries from top to bottom, and then installing new batteries from bottom to top – two on each of the three shelves. I found it necessary to undo the fixing screws on the back of one of the shelves in order to flex it upwards to create enough clearance to get the batteries on the level below out.

Powervault portal for G200 after battery replacement

It’s only been just over a week now, and it’s fairly dull weather at the moment, but it does seem that the replacement has been successful. The dashboard above for example shows both that the system has charged at full power for 4 hours overnight (3 kWh in) and also that 4.75 kWh has been saved (versus 4 kWh working capacity) as to a degree there’s some double-dipping going on with the two partial charges of more than half capacity (one overnight on cheap mains power and one during the day on surplus solar power).

Time will tell what the longevity of these cheap batteries is.