Immersion Heater - PV electricity

[chriswarhurst]

Let me know when you have developed one - i will be very interested

 

[suntrap]

He already has - and all the information about it has been published. Look through this thread if you want to know.

 

[DickiePhitt]

I've been working my way through this thread for days. Really interesting and a credit to the contributors. Its taken days because I'm a chartered (mech) eng so can only deal with things that can be measured in buckets. No, I've 16th Ed and did hobby electronics years ago so I think I have the skills, if not the electronics knowledge, to have a go at building a simple,safe, economical controller for an existing 3kw Immersion heater.

Was surprised at how little the potential savings were, though. Checking the gas meter these past few weeks without heating on, converting to kWh, and plugging in the cost per unit I reckon about 50p per day on HW. In summer its likely HW will max out on the temp stat part way through the day and in winter it'll struggle to put much surplus into the HW tank. So maybe 1/3 of the annual cost will be saved - say £60 - which is worth having but limits options - yeah I know its a crude calculation but probably not wildly out.

Can I just ask a few questions ,which I hope are not too stupid, asked on the basis of 'will it work' not "is it the perfect solution"?
I want to use the Arduino (no experience, I'll have to learn) as it seems to be a well supported, flexible platform for development and communicating. Am I right that 2 suitably sized LEM CT's will hook up fine to the inputs? Think I read here somewhere that something else was needed but cant see where and cant think why.

The Crydom and filter (at the right price) + fuse & MCB looks the proprietry, proven way to control the 3kW load. Since it doesn't like PWM I thought a suitable compromise might be to use an R-2R ladder and an Op-Amp to double the voltage.

Digital to Analog converters (DAC), using R/2R netwroks

Suitably programmed the Arduino would step the output voltage applied to the Crydom which would raise or lower the immersion heater power. Not perfect DAC but a good 80:20 pereto solution I thought.

If this is totally out of court can someone please suggest and illustrate / point me to a suitable way of achieving this? I would need to be taken through Triacs, etc as they came after my hobby days.

 

[methley]

I'm building one right now with an Arduino Uno board which is perfect for this job. For the output, the R-2R ladder uses too many outputs. As the Arduino supports PWM outputs all you need to do is filter the output of a PWM pin and add a voltage doubler to drive the input of a Crydom. The Crydom takes away any difficulty in driving a Triac. It's all there in one package, and is fed simply with a 0-10v input.

For the inputs from the LEM's you'll need to half the voltage, using a simpler resistor potential divider. Note the minimim load on the output of the LEM is 2MOhm.

 

[pauldreed]

My Arduino based device is here if you are interested.
The first version was built with a Arduino Uno, but my current one uses a £4 Arduino chip (Atmega 328PU) along with a few support components, and runs the same software.
The credit however must go to these guys.

 

[DickiePhitt]

Methley, have you successfully got the Crydom working via PWM and filter? Crydom advised me "The analogue DC input is sensed by a micro-controller, a PWM input would either not be recognized or falsely interpreted. The PWM signal would have to be converted to DC signal." which led me to doubt that a filter would give a clean enough input, and as a result might damage or overheat it.

Paul, I've been impressed with the development you've put into this, as witnessed by this thread and others. I'll read up the links and try to get some understanding of what you and these guys have been doing. Might be a bit advanced for me I fear :-(

 

[pauldreed]

Methley, have you successfully got the Crydom working via PWM and filter? Crydom advised me "The analogue DC input is sensed by a micro-controller, a PWM input would either not be recognized or falsely interpreted. The PWM signal would have to be converted to DC signal." which led me to doubt that a filter would give a clean enough input, and as a result might damage or overheat it.
(

I too spoke with Crydom some months back, the problem with PWM is inherently the speed of the sampling rate is too fast for the Crydom Thyristor circuit to recover, and therefore would not work.
I think Methley has taken the PWM output through a capacitor/resistor circuit to smooth and average the peak value, similar to this.
The input would be clean enough, but there is a lot of debate around the switching harmonics/distortions caused by the thyristor.
As for being 'Advanced', this was my first project for over 30 years when valves were the norm, so if I can, so can you, just take your time and if you get stuck there is plenty of help in the forums :->

 

[suntrap]

Was surprised at how little the potential savings were, though. Checking the gas meter these past few weeks without heating on, converting to kWh, and plugging in the cost per unit I reckon about 50p per day on HW. In summer its likely HW will max out on the temp stat part way through the day and in winter it'll struggle to put much surplus into the HW tank. So maybe 1/3 of the annual cost will be saved - say £60 - which is worth having but limits options - yeah I know its a crude calculation but probably not wildly out.

