simple consuming/supplying indicator

[pj]

(Now that I've caught up) This seems like a good basis of a circuit to soak up excess PV generation by driving an immersion heater via a PWM output into an SSR? It could fine tune the PWM to keep the system just exporting, and rapidly track the clouds as they come over.

[ericw]

PJ

It was intended to be the first step in producing a circuit to soak up excess energy. However you cannot use PWM (or phase control) to vary the load. This has been discussed in posts a couple of years back and the reasons are set out here.

For the sake of illustration assume that you have a load that takes 10 amps and your PV is producing 5 amps.
Simplistically you would switch the load on at a 50% mark space ratio to use the PV power.

However if you look more closely, when the switch is on then there will be 10 amps flowing in the load, 5 amps from the PV & 5 from the mains.
When the switch is off 5 amps is fed into the mains from the PV. So all things being equal the net power from mains is zero.

The gotcha is that while old mechanical meters may average it in that way, modern electronic meters don't. They measure (and charge you for) the power that is fed from the mains but ignore anything fed back into the supply.
So the net effect is that you are actually paying while you are dumping energy.

I believe the only solution is switch in varying amounts of resistive load. It has been suggested elsewhere in the forum that there are several advantages in using a set of wireless remote controlled sockets, hacked at the transmitter end to do the switching. This looks interesting, but the the constant switching to track changes in input and/or load will give any relays in the controlled sockets a hard time, probabely shortening their life.

The other thing which has not really been addressed assuming that you will be using it to heat DHW is actually geting hold of suitable heaters and actually installing them.

[ericw]

Thanks everyone for the interest I'll try and answer points that have been raised.

Phase
This is probably better described as direction. The output of the CT is nominally in phase with the mains voltage when the power is being fed into the supply and 180 degrees out of phase when power is being used (or vice-versa if the CT is connected the other way round)
By measuring the output voltage of the CT at the peak of the mains voltage you can determine which of these conditions applies.
The phase input is a relatively high voltage (from a high impedance source) which causes the PIC input to rapidly swing from rail to rail (clamped by the internal ESD diodes), these transitions are used as a time reference to work out when the peak of the mains voltage occurs. On the battery circuit the capacitve coupling gives a 90 degree phase shift and so the peak mains voltage occurs at the transition of the phase input, for the mains circuit you need to intoduce a 5 millisecond delay.

Current sensing
The overall accuracy is determined by how much effort you put into current sensing.
The scheme descibed by Sean is the Rolls Royce of methods but does require an accurate sample of the mains voltage to work. As I need to have a system that didn't have a direct connection to then mains voltage (meter cabinets don't usually have sockets in them) I could not use this method.
If anyone is interested in real accuracy I suggest that the Arduino based system described here http://openenergymonitor.org/emon/input with the addition of my direction sensing would be a good place to start.

The HalfBee Sean mentioned uses the 'Ford' method in which the current is sampled 256 times per cycle, squared and the RMS value. By assuming a power factor of 1 you can then calculate the power involved I think this asumption is probably OK for most domestic whole house situations.

For the demonstator I tried to get the total supply current as low as possible (< 1/2 mA) to give maximum battery life and so used a 'Trabant' method of just doing a single measurement at the voltage maximum. With hindsight it's probably worth revisiting this using the HalfBee code and possibly moving to bigger (D size?) battery.

Display
Personally I think that once the novelty wears off then there is little point in having a real time display. I don't see either anyone rushing to turn off things when the sun goes behind a cloud or SWMBO accepting being told she can only have the iron/washing machine/dishwasher on if the meter is well into the green. However as other people may differ I'll continue. The one enhancement I would have thought most valuble would be to give it a remote wireless display. While it maybe possible to hack into one of the freebie power monitor systems its probably easier just to use a simple RX/TX pair (around £10@ Maplin) just to send a couple of data bytes to a mains powered display, which can be as simple or complex as desired. [Iain http://www.ti.com/lit/ds/symlink/lm3914.pdf contains the circuit for a 20 step centre zero meter]

I intend to re-visit the HalfBee power meter code and using one of these it will be possible to plot/log results using a one wire system as a psudeo temperature from -40 (4kW fed to mains) to +120 (12kW consumed)


Code
I'm quite happy to publish the code I used but I think it may be worth waiting a while to see if any more useful comments appear that could be incorporated. The code currently is in PIC assembler which rather restricts its use to anyone who has the faculities to burn the chips. For a wider audience it would be possible to use a PICAXE which is still fairly cheap and has its own programming environment.
For anyone champing at the bit the present program used is :-

Wait for phase input to go high
     If mains powered version wait 5 milliseconds
Measure voltage & store on resistor end of CT
Measure voltage & store voltage at other end of CT
Subtract the two voltages
Add in offset to give PWM mark/space ratio of 50% for zero current input
Move value to PWM generator.

Build instructions

I'll work on a parts list and the build instuctions.

[jotec]

Many thanks for this Eric.
My use for the circuit would be to alter the revs, hence output, of my CHP prime mover. I would want to drive a servo so that when I am outputting to the grid the revs are reduced. As I claim ROCs I get no export tariff.
The CHP is grid tied with a SMA Windyboy.
Dick

[julian]

I would be interested in the project for exactly the same reason as Dick - to control a diesel generator's output when the usual feedback pathway has been severed by the GTI.

