Home Automation - Heating

[ban-all-sheds]

This, and a recent post about CH systems with a mix of zones, motorised valves, programmers etc, has started me wondering.

If you had a very well insulated house, then given the initial and ongoing costs of installing a gas boiler and wet CH, and given the reality of how efficient gas boilers are when coping with a wide range of demands, would electric heating, which is 100% efficient, costs SFA to service & maintain, and which needs just simple per-room local controls (which could be centrally programmed) sometimes be a better option?

 

[bernardgreen]

If the house was perfectly insulated ( thermally ) then it would not need any heating, in fact it would need cooling to prevent heat from the inhabitants heating it up to 98.4 F ( body heat ).

Houses that are very well insulated do require ventilation and to achieve adequate air changes mechanical ventilation is probably needed with heat recovery.

So while electric heating may be easy to install and maintain the other systems such as ventilation with heat recovery become more complex, intrusive ( noise and inconvenience ) and expensive.

Personally I like a leaky ( well ventilated ) house to ensure plenty of fresh air and avoid allergies and other illnesses which seem to be very common in modern well insulated houses.

 

[Simon35]

Plumbing side i have a fairly good idea of what I'm aiming for, was looking at the electrical side - replacing bog standard thermostats with a computer/brain of some description. Initially i looked at using a RaspberryPi but need to figure out how to get it to control 240v.

My solution might be a bit over the top for 18 devices but I recently realised there is a very simple way of controlling mains devices. Obviously opto couplers would give the necessary isolation but the mains side circuitry is fairly awkward. I do not need microsecond response times and wanted to avoid building from scratch.

An electronic room thermostat (I had an old Drayton Digistat) contains a thermistor likely to be a nominal 47K. It also happens to be the middle of the range of a light dependant resistor. http://www.maplin.co.uk/light-dependent-resistors-35962

It's a direct substitution. Make a hole for some light to shine in and point an led at it. No wiring needs to go inside the housing so no worries about isolation. On mine, 100 microamps through the led is enough to make it swing from end to end. I hold the led on the side of the housing with velcro to make a "connector". I get a response time of about 90 seconds.

Cant you just use electromechanical relays to switch the 240V with the Raspberry Pi outputs? Its what we do every day in industrial automation systems. I dont know what voltage the Pi outputs are though, or what they can drive current-wise..maybe thats why you're faffing about making your own devices...

 

[ban-all-sheds]

If the house was perfectly insulated ( thermally ) then it would not need any heating, in fact it would need cooling to prevent heat from the inhabitants heating it up to 98.4 F ( body heat ).

Yup.

I went into a demo house in CAT in Wales once - thick insulated walls, triple-glazed windows, and (IIRC) airlock doors.

No heating - perfectly warm courtesy of people.

Houses that are very well insulated do require ventilation and to achieve adequate air changes mechanical ventilation is probably needed with heat recovery.

So while electric heating may be easy to install and maintain the other systems such as ventilation with heat recovery become more complex, intrusive ( noise and inconvenience ) and expensive.

Doesn't affect wet gas CH vs electric.

Personally I like a leaky ( well ventilated ) house to ensure plenty of fresh air and avoid allergies and other illnesses which seem to be very common in modern well insulated houses.

In all of our rooms we have the small top-opening casements at (as they can be locked in this position) the slightly open setting - there's probably a gap of 20-30mm between the window and the frame at the bottom.

 

[unclebob1]

Cant you just use electromechanical relays to switch the 240V with the Raspberry Pi outputs? Its what we do every day in industrial automation systems. I dont know what voltage the Pi outputs are though, or what they can drive current-wise..maybe that's why you're faffing about making your own devices...

Think the problem with using direct electromechanical is the isolation of 240v and 12/5v devices and the lack of umph from the Pi to drive the relays... The Pi only takes a few Amps total for all its usage -itself plus its devices, most relays that i saw for 240v need a few amps each just to switch... if im not mistaken. As a result I'm having to faff about trying to find a solution, along with the Pi's temperature sensors having issues with longer distance cable runs not returning results correctly!

 

[bernardgreen]

If you are looking for reliable operation NEVER drive a relay coil from the output pin of a micro-processor. ALWAYS use a buffer chip such as ULN2803 to prevent back EMFs and other spurious signals reaching the processor's pins.

Selecting the right type of relay will give more than adequate separation between mains and ELV ( but beware some relay sockets do NOT provide the same separation as the relays they are made for ).

 

[Simon35]

Cant you just use electromechanical relays to switch the 240V with the Raspberry Pi outputs? Its what we do every day in industrial automation systems. I dont know what voltage the Pi outputs are though, or what they can drive current-wise..maybe that's why you're faffing about making your own devices...

Think the problem with using direct electromechanical is the isolation of 240v and 12/5v devices and the lack of umph from the Pi to drive the relays... The Pi only takes a few Amps total for all its usage -itself plus its devices, most relays that i saw for 240v need a few amps each just to switch... if im not mistaken. As a result I'm having to faff about trying to find a solution, along with the Pi's temperature sensors having issues with longer distance cable runs not returning results correctly!

http://uk.rs-online.com/web/p/interface-relay-modules/5337650/ .....5V coil, draws 7mA....why dont you think a relay will provide isolation between the coil and contacts?

