"By far the most costly" depends on how one values their time. When I had young kids, my time was very valuable and I spent as much time with them as possible so I would gladly spend on a completed and safe device instead of starting with a handful of parts and at least several hours of code-writing, testing, debugging without any assurances that the end result would be better than what I was starting with.
Time is money.

By asking the question, the TS has indicated a willingness to do some of the work themselves. And such a project does not require code writing at all. In addition, many people do not have young kids around.
In addition, there is not particular assurance that the Ring system will perform as desired.

 

By asking the question, the TS has indicated a willingness to do some of the work themselves. And such a project does not require code writing at all. In addition, many people do not have young kids around.
In addition, there is not particular assurance that the Ring system will perform as desired.

Maybe

 

My home thermostat is a simple RobertShaw which I think I bought at Lows Home Improvement. I think I paid about $30 USD for it about 10 years ago. While it has all sorts of programmable features I doubt I ever used them. I set on 72 F and forget it. I change two AA batteries about once every three years. When the new furnace and central air went in we adjusted and balanced the duct work. The thermostat sits on the first floor (two story wood frame building) about 5 feet above the deck with an 8 foot ceiling.

Years ago when I was curious I measured floor to ceiling and found about an 8 degree difference due to air layers (stratification). I can see no reason to develop and program a uC simply to maintain a comfortable house. Though I admit in playing around with home automation it might be nice to grab a smart thermostat. Merely more for amusement. If your thermostat is measuring the wall temperature and not ambient air temperature you may just want to replace it. I see basic heating and cooling home thermostats starting at about $20 USD and up from there. For basic creature comfort you really don't need much. You are not trying to maintain a laboratory temper controlled environment. For heating with a set point of 72 F when the temp drops to 71 it calls for heat and when it gets to 73 it shuts the heat off. The furnace behaves like any newer furnace and does a purge, then ignites (forced air natural gas). During the summer we still set and forget at 72 F and move it to cool. While it has programmable functions we never bothered.

I would just go out and buy an off the shelf turn key thermostat and be done with it. When I read your initial post I wasn't aware you only wanted a basic home thermostat control. I simply see no need for building a PID controller loop for a home application.

Ron

 

If the TS wants a much closer temperature control the first thing is indeed to get a much better thermostat than the one deigned to control a coal furnace. Much better thermostats are available .

 

Stand between a heat source and it’s sensor and you will notice the fluctuations, increasing as you near the source. An air handler with a modulated heat source, sensing discharge temp, will likely give you best results. In floor heating, in a similar way, gives a close sensation of consistent heat.

 

Back at post #1 the TS was explaining about their -proportional control and asking if there was a commercial equivalent available.
Certainly there are quite a few of them available that will work for heating and even cooling, which will use proportional control within some close temperature range and ON/OFF control beyond that range. Red LIon and Omron brands come to mind immediately, other brands are available. These are mostly time-proportional, not PWM versions, and that avoids the radiated interference possible from the fast switching of many PWM circuits. Time proportional with a period of a few seconds works very well for most heating systems.

 

First of all, the "Nest" hardware is some of the most costly hardware available, and no better than others.

For temperature control in a room, on/off time proportional with a period of several seconds is entirely adequate and simpler. The very simplest would be to have the heater blower run constantly and switch the heat element power with a low-hysteresis thermostat in the air stream. The immediate benefits would be no external wiring, and a much more stable discharge air temperature. It would still have temperature variations in the room, but with constant circulation they will be quite a bit less.

Years ago I designed a testing machine that used hot oil about 250 degrees F, with a switching cycle time of about 5 seconds. It was able to hold the temperature within +/- one degree , which was much closer than the +/- five degrees demanded by the customers. And the scheme was simpler to implement than faster PWM control.

 

It's been 145 days between post #26 and #27. I'd imagine the problem has been solved. But if I'm to add 2¢ to this thread - as I recall, I had a bi-metalic strip (BMS) thermostat (coiled) and a soft pull magnet. When the contacts got close enough the magnet would pull the contacts closed. It would take significant heat change to release the magnetic hold. The engineers were smart about this design, they put the contact on the end of a screw. The user could adjust the screw to hold the contacts away from the magnet so it wouldn't take so much heat change to get it to disengage. The end result was the ability to set the amount of hysteresis. As I also recall, some thermostats had a tiny nichrome wire that would warm the BMS and trigger a release. There was an adjustable wiper that could be moved to control how much energy that wire got.

IF you wanted to go with electronic control, my first thought was PWM. Using pin 5 of a 555 timer you could modify the pulse output. A thermal resistor (forgive the brain fart) could tell the chip how much temperature it's seeing. As the voltage on pin 5 went up (or down - I'm inexperienced with this pin) the on time would change to suit the environmental need.

