Archived Forum

To get the ball rolling I have provided a short list of possible ways to use excess PV power.
1.    Tumble drier
If the motor is left to derive power directly from the mains it should be possible to modulate the heating element using either burst or phase control.
2.    A second immersion tank.
 There is more than one way to do this and I shall post some more later.
3.    A huge highly insulated heat dump water tank buried in the garden
4.    Batteries
5.    A large flywheel with a generator/motor attached
6.    A large bank of ultra capacitors
7.    Under soil heating in the greenhouse to allow exotic fruits to be grown

One of the simpler ideas that I thought of is to insulate the underside of the bath by simply filling the space behind the bath panel with loft insulation material and make two or three expanded polystyrene inserts that form a bath sized insulator that simply floats on the surface of the bath water. Now run your evening bath at midday, put the covers on and with luck it should have cooled to a comfortable temperature by bath time and the DHW tank should have reheated using the afternoon sun.

If you are considering kooky "hi tech" solutions such as flywheels and ultracapacitors, then come down to earth and instead build a system that can store and dispense energy more efficiently than water. Heat oil instead, it has a much higher upper temperature limit. Super-heated oil could be used to make domestic hot water on demand, or without tempering, it can be used in oil-to-air heat exchangers for space heating.

Control and distribution systems for hot oil are an off-the-shelf solution. Try building your own super flywheel and find out how easy magnetic bearings inside a nearly complete vacuum are to accomplish.

"One of the simpler ideas that I thought of is to insulate the underside of the bath by simply filling the space behind the bath panel with loft insulation material and make two or three expanded polystyrene inserts that form a bath sized insulator that simply floats on the surface of the bath water. Now run your evening bath at midday, put the covers on and with luck it should have cooled to a comfortable temperature by bath time and the DHW tank should have reheated using the afternoon sun."

I'm sorry Brian but have you taken your medication?

All you need to do is set your tank stat to max, this will heat your water to around 70deg c, assuming it's well insulated that is a foamed tank with an extra jacket on it your water will still be around 60-45deg c the next morning more than enough for 3 to 4 baths and keep the legionella at bay. What you have described is a second hot water tank and called it "your bath".

Stratification is our hero, the best way to more hot water is a longer immersion heater, mine is 36", not that I'm boasting.

As for alternative uses for solar after the water is hot, my favorite is to get an electric car (NOT a Sinclair C5 Brian!) and use the excess to charge that. The added advantage is "sorry boss can't come in to work today, it was cloudy yesterday.

The biggest drawback with solar power is you get most of it when you least need it, when the days are long and you want cool showers, I'm considering installing air conditioning for when I have excess or an electric powered barbeque.

I do hope your post was meant as a bit of fun, if not, oops sorry!

If not "hydrogen cells" that's an option.

I think fronius has missed the humor. ;-)

P.S space heating is pointless for the summer months, that's why I'm thinking air con.

GOT IT!!!!!

I'm rigging my system to charge the flux capacitor Marty!

Richmc

How 'Bout an extra few Beer Fridges !

Rich said:

All you need to do is set your tank stat to max, this will heat your water to around 70deg c, assuming it's well insulated that is a foamed tank with an extra jacket on it your water will still be around 60-45deg c the next morning more than enough for 3 to 4 baths and keep the legionella at bay. What you have described is a second hot water tank and called it "your bath".
Something doesn’t stack up here!
My bath is 1700mm long and a typical bath fill is about 150mm depth of water at 55 deg. This amounts to about 40 litres of hot water then cold is added as required.
To produce 3 to 4 bath fills will require 112 – 160 litres of water at ~ 55deg. The capacity of an average 900mm immersion tank is 117 litres so either you have a very large tank or there is some magic going on.

You also say:
 Stratification is our hero, the best way to more hot water is a longer immersion heater, mine is 36", not that I'm boasting.
I guess you mean stratification is our enemy and this is why you have such a long heater?
As for setting the tank ‘stat to max I don’t like the idea of skin burns just by putting my hand under the tap.
http://www.ciphe.org.uk/Global/Databyte/Safe%20Hot%20Water.pdf
Yes, hopefully my post has a little bit of dry humour but the fundamental point is still true. Even a 4kW PV system will be returning most of the generated power to the grid during full summer sun when the small immersion tank is up to temperature.

My preferred use for 'extra' surplus PV would be to have a box which can plug into any ring-main socket yet only supply power to my secondary load when instructed to do so via an RF link from the router.  I've attached a sketch to show how this might work.  At present, it is the lingo which is my main hurdle. 

In my case, the secondary load could be a storage heater, or maybe just a small space-heater over which clothes could be hung to dry.  If possible, I would like to source any new components from the OEM shop.  But I think that most of the things that I would need are from JeeLabs.

I will put a heater on it and in the summer the heater will be outside.

Robin

That looks very similar to what I have been thinking about.

In my case the Arduino on the left is a stripped down emonTx running Martin's sketch. The emonTx is the ideal platform as it supports the RFM12B directly.

One problem in going this route is I guarantee you will not resist the temptation to add an emonGLCD as this is such a wonderful little toy that gives you a window into the inner workings of your system. You can take it from room to room and keep an eye on what is happening. I even have two of them.

My target extra dump load is the tumble drier which lives in the shed at the bottom of my garden. Some initial test have shown that the radio signal gets there OK so it all looks feasible. The only problem at the moment is figuring out how the tumble drier currently modulates the power to the heater. The plan is to allow the motor to run direct from the mains and control the heater power with the remote device.

\before we get back to the serious stuff...may i have my folly and suggest charging the dilithium crystals!

However this thread has had its humorous replies there lies a real and challenging task. In the spring before it gets too warm then space heating is a real answer!.... Robin  Could you write some code to fire a second triac once the first one had done its deed and heated the DHW, and then fire a 2 - 3kW  resistive heater element ( storage or other wise ) and then in the summer activate / deactivate a relay to say power a dehumidifier  air conditioning unit or a a battery charging  unit, with hysteresis?... this could be really useful.... What say you all?

if i had a swimming pool that would the great just dump all the spare energy into that, however i don't.

I like the  idea of a second triac for tumble dryer, i have been playing about with it on and off, i must confess i don't have a spear one to play with.

I also toyed with the idea of a standalone unit for the kettle, you just have a box in-line which you can dial different power levels 750w, 1250w, 2000w and full (3000w) i have never been in a rush when making the teas, but if was a day like today i could have for example dialed the kettle box to 1250w and the kettle would have taken a bit longer to boil, but i would have been completely free. you could use this for any resistive load, irons, kettles, toasters  etc...

i ultimate goal would be for a battery bank that is charged by the PV system and then used at night. but i think thati s a little way off yet

Kettle circuit is definitely one to consider as I have often unplugged the immersion circuit in the garage and connected a kettle for a non-alcoholic brew.. for a free cuppa.... although I often have the other kind in the evening!

