Both FIT AND export payments to be cancelled next year

[ProDave]

To do the cost analysis, should you be working not on the whole unit price, but just the cost of replacing the batteries at 10 years, which should be less than the whole unit.

 

That assumes in 10 years you will be able to buy suitable batteries.

[Jeremy Harris]

28 minutes ago, Alphonsox said:

Given the prices quoted in the link above I would think there is still some margin for price reduction. If my maths is correct the Sofar system is around £250 per additional kWh of battery. Current automotive industry pricing for Li-Ion batteries is in the £100 per kWh range. Clearly the Sofar batteries come in a nice box and they have nowhere near the same economies of scale but I can see competition and volume driven by lack of export tariff bring their price down quite rapidly.

 

 

I agree, and I think that's one reason the Pylontech battery units have been dropping in price a lot.  They are now £906 inc VAT and delivery for a 2.4 kWh module, and it's only a few months since the same modules were around £1500, IIRC.  I can see these battery modules coming down to perhaps £600 a module before too long.

 

As the only difference between the 4.8 kWh, 7.2 kWh and 9.6 kWh Sofar units is the number of battery modules, and as the battery modules are user-replaceable, it may well be an idea to start small, say with a 4.8 kWh system, then look to add capacity if you find you need it and if the price of the battery modules drops further.

 

6 minutes ago, ProDave said:

To do the cost analysis, should you be working not on the whole unit price, but just the cost of replacing the batteries at 10 years, which should be less than the whole unit.

 

That assumes in 10 years you will be able to buy suitable batteries.

The general rule of thumb is that about ten years is the life of an inverter, driven mainly by the life of the commutation capacitors, so the chances are the whole system would need replacing at around the ten year point.

 

[ToughButterCup]

3 hours ago, JSHarris said:

[...]

Sofar is an inverter and battery storage system with a capacity of 9.6 kWh.  It's not that large and I made provision to install a storage system when we built the house, so there is a base for a slim cabinet with a conduit leading to it plus two runs of SWA running close by.

[...]

 

Could you give us a little more detail about that please? We won't be able to afford batteries for a little while, but I can foresee a time when we might we'll do so.

In which case - rather like provision for a spare this-or-that elsewhere in the house - it would be good to make provision for the time when we can fit some battery capacity.

[Jeremy Harris]

12 minutes ago, recoveringacademic said:

 

Could you give us a little more detail about that please? We won't be able to afford batteries for a little while, but I can foresee a time when we might we'll do so.

In which case - rather like provision for a spare this-or-that elsewhere in the house - it would be good to make provision for the time when we can fit some battery capacity.

 

 

Sure, happy to. 

 

What I've done is run a spare conduit that leads from our meter cabinet to a location around the back of our house where I could either wall mount, or most likely mount in a fire proof narrow shed (I'm thinking of getting one of those 5ft x 3ft steel lean to sheds for the job).  This conduit is separate from the power cables and is there to run a data/sense cable from the meter tails to the battery system, in order to allow it to know when we are importing or exporting.  The Sofar system (and I suspect other similar systems) allow the sense cable to be up to 20m long, so the inverter and battery pack can be around 20m from the meter (or possibly the consumer unit - depends on whether the consumer unit handles all the power - ours doesn't, we have two feeds from the meter that are taken off before the house consumer unit).

 

For the power side, then you need two cables if you want to be able to fit a auto changeover switch to allow the battery to supply back up power to some circuits in the event of a power cut.  I've run two lengths of 4mm² SWA from our meter box location to the location of the battery system (if we fit it).  These are not in conduit, but are buried directly and are separated from the conduit by around 300mm in the trench, just to minimise the risk of any interference in the data/sense cable.

 

I also have another spare run of 4mm² SWA that goes to the water treatment shed, giving me the option of having that on a switched back up supply (not sure it's needed, given our pressurised water storage capacity).

