Archived Forum

Hi Riggers:  IMHO, the simplest approach for diverting surplus PV power to an immersion heater is to measure both current and voltage at the point where power enters the property.  All you need for this is an Arduino, a current clamp and the usual bits and pieces to glue it all together.  The objective of the excerise is to control the flow of energy through the meter, so that's surely the best place to measure - at the meter.  It doesn't matter whether you measure the current in the neutral or the live cable; they should both give the same reading.  Having said which, other designers have successfully built systems which measure power (current?) at two places and subtract these readings rather than just measuring currenta and voltage at one place.

By using the standard V&I sketch and a calibrated system, you then know how much energy is flowing in and out of your house. Chargeable energy at the meter is emon.realPower * time.  By varying the power that is diverted to your immersion, your system needs to maintain a small energy flow in the export direction.  How you go about this will be dependent on how you plan to switch power to your immersion.  If you were to purchase the same Carlo Gavazzi controller as I have done, for around £70 from Farnell or RS-online, you could then use the same sketch as I've already posted some weeks ago on this thread.  Other contributors have gone with different types of 3rd party kit, which may well have different characteristics. 

It would be great if the expensive piece of 3rd party kit could be replaced by something simple such as a triac or SSR (not that I really know the difference).  Various contributors have been thinking along these lines, but I'm not aware of a working system that operates in this way.  If you're prepared to wait a bit, there may soon be a cheaper in-house alternative.  But, for now, I think you would need to decide on your chosen controller and purchase one.  Having done so, I feel sure that anyone who has already built a similar system would be pleased to give you any assistance that you need. 

Hi Riggers,

I'm a retired Electrical Engineer - if you need help please ask (but bear in mind that the more we know about your project and how you're thinking, the better we can help you). I try to keep an eye on the forums and if I see anything that looks as if it might be unsafe, I try to steer the writer in the right direction.

The immediate 'idiot's guide' is DON'T BELIEVE EVERYTHING YOU READ ON THE WEB. Make sure the sources you learn from are reputable, and beware material that is based on practices in a different country to yours, as they might have a different mains power supply, and what might be OK there might not be for you. (I presume you are in the UK as you mention £  - I'm thinking websites that go by US practice - where their '240 V' is actually balanced about earth, what I'd call 120 - 0 - 120 V).

Learn some basic electrical theory, and some basic programming. There is some minimal theory on this site ('Building Blocks') and I usually recommend this http://www.relisoft.com/book/index.htm for learning C++ (the language of the Arduino). Look at the build guides and learn how to identify the components. If you've never done any electronic soldering before, get a piece of stripboard and a mixed bag of components and learn how to make good fast neat soldered joints (you are more likely to damage a component by being slow and allowing it to heat up too much).

Knowing what calypso_rae is planning, my advice would be to hold fire for a little while because I think he has a very cost-effective and sound system in the making. I haven't tried to cost his project, but I think it will come in at a long way under £100 all told, so it should be well worth the wait.

No pressure then, Robert ;)

Hi Robert / Calypso_rae,

May I thank both of you for your valued input, both very very informative, balanced and has set the brain cells going.

I will await the fabulous all singing project that Calypso is producing (even with the added pressure), and practice on my soldering techniques - which i found a really useful bit of advice - thank you.

I could (if required) produce pictures of my PV set up? Yes I'm in the UK (Pulborough area). Have had the PV set up for about 7 months, we blew all our saving on it, and whilst i was worried investing in 'new technology' its the best thing we have ever done, and i just want to get the most out of it - hence using the 'spare' electricity that it produces to heat up the hot water tank. The tank current has a 3KW immersion heater in it, and I appriciate that I made need to either swap it or use a step down transformer to actually engage the immersion heater, espicially as the spare electricity will be about 1.5-2 kw during good sunlight.

I will probably required a wireless switch to the immersion heater, as the tank and the PV system are on seperate floors (tank on first floor, PV in internal garage - ground floor).