We have a 900W kettle which "feels" right when the sun is shining, and you're not in a particular hurry for that cuppa.  But it would be much better to have a standalone RF-controlled box into which you can plug whatever load you wish.  Then you could guarantee that no power is being wasted (but maybe not a good idea for the kettle).

If I didn't already have two Arduinos and a PCB-based router, then I'm sure that an emonTx would be the best way to go. 

Does anyone know how much delay would be incurred over the RF link?  If less than around 10mS, it would actually have no overall effect because the zero/crossing trigger has to wait around 10mS anyway for the next z/c point. 

Yes RF link node would be the most practical for the kettle! Your thinking is well ahead of mine!

......but when it comes to the missus... she wants it..now.... so a lot of education(most eco friendly way) or... an-solar-excess -hot kettle may be the answer... after-all in the summer (sunshine positive) we may have more energy than we could use!......or want to export back to the grid!

....need to think about controlling excess boiling and steam generation!

......Don't need to worry 'bout until at least April... then maybe someone has the solution.... thermostat @ 80deg cent....... maybe........ i'm just as cookoo as the starter of the thread......?..... No intentions to offend here.... just combining a serious thread with a little humour !!!!!.......as a few others have done! Please don't take this as offensive..... it is a breath of fresh air to break from this serious forum, and have a few seconds..of light fun.......but after that it is down to the real stuff!

A very well insulated swimming pool and a heat pump connected to the central heating gets my vote. You need 4.18 J to raise the temperature of 1 g of water by 1 °C.  If you have 50 m³ of water in your swimming pool, you need about 1400 kWh to get it from 0 °C to 100 °C. You should do that in a summer. You then take it down again to 0 °C with the heat pump come the winter.

Just don't go swimming until after Christmas, unless you want to end up like a lobster. 

 "........and then fire a 2 - 3kW  resistive heater element ( storage or other wise ) and then in the summer activate / deactivate a relay to say power a dehumidifier  air conditioning unit or a a battery charging  unit, with hysteresis?... this could be really useful.... What say you all?"

Very good idea Mark, especially as its almost word for word taken from one of my posts on Talking Solar! ;-)

charging the dilithium crystals is not really practical as getting the matter/antimatter mixture wrong would have devastating effects on the neighborhood, might even result in some light flicker.

"I guess you mean stratification is our enemy and this is why you have such a long heater?
As for setting the tank ‘stat to max I don’t like the idea of skin burns just by putting my hand under the tap.
http://www.ciphe.org.uk/Global/Databyte/Safe%20Hot%20Water.pdf
Yes, hopefully my post has a little bit of dry humour but the fundamental point is still true. Even a 4kW PV system will be returning most of the generated power to the grid during full summer sun when the small immersion tank is up to temperature."

Stratification is defiantly our friend my plumber friend has solar water heating and the back up immersion is in the bottom of the tank so if he used a Mk2 it would on a poor day generate a whole tank of lukewarm water whereas in mine I have the top of the tank filled with water at a useable temperature.

As for your concerns about scalding, there just me and the Mrs at home, if we do use hot water in the early evening we are aware, but it there were kids around it would be quite easy to install a pre mixer valve to the tank outlet. I work in the care sector and see these in many homes.

Yes I totally agree about the wasteful export during the summer, I did ponder charging a battery to feed an inverter to use on the lighting circuits but thats for another day.

I think this thread can usefully contribute in several areas:

- How to use surplus power?  There seems to be no shortage of ideas here!

- How to call on surplus power where needed?  A viable RF control method is my main interest.

- How to maximise benefit and/or minimise loss?  By monitoring the meter's LED and reacting accordingly.

Getting the router to power down at night would seem a smart move.  Has anyone tried this?

"Getting the router to power down at night would seem a smart move.  Has anyone tried this?"

Just out of interest I've tried this, 2 Watts. What was that phrase "picopence". powering electronics off and on can cause thermal shock especially in semiconductors, far better to just leave the beastie running.

When the sun comes out, I'm going to set up the saw bench and cut some logs ;-)

Robin said:
- How to call on surplus power where needed?  A viable RF control method is my main interest.

Presumably you are using a 5V Arduino Uno running MK2 software as the controller?

The obvious candidate RF device is the RFM12B which is what you have shown in your earlier post with the schematic. This all looks simple enough except that the RFM12B is a 3V3 device and does not have 5V tolerant I/O. This means that to use it with the Uno will require level shifting interface devices.

The alternative is the emonTx but this has issues with the cross talk between inputs that I have avoided with extensive modifications I would not encourage anyone else to copy.

Then there is the question of what to use at the far end. In your sketch you have shown the JeeNode v6 which is probably the best option (can someone advise if I can purchase a 433 MHz version in the UK please?)

So, I am half way to adding a remote load and will probably go the Jeenode route. I have in mind a box with an antenna and a mains socket into which the controlled load can be plugged.

Just to make you guys green with envy, here in Aus we have the opposite problem.  Because our feed-in-tariff is 2x what we pay for electricity, we want to run all our appliances after the sun has gone down.  The energy coming off the roof is liquid gold if you can export it, but if instead you use it, it's just ordinary electricity....

http://emoncms.org/dBCC/mainpower

dBC said:

Just to make you guys green with envy, here in Aus we have the opposite problem.  Because our feed-in-tariff is 2x what we pay for electricity..........

On the contrary, we get paid for generating it and get to use it if we have a clever controller! We keep our cake and eat it!

On the contrary, we get paid for generating it and get to use it if we have a clever controller! We keep our cake and eat it!

Very generous!

So you get paid to generate it no matter what happens to it, and if you use it, it's free?  Sign me up for that plan!

"Getting the router to power down at night would seem a smart move.  Has anyone tried this?"

Just out of interest I've tried this, 2 Watts.

My rig has two sources of mains power: an AC/AC adapter for the voltage sensor, and a 9V AC/DC adapter for the Arduino.  Although I've never checked how much power these items actually consume, I've always felt that it would be better for them to be off at night rather than on.  Maybe they do only draw 2 Watts, in which case that's around 30 pulses on our meter per night which could be avoided.  Or 1 kWh per month, for no benefit.

While the system is active and diverting surplus power, I'm keen to minimise any loss.   At present, the start of any period of surplus power is likely to involve at least 0.5 Wh of energy loss as the two systems become aligned.  By monitoring and reacting to meter events, I think it should be possible to minimise if not eliminate this source of energy loss. 

We also have a pair of transponders which communicate via the ring main.  One plugs into the inverter, and the other into our broadband router.  I think these take around 4W between them.  Given that I rarely feel the need to check how our panels are doing via the Web, these items seem to have no purpose other than to consume power (~2.5 kWh / month). 

Any suggestions as to how a consumption of 2W could be measured with any accuracy?

"Any suggestions as to how a consumption of 2W could be measured with any accuracy?"