 

The way this lot would work is this.  The inverter/charger in the battery system senses whether the whole site is importing or exporting via the sense lead from the meter cabinet.  The grid connection will be from an additional weatherproof termination by the meter cabinet that will connect to a spare RCBO protected connection to the incoming grid supply, on our side of the meter (I already have a spare slot in the weatherproof CU that we used as our temporary site supply).  That takes care of the basic operation of the battery system, in that it can run in normal mode with just a single connection plus the sense connection to the incoming supply.  If you don't need the option of a battery back up for power cuts, then this is all you need, in fact the power feed to the battery unit can come direct from a spare connection in your consumer unit, via an isolator, just as a PV system would connect.

 

The second power cable is needed if you want the option of running some critical circuits from the battery system in the event of a power cut.  To do this, the supply to those critical circuits (pretty much the ones suggested by @Ferdinand earlier) would be via a separate consumer unit, supplied from the main consumer unit via the changeover switch (either automatic or manual).  The backup power outlet from the battery system would connect to the other side of the changeover switch.

 

With an automatic changeover switch, then as long as the grid is up the switch will direct power from the grid, via a protected feed, to the essential circuits consumer unit.  If the grid goes down, the auto changeover sense that and immediately (probably within a few tens of ms) and switches the essential supplies to the inverter in the battery system.  When grid power returns the auto changeover senses that and switches back to the grid supply.

 

In our case I'd selected the location for a possible battery system at the design stage, and allowed room for it against the retaining wall at the rear of the house, where it's in the shade all the time (good for helping to keep things cool).  Putting in a few extra cables and a conduit run was pretty easy, my only regret is that I didn't include a direct extra run up through the slab to the house consumer unit.  That would have made adding the essential circuits supply neater, rather than the arrangement I've ended up with which is a cable run to the water treatment shed and then joined to ones that runs into the house.

 

Hope this doesn't sound to muddled!

 

 

[ToughButterCup]

Thanks very much. Re and re-read time.

[SteamyTea]

I am seeing that dodgy investment arithmetic again.

If you went out on Monday and invested £4000 in United Bank's fixed rate NISA/ISA, after 5 years you would have £4350 tax free.

 

If you went out and bought a battery storage system, after 5 years you would have £0 plus any savings.

 

Worth bearing in mind that the last 3 weeks I have used less than 100 kWh.  As most of that is during the night, so the cost is about a 3 quid a week.

[ProDave]

Yes indeed Steamy.  Most people when working out the ROI on e.g solar PV tend to forget that the capital has gone, so you need to recoup that before you start making a return.

[joe90]

9 hours ago, SteamyTea said:

I am seeing that dodgy investment arithmetic again.

If you went out on Monday and invested £4000 in United Bank's fixed rate NISA/ISA, after 5 years you would have £4350 tax free.

 

If you went out and bought a battery storage system, after 5 years you would have £0 plus any savings.

 

Worth bearing in mind that the last 3 weeks I have used less than 100 kWh.  As most of that is during the night, so the cost is about a 3 quid a week.

 

The very reason I did not include solar in my build, bought a classic car instead and get a far better return ?

[newhome]

2 minutes ago, joe90 said:

 

The very reason I did not include solar in my build, bought a classic car instead and get a far better return ?

 

My brother collects classic motorbikes but still has PV ?. He got the original FIT though so probably better value. He just paid 5k!!! for a FS1E for his son and was gutted he hadn’t kept his own from back in the day when they were peanuts. Who can ever predict how these things will go though. 

 

 

[newhome]

14 minutes ago, ProDave said:

Yes indeed Steamy.  Most people when working out the ROI on e.g solar PV tend to forget that the capital has gone, so you need to recoup that before you start making a return.

 

That’s what I’m debating as I consider installing an ASHP here. RHI should provide £9100 return over 7 years and the government provides interest free loans so on the surface it seems like a decent deal, although clearly I am taking a gamble that the electricity savings will benefit me as I don’t intend to be here for 7 years. The biggest hurdle to overcome? Getting someone who wants to install one ???

[Jeremy Harris]

10 hours ago, SteamyTea said:

I am seeing that dodgy investment arithmetic again.

If you went out on Monday and invested £4000 in United Bank's fixed rate NISA/ISA, after 5 years you would have £4350 tax free.

 

If you went out and bought a battery storage system, after 5 years you would have £0 plus any savings.

 

Worth bearing in mind that the last 3 weeks I have used less than 100 kWh.  As most of that is during the night, so the cost is about a 3 quid a week.