What type of temperature sensors are you using? PT100's?

Glad I work with PLC's...all this faff isnt an issue. Done a few projects using PC based I/O and software, and they are always a pain.

Dont know how much I/O you need but would a programmable relay, like a Siemens Logo, or many others, not be an easier solution. I think you can get some with web servers if you want.

 

[unclebob1]

Cant you just use electromechanical relays to switch the 240V with the Raspberry Pi outputs? Its what we do every day in industrial automation systems. I dont know what voltage the Pi outputs are though, or what they can drive current-wise..maybe that's why you're faffing about making your own devices...

Think the problem with using direct electromechanical is the isolation of 240v and 12/5v devices and the lack of umph from the Pi to drive the relays... The Pi only takes a few Amps total for all its usage -itself plus its devices, most relays that i saw for 240v need a few amps each just to switch... if im not mistaken. As a result I'm having to faff about trying to find a solution, along with the Pi's temperature sensors having issues with longer distance cable runs not returning results correctly!

http://uk.rs-online.com/web/p/interface-relay-modules/5337650/ .....5V coil, draws 7mA....why dont you think a relay will provide isolation between the coil and contacts?

What type of temperature sensors are you using? PT100's?

Glad I work with PLC's...all this faff isnt an issue. Done a few projects using PC based I/O and software, and they are always a pain.

Dont know how much I/O you need but would a programmable relay, like a Siemens Logo, or many others, not be an easier solution. I think you can get some with web servers if you want.

Isolation etc is what i read on forums and other peoples experiences.
Sensors again - what people have tried and used with the Pi -(using adafruit website for most of the info) - initially sensor i was looking at was LM35, but was recommended to go with the ds18b20 sensor. however then i will not have enough IO's on the Pi for all the sensors and all the relay outputs... so realistically would potentially need something else to sit between the Pi and sensors or Pi and relays to do the controlling.

 

[unclebob1]

Simon35 - can you give me a bit more info on the PLC's you work on - model and or make so i can have a look to see if anything long those lines may be suitable?

 

[Simon35]

Simon35 - can you give me a bit more info on the PLC's you work on - model and or make so i can have a look to see if anything long those lines may be suitable?

Me personally, Siemens S7 and Rockwell...both of which would be massive overkill on capacity and price for your project...but google programmable relays, plenty suppliers make them, Siemens Logo, Scneider Zelio etc etc. You can get them relay outputs so you might find that you directly drive your valves etc.

 

[bernardgreen]

Have a look at the AD590 temperature sensor.

It can be very easily multiplexed for multiple sensor applications. Any one of 64 sensors can be selected to be connected to one analogue input on the processor using nine 4051 cmos analogue multiplexer chips and 6 digital ouputs from the processor.

Data and application here

http://people.ee.duke.edu/~jab/ece154/AD590.pdf

It is advisable to use the current amplifier at the sensor is there are long cable runs involved.

 

[kbdiy]

Not wanting to pour cold water on an innovative idea but am just wondering if the complexity of the OP's proposed configuration is justified by the outcome, i.e. such fine individual control of each room's temperature and will that make it more efficient? Who, apart from the designer would be able to fault-find on it and surely the introduction of so many controls is increasing the fault liability many-fold?

There are enough posts on here just pointing to faults with 2- or 3-port valves, let alone the endless possibilities this plan would generate.

Just curious if the end is likely to justify the means?

 

[Simon35]

Not wanting to pour cold water on an innovative idea but am just wondering if the complexity of the OP's proposed configuration is justified by the outcome, i.e. such fine individual control of each room's temperature and will that make it more efficient? Who, apart from the designer would be able to fault-find on it and surely the introduction of so many controls is increasing the fault liability many-fold?

There are enough posts on here just pointing to faults with 2- or 3-port valves, let alone the endless possibilities this plan would generate.

Just curious if the end is likely to justify the means?

Doubt the end will justify the means, but lets not let that stand in the way of engineering ...not sure it offers much over standard TRV's and it wont be cheap, but I suspect that the OP is fancies doing it as much for the process as for the end result. I'm sure plenty of us have been guilty of that.

If its properly documented, any decent engineer should be able to fault find on it.

 

[SimonH2]

My biggest concern with zoning with valves would be the on-off nature of them - which makes sensitive control difficult. When the stat turns on, the zone valve will open and the rad will get "fully hot" - and then the stat will turn off while the temperature overshoots. Then the cycle will repeat when the room cools down a bit.

Having said that, one of our neighbours from back in the 70's did just that. Each room had it's own stat and controlled a 2-port zone valve - I assume as an S-Plan-plus-plus-plus...

Over on the Navitron forums, there was some talk about experiments with adding a resistor to the sensor in a standard TRV. By putting current through the resistor, heat is applied and provides a falsely warm indication - adding a set back facility. By controlling the power added, it should be possible to achieve a controlled variation in TRV setpoint.

 

[ban-all-sheds]

Doubt the end will justify the means, but lets not let that stand in the way of engineering

Engineering?



"A good scientist is a person with original ideas. A good engineer is a person who makes a design that works with as few original ideas as possible."

Freeman Dyson