I had considered using two clamps on the load and GTI output, comparing between them, and inferring the inport/export status from them, but, from my brief skim through the thread above, it seems like the above unit would serve well also : )

Thank you,
Julian

[pj]

Eric,
I get your point about PWM not working. From descriptions available on the internet, looks like new meters are also doing per cycle power measurement much as Sean described.

Restricting the current (power) in an immersion heater seems an impossible thing to achieve at a reasonable cost. So I agree, to use up excess PV generation needs a number of loads, plus switching between them. To this end, I've contacted an immersion heater maker, and they can provide a single standard fit (2 1/4") immersion, but containing three separate elements, and with each element at a different power rating - 500W, 1kW and 2kW. Thus with three pins from your PIC, plus SSRs for reliability, you can select a load from 0 to 3.5kW in 500W steps. The cost quoted for the heater was £80, inc vat and delivery.

Now we just need your fully developed controller, and for those who wish to use excess PV for water heating, maybe we have a viable solution?

[Davo]

To this end, I've contacted an immersion heater maker, and they can provide a single standard fit (2 1/4") immersion, but containing three separate elements, and with each element at a different power rating - 500W, 1kW and 2kW. Thus with three pins from your PIC, plus SSRs for reliability, you can select a load from 0 to 3.5kW in 500W steps. The cost quoted for the heater was £80, inc vat and delivery.

Hi PJ
Can you supply details of the element manufacturer - the 3 wattage soluton looks really useful.

[ericw]

PJ,

That looks really useful information. How long are the elements?

I was trying to devise a system using towel rail heater elements but 800w seems to be the max generally available.

I think the next problem to consider is the optimum position of the heater(s) to try and quickly get hot water at the top (short elements) but also be able to heat the whole of the cylinder (long elements). The equivalent to this problem has already been the subject of much discussion in the solar thermal part of the forum.

It may be something working on the same principle as the Willis immersion heater could be the answer and would also be a fairly easy retrofit.

I'll work on the circuit, but as it is difficult for me to run wires from my meter cupboard to the cylinder it will probably end up with a simple wireless link. Personally, like you, I would tend to build a unit with solid state relays but I think that for the less electrically competent using commercial wireless remote control sockets ( 3 for <£20 on Ebay) might be a preferred option.

[Richard Owen]

you can select a load from 0 to 3.5kW in 500W steps. The cost quoted for the heater was £80, inc vat and delivery.

If you could chose to connect them in series as well as parallel you would have lots of interesting intermediate steps as well.

[ecogeorge]

3 phase elements (readily available) in sizes of 6,9, and 12kw  are usually 3 elements joined together ie 3x2kw, 3x 3kw or 3x 4kw.
Can get these at reasonable prices if interested.
Elements can be run via 110v "yellow" transformer to reduce load .

[pj]

Davo, the company I spoke to was www.tpfay.co.uk

Eric, the elements were 11", but I think they can be made to any length, and any mix of wattages - see their web site for more details. As far as where the element fits, this thread http://www.navitron.org.uk/forum/index.php/topic,13537.msg151747.html#msg151747 talks about an accelerator tube. If it works well, then the element could be retro-fitted in place of the existing at the top, creating hot stratified water at the top of the tank, but being fed by cooler water from below.

Richard, I'm not sure what the control matrix of SSRs would look like to give selectable parallel plus series connections, but if it could safely be done (i.e. no shorts straight through SSRs) then yes this would give even more granularity.

ecogeorge - yes, I think this is based on a three phase head, but with 240V elements. The key here is to have different wattages in each element, giving a range of powers. For a domestic environment, the max needed is 4kW per the max PV array, which led me to the values above.

P.S. Anti Spam Note - I have no relationship to the linked company, other than as a potential customer Hope this is OK Samantha?

[Davo]

Davo, the company I spoke to was www.tpfay.co.uk

Cheers Eric
An extremely flexible solution at a very reasonable price.

[Automan]

This thread may also be of interest on the new Eco-Eye product going on sale in a few days...

http://www.navitron.org.uk/forum/index.php/topic,15900.0.html

http://www.eco-eye.com/products-smartPV.html

Automan.

[ericw]

If anyone is interested in building a remotely controlled 3 load system, then the local ADSA is selling a pack of 3 remote radio controlled sockets for £10
As the TX has both on and off buttons for each socket it looks like an promising thing to connect to a current sensing interface.

Downside is that the sockets loose their ID after a power out so would need them to be re-learned.

[Davo]

If anyone is interested in building a remotely controlled 3 load system, then the local ADSA is selling a pack of 3 remote radio controlled sockets for £10
As the TX has both on and off buttons for each socket it looks like an promising thing to connect to a current sensing interface.

Downside is that the sockets loose their ID after a power out so would need them to be re-learned.

I have a set of these and they have proved very reliable. Not only that, I was very impressed by the range of operation (including through several hard walls). I wouldn't have a clue how to 'hack' them though.
Not too much of an issue re-learning after a mains failure. They do need to be plugging in 1 at a time when re-learning though (else they will all operate from one button at the same time - which may of course be another use).