 

 

Not dodgy investment arithmetic if you are just looking at the possible savings a battery storage system may provide over not having one, and especially not if you are specifically pointing out that the return on investment is less than zero if taking into account what that investment may have returned if invested in the best possible, most tax efficient, savings scheme.

 

As I tried to highlight, the calculation is extremely complex and subject to a lot of unknown variables, that may have a significant impact on any possible return.  There's also the matter of valuing less tangible returns - for example, we have frequent short power cuts, what value should be placed on having a system that provides power to critical loads during power failures?

 

To chuck in another variable, smart meters will introduce two factors that may significantly impact on bills (assuming that they will eventually become mandatory, as I am certain will be the case.  The first is the adoption of short period tariff changes to match the variation in wholesale price.  Best estimates are that half hourly, or perhaps hourly, "smart" tariffs will vary from around £0.05/ kWh to around £0.30/ kWh, with the highest rates being during peak demand periods (probably early evenings).  Being able to use battery power during high tariff periods significantly swings the return on investment sums for a battery system.

 

Secondly, "smart" meters have the ability to introduce kVA charging, rather than kWh charging, and I very strongly suspect that the suppliers will take up this option.  It used to be the case that domestic use was considered to be around unity power factor (PF), but that's far from true.  Virtually every domestic device, from a washing machine through to TVs, LED lighting, device chargers and power supplies etc, has a PF that's way below unity - often around 0.5 to 0.6.  This places a significant burden on the distribution network and there is pressure building to switch to kVA charging for domestic premises (it's why "smart" meters have the option built in to register kVA as well as kWH).  kVA charging would almost double the tariff for a typical modern home, with the majority of the loads having a low PF.

 

A smart inverter (in fact most inverters) can apply a degree of PF correction by virtue of the way they sync to the grid.  I've not looked in detail as to how effective this is, but it's noticeable that the PF of our house markedly improves when the PV system is generating, which has to be down to the way the inverter works.  Our poorest PF loads are the ASHP and the MVHR (with it's internal air-to-air heat pump).  Both of these use switched mode controllers, and run with a PF of between 0.6 and 0.7 at best.  They aren't as bad as the LED lights, though, they run at a PF of around 0.5, but thankfully their total kVA is relatively low.

 

Finally, you are an untypical example of a domestic user, @SteamyTea.  You are very aware of your energy use and have taken significant measures to both monitor it and reduce it.  At a guess I'd suggest you probably use around half the energy of someone in similar home that was not as energy-conscious.

 

 

[Nickfromwales]

20 hours ago, JSHarris said:

What I've done is run a spare conduit that leads from our meter cabinet to a location around the back of our house where I could either wall mount, or most likely mount in a fire proof narrow shed (I'm thinking of getting one of those 5ft x 3ft steel lean to sheds for the job). 

This was the first thing that popped into my head tbh. What does BR say about its suitability eg odds / mitigation of fire hazard etc? Double firecheck PB'ing in a plant room if within the dwelling with fire / smoke detectors etc? 

I see these / others ( in online marketing ) shown as fitted inside, is this permissible for UK BR?

[pdf27]

55 minutes ago, JSHarris said:

A smart inverter (in fact most inverters) can apply a degree of PF correction by virtue of the way they sync to the grid.  I've not looked in detail as to how effective this is, but it's noticeable that the PF of our house markedly improves when the PV system is generating, which has to be down to the way the inverter works.  Our poorest PF loads are the ASHP and the MVHR (with it's internal air-to-air heat pump).  Both of these use switched mode controllers, and run with a PF of between 0.6 and 0.7 at best.  They aren't as bad as the LED lights, though, they run at a PF of around 0.5, but thankfully their total kVA is relatively low.

Have you got any inverter-driven devices you can check? These should in theory have a DC-link between motor and mains which ought to eliminate the phase shift, but I'm very curious how well they work out in practice. 

[epsilonGreedy]

Is it time to debate whole-house DC lighting again?

 

I got some negative feedback here 4 months ago for exploring this subject but now the maths has changed. With all those PV watts stored locally in a battery does it make sense to loose 10% stepping up the electricity feed to 240 volts AC and then loose possibly 50% again converting that back to DC within some cheap circuitry embedded within each LED light.