At night with a digital multi meter checking the current drawn by the AC ref (naff all) and Arduino supply and adding them together, I know it's a bit sad but someone asked. 8.2mA (1.952W for anyone more pedantic than me!).

Unfortunately, measuring current won't tell you much.  Check out my microwave oven just being a clock:

http://emoncms.org/dBCC/erm

It draws ~0.4A but it's actually only burning about 4W.  It's almost all purely reactive.  It's pretty much a big capacitor when it's in standby mode.

I'm an even sadder pedant: "naff all" is a bad guess: 21.5 mA for the Shop Ideal adapter (unloaded, all primary and magnetising current), and 0.75 mA for the Shop 5 V USB adapter.

Adding them: 5.45 VA @ 245 V.

But what about power factor?  The primary resistance of the adapter is 692 Ohms, but the impedance (V/I) is 11.4 kOhms, so according to me, a power factor of 0.06 which would give you 0.32 W. Assuming the SMPS is unity p.f. (unlikely, it too is probably largely reactive), that still works out to be only 0.5 W.

(For those who are interested, the sum is P = I²R, but R is the real resistance referred to the primary - or a close approximation is the primary resistance only, 692 Ohms. The primary inductance, which makes up a major part of the impedance, doesn't dissipate power.)

Two points Robert, I'm not using the shop bought adapter and secondly I trust my Fluke 77 so 2W it is in my book.

No, you're saying 2 VA, not 2 W. I invite you to do the same sum... and I'll bet the answer is quite a bit less than 2 W.

I would if I could be bothered, but the point is there is no point in powering the Mk2 down overnight, so in the words of Catherine Tate................................

The adapter which powers my Mk2 rig is marked as 9V 200mA, so that doesn't sound as if it could be consuming any more that 2W.  But it is on 24/7, unlike our PV inverter ...

Easy. A good first approximation is measure the current, measure the primary resistance and do P=I²R. That gives the copper loss in the primary. Then add a bit for iron loss - I couldn't reliably estimate it so say under 10%. The Ideal adapter looks to be 0.33 W, so 3 kWh (= 1 bath) per annum.

"so 3 kWh (= 1 bath) per annum." 

Sounds about what I have except in leap years.

As I suggested above a second tank is another option although in my case the second tank holds heating fluid (water with an inhibitor – not corn oil!) which is used to heat the DHW via the usual heat exchanger coil.

A schematic diagram of the plumbing system and wiring is attached for the curious.

The problem with this arrangement is that the heat dump is highly stratified even though a long reach immersion heater is fitted.

Today I completed the wiring and testing of a mechanism to invert the 140 litres of water in the heat dump to ensure that the maximum heat absorbing capacity is achieved. The idea is that a thermostat at the top of the tank can activate the pump when the heat dump reaches 60 deg and this will push water round the bypass back into the bottom of the tank until the thermostat at the bottom of the tank reaches 60 deg or the thermostat at the top cools. The net effect of all this is a solar dump with a potential 15kWh capacity.

Brian.  Have you spoken to anyone else about your obsession for 5 hot baths a day? ;-) !

Sorry, but I just could not resist that!

In truth I have a plumbing system not dis-similar in complexity,  a boiler on the vented system, a boiler on the pressurized system and a Bowman flat plate heat exchanger between the two.

But that is all off topic.

How did you know I had 5 baths a day?

I guess it's moving from one extreme to another. A single immersion tank using gas as the primary heat source can't sensibly accommodate much PV surplus energy but when I built the two tank system I found that all I was achieving was a small amount of water in the top of the PV dump tank which I could deal with by manual intervention of opening valves and starting the pump but now it 's full automated and should provide an even distribution of heat.

But what to do with all the hot water? Hmmmm

One thing that annoys me is that both the dishwasher and washing machine are cold water only. What will happen if I connect to the hot water instead?

I suspect you will succeed in just heating up the plumbing between the hot tank and said appliances.

I think the best idea is to heat the swimming pool - but I for one, do not have one of those.

Any technology like a flywheel in a vacuum on magnetic bearings is likely to cost more than it would save.   Once you start doing the sums on these things you soon realize why it hasn't already been done.

PS. Is our spell checker an American?

PS. Is our spell checker an American?

I think so - it gets a lot wrong.

I think one of the problems is that people get too obsessed with proportional control, and exactly matching the generation curve.

I have an air-con system in the house here (big box on the wall outside, and three cassette units indoors). On a mid summer's sunny day it consumes about 4kw peak, and then settles down to around 1.5kw to keep the house cool. The panels on the roof generate 4kw, so it's a no-brainer - just switch the aircon on! 

There must be other, similar non-propotional uses. Not so good on a winter's day when there's minimum generation, but in the summer just fill your boots!

P.

One thing that annoys me is that both the dishwasher and washing machine are cold water only. What will happen if I connect to the hot water instead?

Our washing machine has hot & cold inlets, but only draws water from the cold one for low temp washes.  I presume this is to avoid overcooking the delicates.  If your wash-load can withstand the temperature of your hot water, then I suppose it would be fine to use hot only.  The cycle time should be reduced if there's no need to wait around for water to heat.

But then you'd be doing hot rinses too, maybe not a good idea.

Big writing in my (Miele) machine machine handbook saying not to connect it to a hot supply. What's really annoying is that the Australian version of the same model has hot+cold fill.

On the other hand, washing machines and dishwashers use so little water these days, depending on the length of your pipe runs it might be LESS efficient to connect them to the hot supply. 

P.

Not a problem in the summer Robin, when everything is solar powered, it's just a different heater. The hard part is getting the other half to wait till the LED is solid on.

Here are a couple of totally untried ideas for consideration. They come from a conversation with calypso_rae. The first is the one to consider if you think stratification is good, the second if you think stirring is good.

(1)
Put the PV dump load in a small vessel, a boiler. [The load could actually be a handwash unit, or a jug/kettle element, or both. If one is a jug / kettle element rated at 1 kW, flicker might not be a concern.]  Connect the tops and bottoms of the boiler and the hot water cylinder and include a low flow rate pump. Position the boiler above the cylinder and use a non-return valve, or anywhere and use a positive displacement pump, so that hot water can only flow from top of the boiler to the top of the cylinder at a controlled rate. Convection must not happen. It may be an idea to keep the speed of the water entering the cylinder to a minimum to prevent stirring, maybe by using a short length of large diameter pipe where it enters the cylinder. The water in the cylinder must stratify for this to work.
In operation, the dump load heats the water in the boiler and the pump delivers the hot water to the top of the cylinder. The flow rate is adjusted according to the energy available to maintain the output temperature at the required maximum temperature. The hot water displaces cold starting at the top of the cylinder and working down as more energy is delivered. Only when the hot water has reached the bottom of the cylinder and the pump cannot hold down the output temperature should the dump load cut out.

(2)
Add a second side-on element and thermostat near the bottom of the cylinder. Using two thresholds and two triacs in the controller, give priority to the top element. When the top thermostat trips, the bottom element takes over and starts convection. If the top temperature drops, the top element will automatically regain control and reheat the top. This should continue until the whole tank is heated more-or-less evenly.