That sounds very plausible and does make it less attractive if you are offsetting the cheapest fuel - mains gas. Same problem there has always been with solar water heating: loads in the summer and not enough in the winter. And you could also argue that offsetting gas consumption with electricity is not very environmentally friendly. I did ponder the idea of overriding the freezer thermostat to reduce its consumption from the grid at night, but of course that's a fairly small fixed load and has the additional complication of not liking to be short cycled.

 

[suntrap]

an Op-Amp to double the voltage.

There is a 5V version available.

 

[pmcalli]

I have now built and tested two immersion controllers so thought I would share the results.

the first solution used a PID controller crydom and true rms ct's which as people here have pointed out is a very expensive solution.

results
The true rms CT's are loaded by the PID controller this does not affect the operation but is undesirable.
the response time of the true rms cts is 300ms so the PID integration time had to be long
very small grid feed in with large load changes otherwise results good.

the second solution uses some inexpensive ops amps the crydom controller and a couple of owl ct's.

results
the cost is very low the owl ct's are obtainable for about £6 each op amps and other components are less than a £1. crydom is about £70 filter £25 power supply £10
response time is set at 100ms and no grid feed in or consumption is measurable.

As there has been many posts on how the inverter and phase controller work to ensure no grid consumption. I have run an experiment whose results may lead to some other solutions.

while my controller is running using an oscilloscope with line trigger and a spare CT I have measured the current from the inverter and the current going to and from the grid.

the result is the inverter delivers current evenly over the whole mains cycle. this means that if the load from the immersion heater is phase controlled then the part of the mains cycle that it is on the current is drawn from the inverter and the grid if the inverter is not generating enough. On the part of the cycle that the crydom is off the inverter feeds into the grid. Over a cycle the net result is zero. the meter response time is more than a cycle and therefore does not record that you borrowed some power and gave it back before the end of the cycle. measuring the current going to and from the grid confirms the result. measuring the meters response time is more tricky as I would have to burst fire a SSR over an increasing time until the meter recorded if i get around to it I will post the results but I expect it will vary with meter type.

I shall be laying out a PCB ( currently on vero board ) and am happy to share the circuit diagram/component list/PCB layout with anybody who wants it.

incidentally I also ran the controller without the i/p filter and I am unable to Bluetooth to my inverter due to the interference showing that an input filter is absolutely essential.

 

[suntrap]

Excellent stuff! Would love to see the circuit diagram.

 

[DickiePhitt]

There is a 5V version available.

Taken a look. Its a burst fire controller, not a proportional controller.

quote "This device regulates the speed of an on/off pulse. Provides a smooth proportional control that minimizes electrical noise by utilizing zero-cross detection and switching, firing only complete AC cycles. Specialized Phase Detection Circuitry allows the MCBC to be connected to only one side of the load, minimizing installation wiring. Particularly suitable for use in heating applications where the electrical noise generated by typical phase angle controllers cannot be tolerated."

So this might remove the need for the filter BUT since it fires complete AC cycles wont it be 3kW on for so many cycles/3kW off for so many cycles so the company meter will register consumption, unless PV happens to be supplying everything.

Very happy to stand corrected, as I'm very new to all this.

 

[suntrap]

Taken a look. Its a burst fire controller, not a proportional controller.

Ah sorry of course - the correct part would be an MCPC2425A or an MCPC2450A, neither of which RS sell.

 

[pmcalli]

Excellent stuff! Would love to see the circuit diagram.

here is schematic for control loop. I haven't finalized the +- 15v supply components so haven't included them will post full schematic when finishedView attachment analogue.bmp

 

[pmcalli]

Ah sorry of course - the correct part would be an MCPC2425A or an MCPC2450A, neither of which RS sell.

£62.03 from Farnell in stock.

 

[pauldreed]

here is schematic for control loop. I haven't finalized the +- 15v supply components so haven't included them will post full schematic when finished

I'm surprised that the value of R1 & R5 burden resistors are so high.
I'm using a CT with 3000 turns and a burden resistor value of just 240R which gives me -2.5V to +2.5V (5V p-p) when reading 5kW.
I found this handy online calculator for CT's/Burden resistors, so you can determine safe peak voltages without damaging your op amps if it helps.
Thanks for sharing your schematic, there are so many good ideas surfacing now.