Edited July 21, 2018 by epsilonGreedy

[Jeremy Harris]

38 minutes ago, Nickfromwales said:

This was the first thing that popped into my head tbh. What does BR say about its suitability eg odds / mitigation of fire hazard etc? Double firecheck PB'ing in a plant room if within the dwelling with fire / smoke detectors etc? 

I see these / others ( in online marketing ) shown as fitted inside, is this permissible for UK BR?

 

AFAIK, the regs haven't caught up with the potential fire risk or risk category that should, perhaps, apply to battery storage, especially that using lithium chemistry cells.  The risk is probably pretty low, even though the perception of risk is higher than it perhaps should be.  My view has always been that the battery storage will be outside, away from the house and in a steel shed  or housing against a concrete wall.

 

6 minutes ago, pdf27 said:

Have you got any inverter-driven devices you can check? These should in theory have a DC-link between motor and mains which ought to eliminate the phase shift, but I'm very curious how well they work out in practice. 

 

I measured the PF of our inverter driven ASHP, it seems to be around 0.7 PF when running on part load.  it's probably a bit better when more heavily loaded, I'd guess.  I've not measured our MVHR, but that uses switched mode control of the fans and the ASHP, so would expect it would be around the same.  The LED lighting is all run on switched mode supplies and is around 0.5 to 0.6 PF.

 

5 minutes ago, epsilonGreedy said:

Is it time to debate whole-house DC lighting again?

 

I got some negative feedback here 4 months ago for exploring this subject but now the maths has changed. With all that PV capacity stored locally in a battery does it make sense to loose 10% stepping up the electricity feed to 240 volts AC and then loose possibly 50% again converting that back to DC within some cheap circuitry embedded within each LED light.

 

With respect, you didn't get "negative feedback" at all, just some uncomfortable facts about the non-availability of DC rated light switches, based on my asking every manufacturer I could find whether they could let me have the safe DC load for their light switches (none could).  The fact is that no one yet seems to make wall mounted light switches that will accept low voltage, relatively high current, DC.  If you want evidence of the difference between switching AC and DC with a standard light switch, and the arcing that occurs under DC loads significantly below the switch rated AC load, then John Ward (who I have mixed views about) has a video on his You tube channel that illustrates the issue well: John Ward DC light switch testing video

 

One way around the DC switching problem (but doesn't address the losses issue - see below) might be to change the relays inside one of the remote switching systems available.  I've just fitted the Quinetic light switch system and found it works well.  The receivers have internal relays rated at 16 A AC, and in all probability you might be able to either obtain the safe rated DC current for these, or replace them with suitably rated relays.  The only other problem to overcome would be to modify the internal circuitry to operate on whatever DC voltage you decided to use, as at the moment the receivers have an internal AC operated  power supply (that may work on DC - I've not looked to see if it's a switched mode or something cruder like a capacitive dropper - if the latter it will only work on AC as it stands).

 

Looking at the losses, then a typical switched mode power supply feeding LEDs will be at around 85% efficient, it could be around 90% efficient, and it's perfectly possible to get up to around 93% to 94% efficient if the manufacturers really wanted to (for example, using synchronous rectification easily gets over 90% efficient).

 

Battery storage systems vary in terms of the DC voltage they operate at internally.  The lowest seem to work at around 48 VDC, some may operate at higher DC voltages.  That means any DC lighting system still has to use a DC-DC converter to provide constant current drive to the LEDs if run from a storage battery, and that is likely to be around the same efficiency as any AC-DC converter, so offers no gain in efficiency terms

[epsilonGreedy]

25 minutes ago, JSHarris said:

With respect, you didn't get "negative feedback" at all, just some uncomfortable facts

 

 

With respect I think this is just a British blind spot, a bit Googling reveals a healthy market for DC led lightning overseas. A humble DC switch on my old boat could comfortably handle 20 watts at 12v volts DC.

 

People on this forum seem happy to shell out £5k to £20k on bleeding edge home technology yet go belly up at the first hint of DC trouble in a £10 light switch. 