#1 is a pretty good description of how my solar thermal hotwater system works.  In my case the "boiler" is the thermal collector on the roof.  My collector (and presumably your "boiler") need to be sealed and capable of dealing with mains water pressure (mains water pressure++ given the heating going on).  And as you say, you need to cut out when things are working too well, or your tank explodes.  Mine cuts the circulation pump off when the bottom of the tank hits 70C.  At that stage it just lets the collector boil (there's a relief valve up there to permit that).

For 1), there's a thing called a Willis Immersion which is essentially a heating element in a small tube. The original is here http://www.willis-renewables.com/immersion-how-it-works.htm but there are several copies on the market too. The original Willis isn't suitable for hard water areas but I believe some of the clones are OK.

P.

dBC: My collector (and presumably your "boiler") need to be sealed and capable of dealing with mains water pressure (mains water pressure++ given the heating going on).

Does that mean there is a heat exchanger inside the tank which is not shown in the diagram?  Or is the whole of your DHW system at full mains pressure? 

The SolarTwin thermal system is similar in concept to RW's scheme No 1.  If there were no FIT scheme, this may be the most efficient means of solar water heating.

Does that mean there is a heat exchanger inside the tank which is not shown in the diagram?  Or is the whole of your DHW system at full mains pressure?

Both systems are available, but mine is the latter (i.e. the whole system is at full mains pressure).  The other style uses some sort of coolant in the run up to the roof and a heat exchanging coil in the tank.   One is called open, and the other is called closed, but I can never remember which is which.

Mine is simpler because you never have to worry about coolant loss etc.    I think you need to use the coolant / heat exchanger model in hard water areas because the collector clags up (not an issue where I live... my 10 year old electric kettle still looks brand new on the inside).

One appliance that does not at first appear suitable for use as a dump load is the domestic electric oven. This is dangerous territory as it’s generally under direct control of the boss but let’s ignore that for the moment.
First, it’s a decent size load ~ 3kW. Second it’s often used during the day when surplus power is available. The problem is a control system is required that will not change the cooking time for a given item or we will be in big trouble.
I have noticed that when the oven has reached operating temperature the on period has less than a 1:1 duty cycle which means the power could be provided by PV if connected to a mk2 or eMC controller.
If the radio controlled remote that has been discussed above has its own processor then the procedure will be for the controller to initially connect the oven heating element directly to the mains until the target temperature is achieved. This can be determined by simply monitoring the current into the element. After this if the oven thermostat switches on the heating element then the power will be regulated in the usual way to ensure no cost cooking.
A fail safe timer will connect the heating element directly if necessary and the management will not complain.
Does this make sense?
 

If the oven is in use, then it will surely just add to the base-load.  Once it gets up to temperature and the element starts to cycle on & off, the Mk2 Router or equivalent will react accordingly.  PV power will only be surplus (and hence diverted) when the oven is not on.

I see the RF link as terminating in a dumb unit which only supplies power when instructed to do so.  So although there is a processor, it will only be used for listening to its associated RFB12B unit.  What an undignified role for such a powerful device.  Good job they're cheap!

The trick, as you say, is to try to use the oven during sunny periods if surplus PV would otherwise go to waste and you'd have to purchase power for doing the baking later. 

Robin said:

If the oven is in use, then it will surely just add to the base-load.  Once it gets up to temperature and the element starts to cycle on & off, the Mk2 Router or equivalent will react accordingly.

But not if the main solar dump has already switched off and there is less than 3 kW of PV available.

My 'quick fix' was to program the gas boiler timer to heat the water from 6am - 7am every morning, which provided enough hot water for showers etc first thing in the morning and probably quarter tank left for during the day - general use.

The Solar fed Immersion heater was set 5 or so degrees higher that the gas thermostat, so that it would always start regardless, and continue to heat the water throughout the day (sun permitting!). Providing that there was enough power generated, the gas boiler does not fire the following morning, and so on. (This happens very regularly in summer, but rarely in winter).

If the sun doesn't shine (which is unlikely in Doncaster), then the boiler heats the water again 6am - 7am.

It's not an efficient solution, but it helps!

https://cosm.com/feeds/34843

Robin

I agree that the oven would normally just form part of the base load and my previous description was not entirely clear so I shall have another go.

First some basic assumptions:

• Electricity is more expensive than gas.
• The immersion tank is normally heated by gas and only surplus PV power is used to heat it electrically.
• When up to operating temperature the 3kW oven only requires about 1.5kWh per hour to cook the food.

Now the conditions:
Let’s assume there is say 1.5kW available from PV.
Any surplus power is being diverted to the immersion by mk2.
The oven is switched on and demands 3kW

The new box of tricks that I shall call the remote controller has a series connected triac which is between the mains Live and the oven heating element and this triac is normally on therefore the user sees no change in the oven behaviour.

Mk2 stops diverting and the oven takes 1.5kW from the grid plus 1.5kW PV and begins to heat normally.
The remote controller has a CT on the cable inside the oven that feeds the heating element consequently it’s aware that the oven has just been switched on. The oven then reaches operating temperature and turns off the heating element and mk2 begins diverting again.

So far everything is as normal. Now the clever bit begins. The oven cools sufficiently to demand more power but the remote immediately switches off its triac and passes back information to mk2 using radio or even via a wired connection to indicate that a priority load is available. Mk2 stops diverting and sends a ‘go’ signal to the remote which enables its triac and a burst of energy is applied to the oven heating element until mk2 sends the stop signal. mk2 continues to send go and stop signals until either a ‘satisfied’ signal comes from the remote or the energy bucket fills beyond a threshold which would indicate that the oen has switched off and normal diverting to the immersion should be restored.

Basically we are now cooking for free instead of paying for half of it. Obviously we shall need to burn a little more gas as a consequence but gas is cheap we are saving electricity at the rate of roughly 1.5kWH per hour.
Hopefully I have been a bit clearer this time. If there are any holes in this idea I am sure that you will spot them – if you don’t then Robert will :)

Sounds like a good sales pitch for a gas oven to me :)

OK Brian, I copy.  You're relegating the normal dump load to second place when the oven is "simmering".  By setting your oven  to be the higher priority dump, with the immersion being second, I see no need for any reverse communication.

Once up to temperature, the 3kW element normally cycles on and off so that the temperature is maintained within certain limits. What you are wanting to do (I think) is to maintain the oven's temperature within similar limits but without the element being on for long bursts.  You therefore want to dump all surplus power to the oven, as directed by the RF control signal, while the temperature remains below the upper limit.  If the temperature ever falls below the lower limit, there is clearly insufficient surplus power so the full 3kW consumption would need to re-start.  Minimising the period for which full power is on would give more opportunity for any surplus PV to get a look in. 