Please let me know if you required anything else Robert?

With kind regards,

Riggers

I think your best way forward is to buy an arduino and a ct and then make up your circuit board with a few cheap components. Get this up and running first and then decide how you want to control the output to the immersion. Your exhisting element should be fine and will heat up a full tank using only 2kw easily with a proportional type controller.if you use Calypso_rae's sketch you can get away with just one ct and position it at your main incoming meter tails.

I would certainly stay with Calypso-rae's control scheme and not even consider adding a step-down transformer (if only on cost grounds) - neither would I use phase control because of the heavy filtering that you would need.

And yes, the 3 kW heater should be fine. The way it is controlled takes advantage of a property of electricity meters - they have a minimum current and a minimum 'packet' of energy that they register. The minimum current will always be exceeded, but not so the minimum energy. The value varies a little from meter to meter but we think it is generally 1 Wh (3600 J). So if you are generating (say) 1.8 kW, the heater is turned on for as many cycles as necessary so that on average your heater will be on for 1800/3000ths of the time so as to exactly balance the generation. And provided the nett energy remains inside that window, you neither import nor export.

I presume your immersion heater is fed by a fused spur off the ring main? It does look as if you are in uncharted territory with having your control and metering in different places. You need to measure the whole house current at the meter/consumer unit, and the voltage (so that you can know the direction of power flow), but you need either a separate power feed from the c.u. to the immersion heater, a low voltage signal cable, or else a radio link is called for. The emonTx already uses radio to send the accumulated readings, you would be sending only an on/off signal so it should easily be fast enough. (Radio implies two Arduino's - one measuring and doing the sums then sending the on/off signal by radio, the second simply receiving the radio message and switching the heater on/off). However, I'd recommend the simplest/cheapest/safest way: a low voltage (could be telephone/burglar alarm) cable from where the monitor/control will be to where the power switch will be.

Initial Shopping list:

Arduino and necessary components or emonTx (with/without radio module - your choice),
current transformer - probably the 100 A version as we have full data on that,
voltage transformer for supply & monitoring,
small piece of stripboard for a few additional components,
suitable box big enough to allow for expansion,
various connectors and hardware etc.

To that you will need to add the power switch. This is likely to be a small piece of stripboard with a handful of components, a triac and a metal box that doubles as a heat sink. Plus of course mounting hardware, connectors, terminal blocks etc. If you went for the radio option, it would need another Arduino, power supply etc as well. Now is probably the time to think where and how you will do the power switch. Ideally it will not be in the bathroom because there you run foul of Part P of the Building Regs that require all work in specified areas to be carried out by registered persons (at great expense no doubt). Don't ask me to elaborate: I've fairly recently come across 3 jobs done by allegedly registered persons/firms - one had minor breaches of BS 7671, one was potentially a fire hazard and one had rendered every bit of the electrical installation in the whole house actually dangerous.

"It does look as if you are in uncharted territory with having your control and metering in different places."

Robert, I have had my metering and control at opposite ends of my house running well now for a few months so I would consider this method quite proven.  I was getting approximately 100 transmissions per minute before adding in temperature sensing which has now slowed things down a bit, but as my electricity meter indicates green when exporting I can see that all is as should be.

I'm using a second emontx for the PWM control as I intitially intended temperature sensing on this board but now not used, a croydom proportional SSR and I have a filter (not snubber). 

In use, on a decent day we hit the software temperature limit of 63.5 deg around lunchtime (a DS18B20 sensor stuck to the tank metal at heater height, set just below the tank stat max cutout) , we can then hot fill a kids paddling pool and have hit the limit again by teatime - just proven last week.  I fitted the temperature sensors as the tank stat hysteresis was too great, and I can now maintain temperature control to I think 0.1 deg C.

I have intended to upload the sketches which I have to say are still in beta although they have proven themselves already, but I keep adding tweaks and extra's as I go along so have to continue testing.  If anyone is interested in seeing the sketches and hopefully improve on them then I will upload them as they are.