However, since this is an old thread (not years old) this is probably a moot point at this time.

 

PWM is way overkill for an electric room heater. Minute by minute cycling can use an SSR and that same 555 timer with a temperature sensing thermistor to set the time cycles. Much less likely to cause RFI and no designing of high power switching systems. And keep the bimetal thermostat as an over-temperature safety shutoff.

 

Temperature sensor? Temperature range? You make no mention of a choice of temperature sensor or actual range you wish to control within. You are describing PID (Proportional Integral Derivative) control. Start with a dimple Google of PID Temperature controllers. You can find inexpensive units like this one likely from China or more expensive versions. They offer a few output types like you mentioned and offer a wide range of input sensors all user programmable. You choose temperature control based on your needs like how tight of a band you wish to maintain.

Ron

I use a ntc thermistor. The range is comfortable human being temperatures. These would be say 18c to 24c. My circuit is user selectable for the temperature. Thanks for the references.
The band I use is one degree centigrade. If you make it smaller than this you would get a lot of on/off switching.
Never been able to find PID controllers in any store. They only have on/off thermostats.
I am guessing your references are devices that the ordinary consumer would find too complicated to use. I bought one made by Berm Industries and made in China. Still can't figure out how to use it. But why bother as my circuit just asks the consumer to set his desired temperature.
Have a look at the Berm model. Sure its cheap but who wants to try to set it up to work.
I now have phase control via a random phase Triac optocoupler and variable duty cycle pulses coupled to a zero voltage crossing triac opto coupler for PID control.
I have both analog and microcontroller designs now.

 

The controller that I had selected for the machine that I mentioned used a thermocouple, but that was to assure the availability of repair parts if they were ever needed. Type J, K, and T thermocouples are well defined and thus far more available than thermistors, which are usually available in many proprietary variations. But for a one-off temperature control system a simple time proportional control based on a sawtooth wave generator and a comparator could be entirely adequate. No need for full PID (proportional, integral, derivative) control, just simply duty cycle with a one minute period. Add an opamp in between so that the sawtooth can be varied in amplitude and average level and the controller can provide simple ON/OFF control above and below the desired close control band. At that stage the entire system can be made with one quad opamp, (LM324), two sections to generate the sawtooth wave, one more section to control the amplitude and offset, and the fourth section to serve as a comparator, with a bit of hysteresis. The output can drive a darlington transistor such as an MPSA13. And the whole package can operate from a single positive supply voltage that can also power the thermistor portion.

 

Part of what needs to be considered is what type of heat source we're discussing. If it's hydronic (hot water) there's a bit of hysteresis in that by itself. You'd need to modulate the temperature of the water. It could be base board heat or in floor heat. I think I've even heard of heated ceilings, though I'm not sure on that one. If not hydronic, it could be electric heat. And that also comes in several variations. Electric base board or electric in floor heat. Even forced electric heated air. Then there's natural gas. That can be used to fire several of the afore mentioned systems. Gas, oil, electric, even coal and wood, those are still around. Each have their own hysteresis in their very nature. You can get heat out of hot water, but when the demand for heat is satisfied, the water is still hot and still warming the room. I had an electric oven I used for curing potted samples for micro sectioning. I wanted a specific temperature but the oven had horrible - what I call "Thermal Inertia" where the heating element would come on and begin to warm the oven. When temp was reached the power would shut off, but the hot element would continue to glow and heat the oven to well above the desired temperature. To achieve a more controlled temperature I had to set it lower than desired, then once it finished heating to that low temperature I would sneak up the temperature setting a little at a time. That got rid of some of that inertia, but the heat profile was much like a sine wave, even after moderating the temperature. So it's an important question to understand what kind of heat source you're trying to regulate. Just like a barbecue, temperature control using charcoal would require close human control. I suppose someone could build a temperature moderating device for a barbecue, but why?! Still, what kind of heat system are we talking about?