I'm no expert on electric ovens, but I would have thought they need to keep the temperature pretty constant.  So the extent to which the system could ride out the effect of a passing cloud, without any adverse effect on lunch, would be fairly limited.

The Mk2 or similar should not need to know anything about what's going on.  It just identifies surplus power at the supply point and makes it available to its dump outlets in the order specified: in your case, oven then immersion.

Sounds like a good sales pitch for a gas oven to me :)

Ha!

Especially as this is a duel fuel range! However, the cook prefers an electric oven and who am I to argue?

OK, I'll bite.

The only difference I can see in operation is you've added a remote controller and a feed-forward mechanism and a lot of complication.

In Robin's fundamental algorithm, the controller/diverter switches the dump load on the measurement of surplus power flow and all other loads - whether manually switched or otherwise - are primary loads. So what's the difference if the oven is being switched by a second triac. I think the answer might be 20 ms.

Here's my take on your explanation:

"So far everything is as normal. Now the clever bit begins. The oven cools sufficiently to demand more power..."

so the thermostat switches on at a random point in the cycle and possibly within that cycle, certainly during the next cycle, the controller detects it and at the start of the following cycle switches the dump load off. The oven has automatically grabbed priority and the consequence is the "energy bucket" has emptied a bit more rapidly.

"...but the remote immediately switches off its triac ..."

when it's demanding energy? But you've already drawn a current pulse in order to detect the thermostat closing ! ! !

"...and passes back information to mk2 using radio or even via a wired connection to indicate that a priority load is available."

but it knows that anyway

"Mk2 stops diverting and sends a ‘go’ signal to the remote which enables its triac ..."

so all we've done here is avoid a 1-cycle overlap?

"...and a burst of energy is applied to the oven heating element until mk2 sends the stop signal. mk2 continues to send go and stop signals until either a ‘satisfied’ signal comes from the remote or the energy bucket fills beyond a threshold which would indicate that the oen has switched off and normal diverting to the immersion should be restored."

And what if there isn't enough PV/surplus to maintain oven temperature and open the thermostat? Your dinner is late and under-cooked.

Sorry, but it sounds an awfully complicated scheme and I really can't see where there is a clear benefit. And the side effect could well be a burst of flicker because the thermostat switches on, the oven triac switches off, the dump load switches off and then the oven triac switches back on again. Unless you haven't explained that you somehow synchronise all the triac switching events to give a seamless transfer of energy.

KISS, I say.

I see no need for any reverse communication.

Yes, I now reckon that's true.

As you say, the tricky bit is dealing with intermittent PV availability. This could be a passing cloud or an extra load being turned on somewhere in the house. Also, my observations today made me realise that the oven door is opened on various occasions for mysterious observational purposes and that tends to allow a large volume of hot air to rise up to the ceiling necessitating another long burn. An additional problem is that some dishes require a lower starting temperature and end with a higher temperature.

Clearly more detailed empirical data is required but if a rough figure can be established for the maximum tick-over energy requirement then the Remote Controller (RC) can keep track of how many Joules have been burnt in say a 5 minute period and if this is outside the acceptance window then simply switch on the triac and let the oven regain control.

In other words, if insufficient energy to maintain tick-over has been supplied then turn on full. OR if more than the tick-over amount has been supplied then turn on full. Revert to normal when the CT detects opening of the thermostat. Also, the RC could test the state of the thermostat by applying the occasional half cycle and checking the CT for current flow.

Today was very sunny at lunch time and it was interesting to see that we were generating 2.8kW and the house was using 3kW so we probably had a 'free' lunch because the oven was burning about 2.7kW.

Having explored the technical feasibility the next thing to do is calculate the cost benefit.

The saving is the difference in price between gas and electricity which is in the order of 8p/kWh so if a typical meal consumes say 1.5kWh then 12p per session will be saved. Assuming 300 meals a year are prepared (probably less) then the benefit is £36 a year. Hardly worth thinking about.

Alternatively, if the main thermal dump has reached temperature when the cooking is being done then the value rises dramatically because the energy would otherwise go to the grid and the benefit is ~ £148 a year. Neither of these figures would justify the effort required but it's an interesting thought experiment.

Robert

An excellent critique!

My eMC controller uses what we now call anti-flicker so the odd half cycle here or there is insignificant as far as an oven controller is concerned. Remember, unlike the immersion heater this is power that we would burn anyway so micro pence perfection is not the objective.

Nevertheless most of what you say looks correct to me so I shall describe what set me thinking in this direction in the first place.

A couple of days ago we had a bit of sunshine and eMC was busily diverting power to the dump load during periods when the oven thermostat was open. When the oven thermostat closed eMC stopped diverting and the oven was taking ~ 1.5kW from the grid. It was pretty obvious that the average PV power was sufficient to meet the average input requirement of the oven but the imbalance of power availability and timing prevented it. Consequently the system was effectively burning electricity and saving gas which makes no sense with all this technology at our disposal.

I certainly agree that simplest is best and my original proposal is flawed but does that mean this task is beyond our collective skills to overcome?

Don't know why you lot are giving yourselves such a headache, I've been cooking countless meals over the last few months on pure solar, and have had some leftover to divert, except on the darkest days. Do what I did and buy a slow cooker, runs at 170W on it's highest setting, around 80W normally.

A standard oven controller has a thermostat with a variable temperature setting and some hysteresis.  In operation, the temperature ramps from one end of the hysteresis range to the other. 

While in the 'simmer' state, we would like to use surplus power at whatever rate is available.  But we must do this while remaining within the hysteresis range of the primary thermostat, else the soufflé will flop.  Tricky!

Isn't the answer then to replace the thermostat with it's relatively large hysteresis with a proportional controller?

Some of the better modern electric ovens use electronic thermostat control and guess what a BTA series triac, so you could interface a Mk2 directly. They do of course have thermo mechanical over temperature cut out.

So control the element with a Mk2 and allow the thermostat to mechanically switch in if the temperature is not getting high enough.

Mind you by the time theres enough sun to make the system viable, we will all be eating salads. :-)

i would say it's far to complicated, my oven has 3 heating elements top, bottom and around the fan. not all used at the same time i may add, but for different cooking functions.

I'm sticking with my kettle box, just turn the dial to the desired wattage.

I'm sticking with my kettle box, just turn the dial to the desired wattage.

If it uses phase-angle control, that sounds like the perfect solution!

Having just got my energy dump to immersion heater system up and running properly, I've been thinking some more about this not so little problem of what to do with the rest of the surplus and so would be interested to hear what the rest of you think of the following control method:

Others have mentioned the possibility of having a wireless controlled dumper that can be (plugged into any ring main outlet and) used around the house.  So far, I suspect most of the thinking has been to dump to the immersion primarily and then dump to this other box as a secondary load.