What's the rated power of your PV installation?  That will put your timings into context.

eof, as your post was immediately after mine I'll assume you directed your question to me?  I'm unsure though why my PV size would affect transmission timings?

I have a 3.84kW system.  The transmission rate I mention, although not clear on re-reading my post was meaning the amount of wireless transmissions that my emontx which does all the measurement on sends.  I use 3 ct's to measure PV generation, house use and heater power all in the meter cupboard.  I have a type 1 setup and have taken my immersion power from an additional MCB installed in my generation DB.  I think the amount of transmissions is related to the time taken to get through the sketch loop as every 40 seconds I call for temperature sensor readings, and this can add an extra second onto that loop cycle.  This has not been a big deal as in the extremely rare event I switched on a high load appliance at precisely that second then I'd guess it may cost me less than a penny in import before the system cut back on heater power.

It won't affect transmission timings - I meant your timings of how long the tank takes to warm up.

Logic,

I hadn't actually picked up the fact that you were using  distributed system. How are you transmitting the data - wired or wireless?

(In any case, I think Calypso-rae's system requires a bit more than 100 messages per minute).

The best thing I can do for anyone to see the performance of my tank against solar output is to link my website.  This as you'll see is only a 'get it online' quickie as I'd never done a website before.  It is a Megaflow 210 litre tank previously heated indirectly by oil, and now using the single 3kW heater.  There has only been a couple of times during some foggy days recently that the oil has had to be used to top the tank temperature up, but in any case if the temperature is not greater than 35 deg by 5pm the oil will boost the temperature (the boiler control is thermister rather than thermostat control).

See http://tinyurl.com/c3mltse which should take you to the multigraph, and clicking the 'energy' link top right should display the live dials.  (Please ignore the kWh totals as I'd previously deleted/reset the feeds, and that emonsms glitch a week or so ago added a few thousand to my consumption - I've yet to figure how to do addition/subtraction in the emoncms to correct this) Power 1 is Solar, Power 2 is Useage and Power 3 is the Heater.

Robert, for the transmission I use a 2nd emontx to receive the existing wireless transmission from the emontx fitted at the consumer board.  That then uses the standard kit from openenergymonitor to do the rest (plus a 3rd ct - not strictly necessary, but I felt more confident taking an actual heater reading rather than derived value from the generation minus consumption calculation ).

I am also close to completing my system that has distributed measurement and control.  I have an emontx with 2xCTs measuring consumption and generation at the consumer unit.   As my boiler has been spurred off my immersion feed at some point (not ideal I know), I cannot control my immersion at this point, so I will have a second emontx, with a CT (not entirely needed but I do want to know when the immersion thermostat has turned off, and how much energy I actually dump into the water tank) in the airing cupboard, controling the immersion via a Crydon SSR.  I also have an emonlcd and an emonbase uploading to emoncms.  This has all worked well on the bench driving a kettle, so I've now started boxing everything up ready for installation.

Lloyd

Logic,

what are you using for your filter?

Thanks

Lloyd

Hi All,

Have been watching this thread for a while, and it's great to see the open-design philosophy in action.

"...you need either a separate power feed from the c.u. to the immersion heater, a low voltage signal cable, or else a radio link is called for."

Robert - I was wondering if anyone had tried using X10 to transmit the signal from the metering to the controller?  Although it's another interface to learn and integrate with, it might be a further alternative to consider.

Andy

andyjc 

"I was wondering if anyone had tried using X10"

I have not noticed it mentioned, and having just read up on the protocol, I think speed (or lack of) would be an issue. (Reason: 3/4 second for X10 to transmit one command? - The meter 'window' is 3600 J and 1 immersion-heater-second = 3000 J, so not much room for error). My thinking was KISS. The output is most likely driving a LED in an opto-isolator, so a simple twisted pair (with adjustment to the current limiting resistor if appropriate!) would be simple, effective and not suffer the propagation delay that a radio protocol might. And it would be cheap.