 

Part of what needs to be considered is what type of heat source we're discussing. If it's hydronic (hot water) there's a bit of hysteresis in that by itself. You'd need to modulate the temperature of the water. It could be base board heat or in floor heat. I think I've even heard of heated ceilings, though I'm not sure on that one. If not hydronic, it could be electric heat. And that also comes in several variations. Electric base board or electric in floor heat. Even forced electric heated air. Then there's natural gas. That can be used to fire several of the afore mentioned systems. Gas, oil, electric, even coal and wood, those are still around. Each have their own hysteresis in their very nature. You can get heat out of hot water, but when the demand for heat is satisfied, the water is still hot and still warming the room. I had an electric oven I used for curing potted samples for micro sectioning. I wanted a specific temperature but the oven had horrible - what I call "Thermal Inertia" where the heating element would come on and begin to warm the oven. When temp was reached the power would shut off, but the hot element would continue to glow and heat the oven to well above the desired temperature. To achieve a more controlled temperature I had to set it lower than desired, then once it finished heating to that low temperature I would sneak up the temperature setting a little at a time. That got rid of some of that inertia, but the heat profile was much like a sine wave, even after moderating the temperature. So it's an important question to understand what kind of heat source you're trying to regulate. Just like a barbecue, temperature control using charcoal would require close human control. I suppose someone could build a temperature moderating device for a barbecue, but why?! Still, what kind of heat system are we talking about?

Tony is reporting the effects of time delay on a feedback system. Those effects are very much real and can be modeled. They are also why much more complex control schemes exist, PID being a large set of them. That is why "feedback systems" is a two- semester-five credit hour upper level course in engineering schools. Just creating the math models is a challenge.

 

Many heaters use a mechanical type thermostat which is a simple bi-metallic bar that bends with heat and makes or breaks a circuit. These are terrible for comfort because you are too hot when on and too cold when off. The digital control ones are not much better. They are either on or off.
So I made my own what I call 'temperature controlled power'. It works like this. You set a temp of say 21c. At 20c maximum power is applied to say a 1000 watt heater. At 20.5c power applied becomes 500 watt and so on. Proportional power is applied according to the temperature. If the temp reaches 21c the heater would turn off. Practically what happens is that the ambient temp in the heated space stays between 20c and 21c.
While it works perfectly I have ended up with a birds nest of wires. I want something more compact so I am looking for a commercial product which will do the same thing and I wont have a jumble of wires.
I think RS Components stock number 188-5165 temperature controller may do the job. This has an SSR output as well as a ordinary relay output. I have not been able to find out yet if the output of the SSR is zero crossing or phase control. I am guessing it would be zero crossing as most are. Then all you have to do is apply PWM to the SSR input .
So does anybody on this forum know of a good commercial product that will have 'temperature controlled power'. I guess its easy to see why mechanical thermostats dominate the market because of the cost of 188-5165 for example. My own circuit is very simple and should be able to be made commercially fairly cheaply.

Back to the beginnings. Based on the very first post and reference to the heater I can only assume an electric heating system consisting of electric elements. Before even going any further a room, any standard room in a living quarters is going to have air stratification layers. Short of constantly moving the air around like a ceiling fan there will be air stratification layers. In my home I have pretty standard 8 foot (2.4384 meter) ceilings in a two story wood frame building. First floor living room is where the thermostat is about 5 feet above the floor. During winter heating I can measure anywhere between 6 to 8 degrees F differential between floor and ceiling. Natural gas fired heat but that is here nor there. My thermostat while a newer electronic type is still the traditional On/Off control. With this system when I set it on 70 degrees F it calls for heat the moment the thermostat drops to 69 F and shuts down at 71 F. With good insulation simple On/Off control does just fine. When the system went in the ductwork was adjusted for airflow in each room.

Now if we want laboratory type environmental control in a room it's another story altogether. It starts with not just the heat source but the entire room construction. We are also controlling beyond heat and cool but extending into humidity and the flow control to eliminate stratification of air layers. You want all the air in the room to maintain a set point temperature. There is much more to things than a thermostat or using PID control of an electric heating element. The fan moving the air never shuts down. Really tight temperature control generally uses a 4~20 mA loop driving a SCR bank. More than anyone will use for a residential heating system for creature comfort. A good controller will have an auto tune feature to develop a suitable algorithm. It will, as someone mentioned earlier "learn".

Years ago a good well made PID controller would cost hundreds of US dollars. Today I see PID controllers user programmable off the boat for under $30 USD, no clue how good they are but paired with a good SSR and having an SSR output they should do fine for controlling an electrically heated room or house. They offer a wide range of input sensor types too and are also commonly called "process controller" because they extend well beyond temperature control.

Ron

 

Back to the beginnings. Based on the very first post and reference to the heater I can only assume an electric heating system consisting of electric elements. Before even going any further a room, any standard room in a living quarters is going to have air stratification layers. Short of constantly moving the air around like a ceiling fan there will be air stratification layers. In my home I have pretty standard 8 foot (2.4384 meter) ceilings in a two story wood frame building. First floor living room is where the thermostat is about 5 feet above the floor. During winter heating I can measure anywhere between 6 to 8 degrees F differential between floor and ceiling. Natural gas fired heat but that is here nor there. My thermostat while a newer electronic type is still the traditional On/Off control. With this system when I set it on 70 degrees F it calls for heat the moment the thermostat drops to 69 F and shuts down at 71 F. With good insulation simple On/Off control does just fine. When the system went in the ductwork was adjusted for airflow in each room.