Can I suggest that it might be more fruitful to turn this around the other way and in the process make the control signal to the wireless controlled dumper proportionate (phase angle controlled).  Proportional control may be easier (smoother) to work with over the intrinsic delay that wireless control will introduce.  Such an approach could mean that we gradually ramp up power to the wireless controlled dumper as the sun comes up in the morning with less likelihood of consuming grid power (except for passing clouds or fridge cutting in, of course).  I envisage that the most suitable load would be a convector heater for space heating during the nippy Spring mornings that are rapidly approaching.   (Once the room being heated is up to temperature then, of course, the thermostat on the convector heater would turn it off and all the surplus power would be available to supply the other load)

During the above process it may be desirable to use the (secondary) immersion load to help balance the energy bucket using the burst fire and/or anti-flicker arrangement like already implemented on an instantaneous basis.   It may even be possible to use such instantaneous control in a transitory manner to preempt any delay in wireless transmissions to the remote load.

Tests I did for Robin a few days ago showed that the radio delay (measured from when a message with a 1 byte payload had been prepared prior to transmission to the message being received but not decoded) was 3.8 ms - with about 0.2 ms of jitter. So not practical for timing the phase angle even if you were allowed to monopolize the channel, but certainly good enough to send a demand signal to be acted on for the next half-cycle, when you've decided you need to change the setting.

If the remote load were to use p-a control, then the detailed management of that process could be entrusted to the remote processor.  The base-station would only need to send occasional instructions to increase/decrease power.  There should then be no problem with transmission delays.

I have a redundant £££ controller which can operate in p-a mode with full EMC protection.  This may finally be a good use for it!

The priorities of the two dump loads could be made switchable in a similar way to the feature that I've recently added to change between normal and a/f modes on the fly.  As shown in the video and schematic that I posted yesterday on my Mk2i thread .

I suspect when doing RF based control a cheap SSR would be more practical than a triac, as it takes care of the zero crossing switching. You'd only need to concentrate on switching on and off - something like a JeeNode would be fine for that.

P.

Robert

I am really surprised the transmission delay appears to be so short.  Promising then.

Robin

Sounds good. Longer intervals between transmissions may be possible if along with transmitting basic signal to consume power at a given level to the remote controller one transmitted a rate-of-change signal to ramp up/down power use, based on information from the PV generation channel.  Additionally, the remote controller might be permitted to make a choice of how much of this input to apply based on previous instructions.  In other words if the ramp alternates +/- then cool it!
Lots of scope for instability here I know but just thoughts out loud!

Robin

You must be the first person ever to have a 3ph meter in the conservatory!  Looks great - I am sure it will be a big talking point for the next WI coffee morning!

Paul

Take a look at the Sharp S216S01 & 02 if you haven't already.

As of 9:15 this morning I could have had 500w of heat gently playing on the lounge wall thermostat.

At present, 9:50 this thermostat is still instructing the oil-fired boiler to cram heat into the under-floor pipe-work.
I don't need this.   At noon we will be roasting.

Paul: I suspect when doing RF based control a cheap SSR would be more practical than a triac, as it takes care of the zero crossing switching. You'd only need to concentrate on switching on and off - something like a JeeNode would be fine for that.

Sorry Paul, I'm not sure what point you're making.  The standard trigger takes care of zero-crossings, so I don't see why a triac is any less able to do remote switching than an SSR.   Whatever form of controller is at the remote end - Arduino,  emonTx or JeeNode - they will all the same Atmel processor on board.

RobP: Longer intervals between transmissions may be possible if along with transmitting basic signal to consume power at a given level to the remote controller one transmitted a rate-of-change signal to ramp up/down power use, based on information from the PV generation channel.

I don't see that "info from the PV generation channel" is any more relevant than info from another channel.  My Mk2 setup only has one 'channel', which measures the net flow of energy at the supply point.  Surplus power is the only parameter that's important. 

There is a p-a version of my Mk2 router which proportionally alters the control signal depending on the level of the bucket.  This is a reactive system with a short response time.   I don't see how you can meaningfully predict how the amount of surplus power is likely to change; you can only react to what it is now.

Sorry Paul, I'm not sure what point you're making.

Robin, the point I'm making is that you don't need to do any voltage sensing at the switching end of the RF link in order to switch at the zero point of the cycle. All you need is a simple receiver and SSR. If using a triac based design (the standard trigger) you'd need to sense voltage too, and more local processing. You can't rely on the RF link to hit the zero crossing point, because of the latency - especially if the transmission is delayed because of a clash.

P. 

I am also puzzled by Paul's comment.

The MOC3041 Zero-Cross Optoisolator Triac Driver that is used in these PV dump controllers looks after zero cross detection for you so there is no need for further hardware. Or have I missed something?

Paul is correct to point out that the operation of the MOC3041 or equivalent will be less well controlled if it is fired at a random part of the cycle.  But if the trigger were to just miss one opportunity to come on, it will surely do so at the next z/c point approx. 9.9 mS later.  I don't think the odd half-cycle's worth of jitter (30J max) would make much difference for this application. 

I thought that Paul was saying the SSR has zero cross detection and a triac doesn't - which is true if you don't fit the MOC 30XX

Yes I agree that missing one half cycle is no big deal especially in A-F mode.

If I've read the above posts correctly, then what is being discussed is more or less how I was controlling my immersion heating all last summer.  I used the emontx to transmit solar and house use value's to a jeenode (also initially a 2nd emontx) which then did the processing to control a Crydom proportional controller.  I just altered the code slightly to send approx 60 messages a minute which was enough to keep my suppliers meter illuminating the export LED (green in my case, red flash is import).  If a 2kW heater was switched on then the green LED only went out for a brief moment before the jeenode received an updated message so it knew to turn off/down the immersion power (obviously when solar generation exceeded the load).

Although I trialed faster transmissions, I found even 1 per second to be adequate as I wasn't worried about losing the odd penny - in my house it's only a handful of instant high load switch on's in a day, and the system handled fridge type loads with no issue's or even cloud cover etc.   I was going to adapt the code to be more intelligent in that it would only transmit so frequent if there was a change in power, but I've now installed a wire and have all the processing done on the single emontx.

The point I was (clumsily) trying to make is that with an SSR you could simply connect it to a digital output of (say) a JeeNode, and that's it. Bob's your uncle.

A triac needs bits around it to do the zero crossing stuff, hence the driver etc.

I was trying to think of a way to keep the switching end of the RF link as simple as possible.

P.

I've never quite worked out what an SSR is, in terms of its cost and what it does. 

A standard triac with a z/c trigger and three resistors shouldn't come to more than a fiver.

Plus a markup for the name on the outside, the CE mark, the website, the salesman, ... ?

"Plus a markup for the name on the outside, the CE mark, the website, the salesman, ... ?"

Not to mention some SSRs are decidedly dodgy, some of the cheaper one from China are just bad copies, and anyone over there can print CE on a counterfeit. Not all SSRs are zero crossing either, you have to choose them wisely. I prefer to use a known good triac and zero cross trigger (MOC3021) AKA Robins design.