@Riggers: I will probably required a wireless switch to the immersion heater, as the tank and the PV system are on seperate floors (tank on first floor, PV in internal garage - ground floor).

Our DHW tank and Consumer Unit are also on separate floors.  To avoid having to run an additional lead upstairs, I've connected into the power cable for the immersion as it leaves the CU.  All that was needed is a 4-way junction box (for L, N, E and S/L), and a spur to the ac controller which is mounted on the garage wall.

Loyd, The filter I got was http://uk.rs-online.com/web/p/products/292-824/ and this I chose this after consulting Shaffner about my application.

Robert, if 100 messages per second isn't enough then I'd be fairly confident that a stripped down sketch only monitoring the incoming meter tail would be faster, as I was running the full emontx sketch as got from this site with an additional ct.  I had to remove the delays and then modify the emonbase to prevent as many updates to emoncms.  Unfortunately I'll not be home for nearly 2 weeks to be able to test htis theory myself.

calypso_rae, I too originally thought of installing the controller next to the consumer board but I wondered about the effects of it then being classed as fixed installation.  While I still have it mounted on a metal plate for testing and fed from temporary cable (lazyness as I have everything to put it in) I was intending powering it from a dedicated socket outlet with the thought of it then not being fixed installation i.e. exchanging the immersion FCU for a single switched socket outlet.  Maybe somebody could correct me on this?

The feed for my immersion still comes out of the CU and heads off upstairs to the immersion, as before.  It now, however, goes via an additional approved switch on the garage wall which is controlled by an Arduino.  If anyone were to ever complain, I could simply remove the extra wiring and the circuit would then be just as it was before. 

Personally, I don't see that it makes any difference whether you intercept and control the immersion's supply cable locally to the CU, or do so at the far end and control the switch remotely via a radio link ;)

logic,

I'm not worried about the sketch - we know the standard sketch measuring voltage and one current I goes at about 50 samples per cycle, and I agree that it looks as if the radio link is quite capable of enough speed, provided the message length is short. As only On/Off information is needed, 1 byte should be enough!  But I was worried about the transmission delay between the data going in to the rf module and coming out the other end. There must be a delay that arises from the overhead due to the radio protocol (node and network numbers etc.), and I'd like to figure out a way of testing to determine that.

"I haven't tried to cost his project, but I think it will come in at a long way under £100 all told"

I very much doubt it especially when VAT, CT, SSR (especially), fuse, isolator, case and an EMC filter is added in. Although Arduino items are cheap the peripheral components needed soon add up. It does depend if you are making something barely minimal that will work or if you are making something safe that might pass CE/IEE standards. The latter probably adds £30 plus.

The triac will be switching on/off for whole cycles at zero-crossing points, so the EMC should be minimal. 

I've not attempted to work out any costs for a completed unit.  The main components will be the Arduino, current clamp, triac and trigger which together should cost no more than £30.  I have no idea about the cost of other components, such as the case, fuse, filter, cable clamps, connectors etc. 

For a DIY build, the total cost should be fairly small; far less than when having to purchase a pricey 3rd party AC Power Controller.  But for a commercial version, there could well be additional factors that would need to be considered.

I can't find a current transformer for significantly under £10 at the moment.  It's cheaper to buy an "unfashionable" complete energy monitor.  I bought a Current Cost unit for £5 all-in, which is actually a lot more impressive than the Owl in terms of resolution and update rate (and has the coolest-running linear power supply I've ever come across!).  Unfortunately it doesn't have the Owl's easily accessible totals which I find useful.

I bought a couple of standard 30A current-clamps from Cool Components as featured in the Building Blocks section.  I think they were £6.99 each plus postage (but out of stock when I last checked ...)

Yes, I saw that link at some point.  As far as I remember, once you've paid postage you don't get much change from a tenner.