[pdf27]

49 minutes ago, JSHarris said:

I measured the PF of our inverter driven ASHP, it seems to be around 0.7 PF when running on part load.  it's probably a bit better when more heavily loaded, I'd guess.  I've not measured our MVHR, but that uses switched mode control of the fans and the ASHP, so would expect it would be around the same.  The LED lighting is all run on switched mode supplies and is around 0.5 to 0.6 PF.

That's something for me to check on Monday then - I'm working on an extremely lightweight megawatt-scale hybrid power system with a DC link. The working assumption at the moment is that the generator will be working at unity power factor thanks to the rectifier, but if that isn't true on some systems then it may not be on others - and that could have a huge impact on the design. Thanks.

[Jeremy Harris]

27 minutes ago, epsilonGreedy said:

 

With respect I think this is just a British blind spot, a bit Googling reveals a healthy market for DC led lightning overseas. A humble DC switch on my old boat could comfortably handle 20 watts at 12v volts DC.

 

People on this forum seem happy to shell out £5k to £20k on bleeding edge home technology yet go belly up at the first hint of DC trouble in a £10 light switch. 

 

Why emphasis "people on this forum" in such a derogatory way?

 

The hard facts are that unless you have a low loss DC supply that is at the correct voltage to run LED DC lighting, then you are still going  to have the same order of conversion losses in converting that DC to the constant current drive required for LED lighting.  LEDs aren't voltage driven devices, they are current driven, so some form of constant current driver is required and it doesn't really make a jot of difference in efficiency terms as to whether that constant current driver is supplied with AC or DC.

 

Boats are a subject to a different regulatory regime, and boat switches are specifically designed to switch high current low voltage, DC loads, and usually be weatherproof and sealed so that they cannot initiate a gas explosion (at least the better ones are).

 

The hard fact is that I could not find a single manufacturer of approved house wall switches that either manufactured a switch designed to switch DC loads, or that would provide a de-rating factor for any of their switches if used to switch DC.  The major problem is that all domestic light switches, approved to EN 60669-1, have very slow moving internal contacts.  When you switch off a normal 10A rated wall light switch the contacts are really slow to open - this doesn't matter much for AC, where the voltage across the switch drops to zero every 10ms anyway, and it makes the switch cheaper to manufacture.

 

If you look inside a boat or car DC rated switch you will find it has a more complex contact mechanism and includes and accelerator spring to make the switch contacts open very much faster than those of a domestic light switch.  This toggle action is there deliberately to ensure the opening arc is very quickly quenched, by moving the contacts apart vary quickly, and also by including a significantly greater gap between the opened contacts, so that no arc can be sustained during switch off.

 

It would be possible to design wall light switches to operate like this, but may well increase the depth of the switch.  For a host of reasons, domestic light switch manufacturers strive to keep the switch depth as slim as possible.  Including a toggle-type switch contact mechanism, with an accelerator spring to separate the contacts as quickly as possible, would mean turning the contact orientation within the switch through 90 degrees and almost certainly result in deeper switches.

 

If you really wanted to switch, say, 12v DC with wall switches, there's no reason I can see not to buy blank wall plates, fit deep wall boxes and fit approved DC toggle switches to the blank plates.

[Jeremy Harris]

3 minutes ago, pdf27 said:

That's something for me to check on Monday then - I'm working on an extremely lightweight megawatt-scale hybrid power system with a DC link. The working assumption at the moment is that the generator will be working at unity power factor thanks to the rectifier, but if that isn't true on some systems then it may not be on others - and that could have a huge impact on the design. Thanks.

 

 

Be interested to hear your findings. 

 

I should add that my measurements of PF, power etc were done using a relatively cheap  HOPI meter: Hopi meter and I would give a strong health warning to anyone thinking of getting one of these that, although they are pretty accurate and reliable, there is NO WAY they meet any normal UK/EU safety standards, so use with extreme care!  Fine in the hands of someone that understands all the potential risks they present, but it still needs treating with caution.

[pdf27]

14 minutes ago, JSHarris said:

Be interested to hear your findings. 