"some of the cheaper one from China are just bad copies"

While it is an expensive Crydom SSR I use for my immersion heating, I have looked at cheap chinese SSR's on eBay for other purposes as the prices are so good.  As per Robins post - "A standard triac with a z/c trigger and three resistors shouldn't come to more than a fiver." - then if it is so cheap and simple then how are the chinese made SSR's likely to be so bad?

I'm asking from a totally uneducated perspective.

Dennis

Most of the power to heat our DHW during the last 7 months has been switched by a BTA41 triac.  This has always been used in zero-crossing mode alongside an MOC3041 trigger.  I also have a smaller triac, a BTA16 which has been used in a similar manner and in phase-angle mode with an MOC3021 'immediate' trigger.  My heatsinks are pretty basic and neither device has ever complained. 

Meanwhile, my £70 Carlo Gavazzi AC Controller sits idle in its box.  Requiring a 0 - 10V control signal and a 10V AC supply, or a current source, it's also a real pain to interface with.  The trigger & triac combination is cheap and effective, and much simpler to use, albeit without the kite mark.

At least some of the extra cost with good quality commercial units is the protection circuitry. All the ones I've seen have fast, usually solid state, fuses and thermal monitoring and shut down for the triac heatsink. The latter could potentially prevent a fire so it's probably a good idea even on a homemade unit.

Martin said:  ...thermal monitoring and shut down for the triac heatsink. The latter could potentially prevent a fire so it's probably a good idea even on a homemade unit.

Would the ATtiny84 be likely to have enough going for it to act as the 'brains' for such a wireless controlled dumper?

Maybe using the RPi adapter from the shop? 

http://shop.openenergymonitor.com/rfm12pi-raspberry-pi-expansion-board-pcb-only/

A sunny day today!

My slightly OTT heat dump worked very well today diverting 8.7kWh of the 11.5kWh we generated from what is is probably a 3kW peak system.

So, how well did the less OTT systems manage?

19kWh on a 7k2Wp system with some shading at this time of year.  Didn't have the dumper turned on though.

The PV Energy Dump and How To Use It

Hi
I have a 3.8 KW PV system and using (Ians) Series 530 Fusion system to heat my economy seven tank using the bottom immersion and the top immersion on cheap rate electricity at night. This has been running trouble free for a month.

As an aside initially for the first week of testing I had a tank full of hot water every evening even when there was little sun I thought this is good must be good Lancashire Sun.

I then went on Holiday to Benidorm with a couple of other Bus Pass carriers for a week (Wife stayed home looking after dog) The sun shines there and I like Brandy and coffee in a morning on the front then sit in a bar watching the grand prix (Motability Scooters) whilst drinking cold lager 1.5 euro a pint. Also no PV panels
To talk about.

When I got back wife says it is a good system always had hot water and no problems except why do we have those flashing lights!

Then I looked at the graphs from emoncms and noticed the tank started getting warm at the bottom when it was still dark I checked the Motorised valve which controls the central heating flow to the tank and found it sticking partly open. Back to reality.

I still believe that during summer The PV dump will easily top up the tank. So reading this thread I thought of Air Conditioning. I have a conservatory which gets to hot to sit in during the afternoon in summer so my thoughts are install an air-conditioner and use the LED that lights on the TX board when electricity is being given back to the grid to operate a MOC 3041 to isolate the system and switch another BTA41 Triac using the normal ring main to provide the power for the air conditioner which would still be controlled by its own thermostat. This could run along side the normal PV Dump.

This may be a rubbish idea so I would like your comments please.

David  
 

RobP said: Would the ATtiny84 be likely to have enough going for it to act as the 'brains' for such a wireless controlled dumper?

I'm sure it would but you have to jump through a few hoops to program them as I found out when I changed the baud on my RPi base station. For the extra 50p you'd probably be better off using the emonTx bare PCB. I've now got 6 of those in my system!

Has anyone mentioned the possibility of using surplus PV to pump rainwater up to the top of the house for flushing loos etc.?  Rainwater harvesting is all very well but it needs to regain some altitude to be really useful. 

I think I've heard somewhere that the cost of the electricity to pump rainwater up to the top of a house is more than the value of the water that is saved.  An ideal candidate for surplus PV then. 

I wonder how a pump would cope with being cycled on & off as conditions change, or what power rating might be needed.  A 20' high column of water would require some oomph to get it moving.

I think I've heard somewhere that the cost of the electricity to pump rainwater up to the top of a house is more than the value of the water that is saved.

That seems a bold claim, but I suppose it depends where you live.  I pay:

$3.2692 per KL of water (or 3.26 cents per 10L)

$0.25c per KWh (when it's cloudy or at night)

$0.50c per KWh (lost export revenue when it's sunny)

It takes my rainwater tank pump 82 seconds to fill a 10L bucket 1 floor higher than the tank, and that's thru a cheap garden hose with a nozzle on the end.  It may well be better through proper plumbing.

The pump uses about 300W, which works out at about 6.8WH for 82 seconds  (10L), or 0.34 cents if the sun is shining.

So unless I've dropped a decimal place somewhere, the water here costs about 10x what the electricity costs to pump it (20x if I do it when the sun isn't shining).

I wonder how a pump would cope with being cycled on & off as conditions change

I can't imagine a standard pump would like it much; they go to some lengths to prevent it cycling too often (pressure tanks etc).  I imagine they have pretty big start-up currents which you'd end up seeing a lot of if you cycled very often.

Thanks  dBC, plenty of useful info for us to be thinking about there ...

I think I've heard somewhere that the cost of the electricity to pump rainwater up to the top of a house is more than the value of the water that is saved.  An ideal candidate for surplus PV then. 

NOOOOOOO Robin, me and my 3000L rainwater harvesting system totally disagree I think you must have stumbled across an urban myth there, possibly manufactured by a water company. My system has a 125L header tank in the loft and is rigged so the pump switches on when there is about 20L left in the header so runs for a couple of minuets a couple of times a day. I don't know how the rest of the UK work but here under the thumb of Southwest Water we are charged more for our water than anywhere else in the country. we are charged an eye watering £1.96 per cubic meter on top of that the sewerage charge is £3.30 per cubic meter, yes that's right we are charged twice whatever we use the water for so that is £5.26 per cubic meter the government have told SWW they must not increase water charges so guess what the sewerage charge has gone up 10% in the last year, thats what happens when you let a foreign owned company run your utilities.

Without doing the maths I know I'm saving 70% off my water bill by harvesting rainwater and getting to wash the car,do the laundry, water the garden and even poo for free. I could put a timer on my 800W pump so it didn't come on at night but then I'm in danger of  using expensive metered water if it calls at night.

So there you are, rainwater harvesting = very good idea.

It costs 9.81 J to lift 1l of water ( = 1 kg) through 1 m. That's pure physics.