Turns out they were £6.99 each ex-VAT, plus postage.  Total bill, £19.97 

 Wow, what an interesting thread!  I have been reading up like mad on how electricity works with the sole goal of diverting as much of my surplus PV Solar electricity to my 3KW immersion in my 200litre storage tank.  I'm grasping the concept of using some sort of PAC to control the voltage going to the immersion and I think I can cope with the programming of the arduino (well, at least hacking someone elses code).  What I can't grasp at the moment is how to control the output of a PAC.  Calypso_rae - I note you mentioned sharing your sketch at some point - I can't see it? Would you mind PMing me a copy so I can try to get my head around the control script for a PAC.

Welcome to the forum.

Calypso_rae has not published his sketch yet, so you won't find it. But it doesn't need an external programmable controller, the Arduino handles everything with the addition of a small handful of components to switch the power. It was only put into operation a day or two ago so it is still being tested. It is his intention to publish as soon as it is proven.

As they say: Watch this space...

Hi gsbeaton,

Code for my Mk1 PV Router was published on this thread on 28th May.  It uses the standard sketch for V&I to determine real power, the AC control being entrusted to a robust item of 3rd Party kit.  This was an excellent way for me to start, and plenty of hot water has been generated during the six weeks (it feels much longer!)

Since then, I've been playing around with other ideas including the use of a triac & zero-crossing trigger to distribute the power.  I've also developed some new code which appears to provide significant benefits over the standard approach.  As Robert has mentioned, my early trials have produced encouraging results.

I think it would be helpful for this new development to be presented in a way that is consistent with existing EOM code.  Maybe it should go into a library?  Guidance has been sought ...

 Agree with what you said - measure the export at the point of power entry. Looks like the immersion switch from http://solarimmersionheaterswitch.co.uk/ uses the same technology - as they have only one sensor clipped on to the incomming grid lines to measure house hold load and production. 

Had a chat with the support guy last day - looks like a Mark III version with differential controller is on the way - dont know if that os true.

Watch this space: http://openenergymonitor.org/emon/node/841

(Or this is where Calypso_rae's sketch alluded to above is published).

 Hi,

I have a 3 phase system with 3 4kw arrays and I want to measure export on each phase and use the surplus to control a battery charger, with inverter to put surplus back at night or when load is greater than solar delivery, and dump power into an immersion when batteries charged. Since house consumption is only about 10kwh per day and panels generate75kwh per day it should be possible to avoid importing altogether, apart from a trickle or on high demand.  Principle is fine, practice that works reliably less so.  Am looking at either a UPS with a separate battery charger assuming I can't configure internal one to control charge current (have found one that will run the house) or a number of grid tie inverters and chargers.  Ideally want intelligent switching control but buying some cheap chargers and inverters and switching in with relays to nearest 500w would do.  Any thoughts on best starting point for monitoring and control logic.  Anyone with an interest in building the project let me know.  Cost of working system should obviously be sensible compared to cost of recovered electricity.

The unit sold at www.totalpowershop.com , their TP-S610A starter kit (they call it a Renewable Energy Monitoring Unit or their REMU I) is able to handle a 3 phase system.

I use it to monitors how much energy my house is consuming and how much power I am getting from my solar panels or any other renewable energy source. I've set it up so it automatically switches on my 1kw immersion heater (and up to 2 other appliances) when enough surplus energy is being generated.

The 1 phase package is only £199. I think it will pay for itself as well, although it may take a year or two to get my money back. In the mean time it will make sure that my home is not consuming anymore power than I want to. The control unit will also record my home’s energy production and consumption and download it to a pc.
 

The standard circuit for the MOC3041 zero-crossing trigger has three resistors, one of which is 330R.  As I didn't have one of these handy, I decided to use a pair of 680R resistors in parallel.  I soldered them to the board, and it worked fine.

A couple of hours later, while doing something else, I suddenly realised that I'd wired them in series rather than in parallel.  I've since replaced them with a 150R & 180R in series, and this combination works fine too.

Talk about a wide margin of tolerance!

(As far as I'm aware, the z/c trigger/triac idea was first posted on this thread)