 

I should add that my measurements of PF, power etc were done using a relatively cheap  HOPI meter: Hopi meter and I would give a strong health warning to anyone thinking of getting one of these that, although they are pretty accurate and reliable, there is NO WAY they meet any normal UK/EU safety standards, so use with extreme care!  Fine in the hands of someone that understands all the potential risks they present, but it still needs treating with caution.

I'll share what I can when I get an answer, but since the project has only just been made public (https://www.safran-electrical-power.com/media/safran-lance-ses-nouvelles-gammes-de-generateurs-et-moteurs-electriques-geneus-et-engineus-20180716) there will be quite a lot I won't be able to discuss yet.

[epsilonGreedy]

43 minutes ago, JSHarris said:

Why emphasis "people on this forum" in such a derogatory way?

 

 

I cannot think of a less offensive collective phrase for the people on this forum.

 

43 minutes ago, JSHarris said:

The hard facts are that unless you have a low loss DC supply that is at the correct voltage to run LED DC lighting, then you are still going  to have the same order of conversion losses in converting that DC to the constant current drive required for LED lighting.

 

 

Every boat floating in a British marina provides an example of such a low loss supply and probably every caravan too. Live-aboard boaters count their daily amp hours and struggle to match PV amps generated with consumption, there is nothing exotic to matching battery volts output with the consumption needs aboard. Larger american yachts tend to favour 24 volts which is just two commodity batteries in series.

 

43 minutes ago, JSHarris said:

If you look inside a boat or car DC rated switch you will find it has a more complex contact mechanism and includes and accelerator spring to make the switch contacts open very much faster than those of a domestic light switch.  This toggle action is there deliberately to ensure the opening arc is very quickly quenched, by moving the contacts apart vary quickly, and also by including a significantly greater gap between the opened contacts, so that no arc can be sustained during switch off.

 

 

All of which illustrates a low cost commodity solution exists for a very low tech problem.

 

The worlds of boating and RVs have proven cheap solutions for this whole challenge. It therefore seems strange to me that the forum is predisposed to focus on a few high margin fringe suppliers who target their products as price insensitive self builders. 

Edited July 21, 2018 by epsilonGreedy

[Dreadnaught]

12 minutes ago, epsilonGreedy said:

The worlds of boating and RVs have proven cheap solutions for this whole challenge.

 

Would the use of such solutions for boats and RVs meet Building Regulations and electrical safety standards required for UK buildings?

[newhome]

3 hours ago, JSHarris said:

assuming that they will eventually become mandatory, as I am certain will be the case.

 

Hopefully not anytime soon ...

 

https://www.bbc.co.uk/news/business-44903471

 

 

[Jeremy Harris]

 

1 hour ago, epsilonGreedy said:

 

I cannot think of a less offensive collective phrase for the people on this forum.

 

 

Every boat floating in a British marina provides an example of such a low loss supply and probably every caravan too. Live-aboard boaters count their daily amp hours and struggle to match PV amps generated with consumption, there is nothing exotic to matching battery volts output with the consumption needs aboard. Larger american yachts tend to favour 24 volts which is just two commodity batteries in series.

 

 

All of which illustrates a low cost commodity solution exists for a very low tech problem.

 

The worlds of boating and RVs have proven cheap solutions for this whole challenge. It therefore seems strange to me that the forum is predisposed to focus on a few high margin fringe suppliers who target their products as price insensitive self builders. 

 

I'm struggling to get my head around the concept of low loss, when applied to a 12 VDC supply for household LED lighting.  LEDs are, as mentioned before, current driven devices, so all LEDs that have a voltage rating have an integral driver circuit of some description to run the LEDs at a current that's within their acceptable operating range.  12V LEDs, for example, have to have some form of driver in order to convert 12V DC to the constant current needed to run the LEDs themselves.  Cheaper 12V DC LEDs may just use a dropper resistor, which is not as inefficient as it seems, as three white LED chips connected in series (inside the lamp itself there will be many LED chips) will operate at a forward voltage drop of between 3V and about 3.5V. depending on temperature mainly - as they warm up their forward voltage drop decreases.  This means that arrays made up of three LED chips in series will run at a variable voltage of between 9V and 10.5 V, but ideally at as close to a constant current as possible.  More expensive 12 VDC LED lamps will usually use a switched mode DC to DC constant current driver circuit. 