So it looks to me as if it costs about 60 kJ to lift 1 m³ up into a loft - about 0.016 kWh. You need to add in friction losses in the pipe and pump efficiency of course, so say 0.033 kWh to account for losses. 'Nuff said?

(And if you don't think I got it right - I didn't - put the numbers in here: http://www.engineeringtoolbox.com/pumps-power-d_505.html)

That's amazing.  0.033 kWh would keep my 3 kW kettle going for 40 seconds, not long enough to boil up a cuppa.  Yet the same amount of energy can hoist a ton of water from ground level into the loft. 

So this option for using surplus energy does not look to be very practical, unless you happen to have a Dinorwig handy !

So if I understand you Robert that is .039KWh per cubic meter to get it pumped to the loft so working on my suppliers cost of 12.24p per KWh each M cubic pumped costs me 0.477p. I think I'm on to a winner.

the starting temperature of the tank could be as high as 50 deg

You don't particularly want it sitting around for a long time at low temps either.  20C - 45C is ideal for legionella.  The standard here (AS3498) requires that in the 7 days prior to coming out of the tank, any hot water must have been "treated" as follows:

TABLE 1
                MINIMUM EXPOSURE PERIOD
              Temperature           Minimum exposure period
             70°C or greater           1s
                  66°C                        2 min
                  60°C                        32 min
                  55°C                        6h

I think you've got it!  ½p to pump up water that would have cost you £5.26 to buy in ready pumped.

do the laundry

How high is your loft?  I think my washing machine's minimum inlet pressure is 100kPa, which means you'd need about 10m of head.

do the laundry

Most washing machines over here are plumbed into the DHW system which is generally at the pressure of the header tank on the upstairs floor.  So that's only around 5 metres of head.

I'm intrigued by the idea of using a rainwater-fed supply for the washing machine.  Is this for the hot or cold, or both?  Sounds like a lot of extra plumbing - a Dick Strawbridge house maybe?

Yes, it may be cheap to pump water but it's not cheap to heat it.  I guess if you have to heat it anyway then it doesn't much matter who heats it.  My hot water is essentially free (the heat part not the water part) so the last thing I want is the washing machine or dishwasher drawing cold water (even if it is free rainwater) and then heating it up, so both of mine are plumbed to the hot water system (mains pressure, town water, solar-thermal heating).

DHW system which is generally at the pressure of the header tank on the upstairs floor

I wonder what's the history behind our divergence on that. The last time I saw one of those was in my grandfather's house in the 70s.  I'm pretty sure  they've all long since been replaced with mains pressure hot water systems here.  Maybe the "install a Rheem" marketing jingle was the secret ;-)....

http://www.youtube.com/watch?v=wbPJoCPC4OE

The issues originally posed in this long thread (put in my words) were:

1. Limited energy store due to a 55°C max hot water tank temperature.

2. Limited energy store due to stratification.

3. Limited energy store because tank may already be at 50°C

4. How to use surplus PV power.

To add my two pennies worth:

1. There is no need to limit the temperature to 55°C.  Fitting a simple thermostatic mixing valve (TMV) to the HW tank outlet will limit the temperature at the taps   This approach commonly used on Solar Thermal systems allows much higher tank temperatures (70°C +) , prevents scalding and indeed help minimise wasted hot water.

2. Stratification may be controlled on un-vented sealed systems by fitting a pipe between the HW outlet (before the TMV) and the cold water feed at the base of the tank.  A valve fitted in this pipe could be used to control the stratification and hence ensure there is sufficient temperature in the top of the tank when needed, but also allow circulation and thus more uniform temperature distribution when there is more PV energy available.  I don't believe a circulation pump is required as the water would naturally circulate.  On vented systems its not so easy as hot water would tend to rise up the cold water feed pipe.  A non-return valve here would pressurise the system which is not acceptable.  By the time you've finished you end up with most of the components of a sealed system.

3. I think a HW control system strategy can be written that accounts for the likely solar energy and compares this to the hot water available.  If at a pre determined time before the expected hot water use (morning / evening showers / baths) there is not enough hot water temperature,  then the boiler can provide the heat otherwise just rely on the solar PV energy. 

4. I take use surplus PV power to mean, use in a beneficial way; beneficial to who I guess is one question.  I would say it benefits us all by sending the energy to the grid, but if we assume we're trying to maximise the personal benefit first then it makes sense to store excess PV energy if it can be used to replace other energy sources.  That said I think the HW tank as a thermal store works in Spring in Autumn, but in Summer there will likely be PV energy that exceeds the majority of households hot water requirement.  The most sensible suggestion I saw here was to recharge an electric / hybrid vehicle, but I cant see the point in using excess energy to power a tumble dryer.  Surely if there is sufficient solar energy for that you could try putting your clothes on a line outside :-)

The most sensible suggestion I saw here was to recharge an electric / hybrid vehicle, but I cant see the point in using excess energy to power a tumble dryer.  Surely if there is sufficient solar energy for that you could try putting your clothes on a line outside :-)

Sure, but maybe the vehicle owner wants to drive their car while the sun is shining rather than keeping it on the drive.  And maybe they will be out at work so it's not possible for them to be at home to bring the washing in between showers.  An RF-controlled heater over which damp washing can be safely dried may be a simple & effective approach. 

How about using a small (preferably mobile) storage heater?  I believe the technology has advanced somewhat since the days of mega-heavy boxes of bricks on the wall.

Most washing machines over here are plumbed into the DHW system which is generally at the pressure of the header tank on the upstairs floor.  So that's only around 5 metres of head.

I'm intrigued by the idea of using a rainwater-fed supply for the washing machine.  Is this for the hot or cold, or both?  Sounds like a lot of extra plumbing - a Dick Strawbridge house maybe?

My washing machine uses harvested rainwater, the only problem with feeding one from a loft supply is that most modern machines use very little water , by the time it gets to the machine its still getting the cold stuff in the pipework, also a hot/cold fill machine will be hard to find nowadays. Cold fill is OK as long as you have excess solar to power the heater in the machine, not a problem with mine as it has a delay timer. Mine is a Hoover with a +3AAA rating (other brands of washing machine are available). 

richmc, Do you heat harvested rainwater separately for your washing machine, or does all of your DHW come from the same tank? 

If the latter, then does it matter that the header tank is in the loft?  It's surely only the DHW-tank to washing m/c distance that matters, not the source of the cold stuff.

Neither Robin, I see no point in heating water for the washing machine, I wait till there is enough solar to let the machine do it for itself, it only runs at around 2KW so there is usually some time of the day when there is enough excess for the machine. Washing machines only use hot water for the initial fill, all the rinses are cold. to get enough head  to give enough pressure for the machine would only mean you are filling the pipe work with hot water not the machine. The point being the source of the cold (rain) water is free, and the electricity (solar) is also free if I get the timing right. The only thing I'm paying for is the detergent. Like when I wash the car, rainwater and my "solar powered" karcher.

Some possibly pertinent comments in these threads: External immersion heaters and How about energy storage?