 

[pmcalli]

I'm surprised that the value of R1 & R5 burden resistors are so high.
I'm using a CT with 3000 turns and a burden resistor value of just 240R which gives me -2.5V to +2.5V (5V p-p) when reading 5kW.
I found this handy online calculator for CT's/Burden resistors, so you can determine safe peak voltages without damaging your op amps if it helps.
Thanks for sharing your schematic, there are so many good ideas surfacing now.

r1 and r5 are not burden resistors they hold the i/ps to zero as a precaution if the CT's are removed. the owl ct's give 0.1v per amp which implies they have a burden resistor built in I presume for safety reasons.

 

[pauldreed]

r1 and r5 are not burden resistors they hold the i/ps to zero as a precaution if the CT's are removed. the owl ct's give 0.1v per amp which implies they have a burden resistor built in I presume for safety reasons.

Ah that makes sense. It might be worthwhile including 'nominal' burden resistors in your schematic, because if someone else builds your project using standard CT's (no internal burden resistor), the excessive voltage may damage the op amps.
I have been caught out twice now, whilst disconnecting my CT's from the PCB, have had unwelcome reminders at what high voltages are developed across the coil whilst no resistor is there to reduce it. The first time, by reaction, I snatched my hand away and cut it on an adjacent metal frame. The second time, I was miffed because I hadn't learnt from shock#1!!

 

[pmcalli]

Ah that makes sense. It might be worthwhile including 'nominal' burden resistors in your schematic, because if someone else builds your project using standard CT's (no internal burden resistor), the excessive voltage may damage the op amps.
I have been caught out twice now, whilst disconnecting my CT's from the PCB, have had unwelcome reminders at what high voltages are developed across the coil whilst no resistor is there to reduce it. The first time, by reaction, I snatched my hand away and cut it on an adjacent metal frame. The second time, I was miffed because I hadn't learnt from shock#1!!

good point I have been meaning to determine the value for a while but the only way I can think of is to put in a step impulse via a resistor and measure the initial voltage. I think all the energy monitor CT's have the burden resistor plus possibly a zener built in so joe public doesn't have your electrifying experience. I am just splitting the schematic up into a block diagram and a schematic to just cover the PCB which now incorporates the +- 15v generation so will incorporate the burden resistor into the block diagram schematic.

 

[nbown]

i have been reading though this tread with great interest, i have just a one question has any thought about using a
900W Heating Element with Thermostat the type that goes in towel rads? just pluged into the feed?
thanks

 

[suntrap]

I presume the main trouble with those is that they don't fit into the hole in a hot water cylinder.

 

[inie meanie]

i have been reading though this tread with great interest, i have just a one question has any thought about using a
900W Heating Element with Thermostat the type that goes in towel rads? just pluged into the feed?
thanks

Hi do you have a link to a product?

 

[DickiePhitt]

I'm pondering the best/easiest/most cost-effective route to inputting real time data into whatever microprocessor does the clever stuff (probably Arduino/Nanode).

The LEM CTs look expensive but I'm sure they are wonderful.

Anyone have any experience of Efergy units (cheap on Ebay) but at 90A won't they reduce accuracy?
Efergy Standard Sensor - Jackplug Fitting | eBay
Likewise I see owl CTs at a reasonable price on Amazon but cant find the spec.
OWL CMA113 Standard Sensor Fior Economy 7 Users or 3: Amazon.co.uk: Electronics
Also come across a SCT-013-030 which seems only available in USA. Anyone have any details / UK supplier?
Non-invasive AC current sensor SCT-013 (30A max) Non-invasive AC current sensor SCT-013 (30A max) [SEN005] - $9.90 : iStore, Make Innovation Easier

Finally a (silly?) thought. As I have a CC EnviR unit and additional clip-on CTs with wireless senders, which I understand operate on433MHz SRD band, the same as an Anduino shield. Could I not just pick up the existing transmitted data using the shield? Maybe a better Q for FreeEnergyForum

 

[pauldreed]

I'm pondering the best/easiest/most cost-effective route to inputting real time data into whatever microprocessor does the clever stuff (probably Arduino/Nanode).

The LEM CTs look expensive but I'm sure they are wonderful.