Now if we want laboratory type environmental control in a room it's another story altogether. It starts with not just the heat source but the entire room construction. We are also controlling beyond heat and cool but extending into humidity and the flow control to eliminate stratification of air layers. You want all the air in the room to maintain a set point temperature. There is much more to things than a thermostat or using PID control of an electric heating element. The fan moving the air never shuts down. Really tight temperature control generally uses a 4~20 mA loop driving a SCR bank. More than anyone will use for a residential heating system for creature comfort. A good controller will have an auto tune feature to develop a suitable algorithm. It will, as someone mentioned earlier "learn".

Years ago a good well made PID controller would cost hundreds of US dollars. Today I see PID controllers user programmable off the boat for under $30 USD, no clue how good they are but paired with a good SSR and having an SSR output they should do fine for controlling an electrically heated room or house. They offer a wide range of input sensor types too and are also commonly called "process controller" because they extend well beyond temperature control.

Ron

I am suspecting that the thermostat is on a portable electric heater with a fan, and if set at 70, it will heat until 75 is reached and then switch off until 65 is reached, maybe, or the swing may be a bit more. And because of stratification the actual temperature gradient may be greater yet. So a simple ON/OFF proportional time system will be a very huge improvement, and with a 1 minute frequency it will keep the air moving and not generate much RFI.

 

I am suspecting that the thermostat is on a portable electric heater with a fan, and if set at 70, it will heat until 75 is reached and then switch off until 65 is reached, maybe, or the swing may be a bit more. And because of stratification the actual temperature gradient may be greater yet. So a simple ON/OFF proportional time system will be a very huge improvement, and with a 1 minute frequency it will keep the air moving and not generate much RFI.

Now that you mention it the thread starter did mention 1 KW which coule likely well be a space heater. I wonder where it would be placed? Really, less knowing a whole bunch more here we sit.

Ron

 

First of all, the "Nest" hardware is some of the most costly hardware available, and no better than others.

This is a complete misunderstanding of what the Nest does.
The Nest uses its internet connection to query the current outdoor temp, it also determines the time-constant to heat your house a given temp and continues to fill out its database about your house's thermal time constants at each outdoor temp. Then it can intelligently keep your house within a very narrow range. Rules are in place to prevent short-cycling of air conditioner units and lag to start high-efficiency furnaces as they have pre-tests for powered exhaust vents.
some of the reviews that test the Nest in one afternoon are useless. It is advertised as a "learning" thermostat. It needs to be installed while the outdoor temp changes over a significant range. It doesn't do much more than a standard thermostat in a one-day test. It can also turn on the furnace fan only to circulate air to balance temp if steep gradients are detected on remote Nest Temp sensors (optional).
Claiming "no better than others" is simply funny

 

“Comfort” depends on temperature & humidity to control it correctly.

 

This is a complete misunderstanding of what the Nest does.
The Nest uses its internet connection to query the current outdoor temp, it also determines the time-constant to heat your house a given temp and continues to fill out its database about your house's thermal time constants at each outdoor temp. Then it can intelligently keep your house within a very narrow range. Rules are in place to prevent short-cycling of air conditioner units and lag to start high-efficiency furnaces as they have pre-tests for powered exhaust vents.
some of the reviews that test the Nest in one afternoon are useless. It is advertised as a "learning" thermostat. It needs to be installed while the outdoor temp changes over a significant range. It doesn't do much more than a standard thermostat in a one-day test. It can also turn on the furnace fan only to circulate air to balance temp if steep gradients are detected on remote Nest Temp sensors (optional).
Claiming "no better than others" is simply funny

Certainly it has lots of "features", but quite a few of them duplicate what cheaper systems do. And my big asertion is that "Net" is very costly. I noticed that was not challenged at all.

 

Certainly it has lots of "features", but quite a few of them duplicate what cheaper systems do. And my big asertion is that "Net" is very costly. I noticed that was not challenged at all.

In the case of Nest, you get what you pay for. I think the word "costly" is a useless description. Costly compared to a bimetallic strip - well, obviously. Costly vs benefit, no. It is a very good value from my point of view.

I installed a nest in one of my rental properties that has a 2-story family room (vaulted ceilings). Renters had complained about that room feeling drafty. I installed a nest and a remote nest sensor in the master bedroom - we haven't had a complaint since.