 

Both systems drop the 12 VDC coming in to a variable output voltage, but a near-constant current, that matches the rating of the LEDs.  For example, I have some 12 V MR16 downlighter LEDs that have 24 off 5730 LED chips, wired as an 8 x 3 array (so 8 parallel connected rows of 3 LEDs in series).  Each 5730 LED has a constant current requirement of around 150 mA.  To get approximately 150 mA through each of the 8 row of series connected LEDs, each row of three LEDs has an 18 ohm resistor in series with it.  This isn't ideal, as it means that the LED current will vary a bit with temperature, but it's safe enough, as the maximum allowable current for a typical 5730 LED chip is around 200 mA. 

 

Analysing the efficiency of this arrangement for a constant 12 VDC supply we get:

 

For LEDs operating at their lowest forward voltage (Vf) of 3.0 V, then the row of 3 in series gives a voltage of 9 V, and the 18 ohm resistor therefore has a voltage across it of 12 V - 9 V = 3 V.  The current flowing through the 18 ohm resistor (and hence the row of three LEDs) will be 3 V / 18 = 167 mA, within the 200 mA maximum allowed.  The power loss in each of the eight 18 ohm series resistors will be 0.167 A² * 18 = 0.502 W.  The power in each LED will be 3 V x 0.167 A = 0.501 W.  A row of three LEDs will give a power of 3 * 0.501 = 1.503 W.  The efficiency at this operating point for a typical resistive dropper LED driver will be around 75%, a pretty low figure, but this type of 12 V  LED is common and cheap to make.

 

For LEDs operating at their highest Vf of 3.5 V, then the row of 3 in series gives a voltage of 10.5 V, and the 18 ohm resistor therefore has a voltage across it of 12 V - 10.5 V = 1.5 V.  The current flowing through the 18 ohm resistor (and hence the row of three LEDs) will be 1.5 V / 18 A = 0.083 mA, well below the best operating current of 150 mA, so less light will be produced.  The power loss in each of the eight 18 ohm series resistors will be 0.083 A² * 18 = 0.124 W.  The power in each LED will be 3.5 V x 0.083 A = 0.29 W. A row of three LEDs will give a power of 3 * 0.29 = 0.8715 W, a fair bit less than for the 3V LED Vf case.  The efficiency at this operating point for a typical resistive dropper LED driver will be around 87.5%, about the same as a switched mode driver circuit, but the power of the LEDs is much lower than their are rated for.

 

If we now look at either an AC-DC constant current driver, or a DC-DC constant current driver (doesn't make a significant difference, both will be around 85% to 90% efficient) then the primary difference is that no matter what the LED chip Vf is, the drive current will be a near-constant 150 mA.  This means that maximum light is obtained from the LEDs, and overall the efficiency is almost certainly going to be at least as good as, probably better than, the simple resistive dropper.

 

So, until someone comes up with a true, high efficiency, 12 VDC light source, that doesn't require any form of energy-sapping driver, it's pretty hard to beat either an AC-DC constant current LED or a DC-DC constant current LED.

 

If you know of any 12 V intrinsically high efficiency light sources that don't need the drivers that all LEDs need, then perhaps you can let us know about them, as that would then swing the argument in favour of using a 12 VDC supply with 12 VDC switching.

 

1 hour ago, Dreadnaught said:

 

Would the use of such solutions for boats and RVs meet Building Regulations and electrical safety standards required for UK buildings?

 

Possibly.  The building regs approved docs and BS761 are a bit silent on ELV DC switching standards, AFAIK, but if a building inspector was satisfied that a switch used in a  house was suitable for the purpose, then there's no reason not to allow it, AFAIK.  It should be easy enough to provide evidence that a switch was suitably rated for DC use at the current it was being used  at, and I reckon most building inspectors would be OK with it.  In general they don't seem that bothered by ELV stuff, anyway, AFAICS.

[ProDave]

I suspect few building inspectors pay much attention to the electrics, other than check socket and switch heights are correct. They rely on an electrician giving them an EIC.  If the electrician was satisfied that the switches used were okay for DC and was happy to issue an EIC then I really don't think there would be any building control issues.