Anyone have any experience of Efergy units (cheap on Ebay) but at 90A won't they reduce accuracy?
Efergy Standard Sensor - Jackplug Fitting | eBay
Likewise I see owl CTs at a reasonable price on Amazon but cant find the spec.
OWL CMA113 Standard Sensor Fior Economy 7 Users or 3: Amazon.co.uk: Electronics
Also come across a SCT-013-030 which seems only available in USA. Anyone have any details / UK supplier?
Non-invasive AC current sensor SCT-013 (30A max) Non-invasive AC current sensor SCT-013 (30A max) [SEN005] - $9.90 : iStore, Make Innovation Easier

Finally a (silly?) thought. As I have a CC EnviR unit and additional clip-on CTs with wireless senders, which I understand operate on433MHz SRD band, the same as an Anduino shield. Could I not just pick up the existing transmitted data using the shield? Maybe a better Q for FreeEnergyForum

Try these CT's, they are each rated 3000:1 coil ratio, and at £16.50p FOR 2, they are tremendous value, and from my experience produce very good results. I'm well pleased with them.
As for your second point, I am doubtful. It took 2 weeks for CurrentCost to decide whether to tell me the spec of their CT's (their intellectual property rights stuff!), when I could have simply taken one to pieces and counted the turns ratio. So I am very doubtful that you would get any cooperation in establishing the communication protocol between the two. Nice lateral thinking though!

 

[nbown]

Hi do you have a link to a product?

here you go, 800W 900W Heated Towel Rail Electric Heater Element | eBay

i was thinking of having two of these, one the side and one on the top of my cylinder. just got on think of a way to kick them off. my idea is the have a current clamp on the ac output on the invertor, and when a set point for current is meet, then is would switch it on. i have 3 set points, 1x 900w, 2x900 and than the 3kw immersion.

 

[JohnBoyZ]

I moved the RCBO for the solar feed to be next to the main incomer in the cs unit and then put the CT around the the incoming meter tail and the solar feed live which are now right next to each other. second CT just goes around the solar feed live all nicely contained in the CS. My CS is a large MK sentry so there is plenty of room.

Hi, Ive been dircted to this thread from a discussion on the PV forum and have spent hours sifting through all the pages and some very interesting idea. I'm a little confused about your CT's and wondered if could you elaborate a bit on how everything is connected in your system. Do you simply have one CU with PV fed in via RCBO and immersion fed from MCB or is the immersion feed separated off from the CU. I've been trying to work out what the result is from your CT set up with up one on PV+Incoming and another on the PV?
I've currently got everything connected into the CU so have thought of putting a CT around all the house circuits (excluding immersion) and another on the PV thereby enabling me to calculate the excess power or maybe adding a Henley block to my set up to keep everything separate but is your method a cunning way to get round all that?

Its probably all very simple but its late and my brains dead
Cheers
John

 

[inie meanie]

Hi John
Have a look at Picture 1 of 5 from Power controller Sorry but it's just a sketch but it will give you the idea. The picture shows one way of connecting. The other will take the feed from the house dist board to the triac, but the circuit used is not passive but reactive , in regard that it balances the two power levels. I haven't posted that version yet but I will in a few days. I will add it to the collection of pics. I can say that I have tried the method of fitting all the cores through one ct and it's not easy. It is so much better to fit a Henley block.
Regards

 

[JohnBoyZ]

Hi John
Have a look at Picture 1 of 5 from Power controller Sorry but it's just a sketch but it will give you the idea. The picture shows one way of connecting. The other will take the feed from the house dist board to the triac, but the circuit used is not passive but reactive , in regard that it balances the two power levels. I haven't posted that version yet but I will in a few days. I will add it to the collection of pics. I can say that I have tried the method of fitting all the cores through one ct and it's not easy. It is so much better to fit a Henley block.
Regards

Hi innie,
Thanks for your diagram, it is how I envisaged things when using a Henley as it keeps every thing simple. I was intrigued with pmcalli's set up (one CT around both the the incoming meter tail and the PV and the second CT on the PV), I cant quite see how to calculate the excess PV power with that set up.
And by the way well done to you inie and all the others on this thread for coming up with so many subtly different concepts and getting them up and running, its very impressive.

EDIT:
@inie Had a look at your other diagrams and noticed your circuit diagram using a trial version of smart draw, have you tried the free circuit schematic an PCB design CAD from express PCB, I've used it, its very good and free
http://www.expresspcb.com/ExpressPCBHtm/Free_cad_software.htm



Cheers
John

 

[inie meanie]

Thanks I will down load it now and have a go.

 

[inie meanie]

PV is dead. RIP PV :- (