Currently I have a Honeywell thermostat in my living room. This device is somewhat intelligent in that it keeps statistics on how the heating performed so as to be able to control it more efficiently.
For example, when I want it to be 20 degrees celsius in my living room at 18:00 hour, the thermostat ‚knows‘ it takes an hour to get from the current 16 degrees to the desired 20. So at 17:00 hour, the thermostat turns the heating on so that at 18:00 the temparature will be 20 degrees.
Does anyone know how a FHT 80B copes with this? I mean, it also has a weekly schedule to set desired temperatures and the desired times. But is it also smart in knowing that it should start earlier with heating because it takes time to warm the house?
I can not seem to find this sort of technical data anywhere on the internet
I tend to agree with you, because when it would have had it, they probably advertise with it
I wonder how people who use these devices control their heating then? Do they set the time slot for 17:00 at 20 degrees so as to assure that it will be 20 degrees at 18:00? Or does IP-Symcon have software modules for this?
Is there a way to log the FHT80B data in IP Symcon and show the statistics afterwards? That way one could easily see if there is any ‚intelligence‘ present in the FHT80B.
What they do have is a learning algorithm that enables them to adjust to the output of the heater(s) relative to the room. You may have noticed the effect that they will either over or understeer during the first couple of days controlling a new room and/or a new heater or perhaps even during radically changing outside temperatures.
What they definitely do not have is the ability to reach a given temperature at a given time. Instead they will simply start regulating towards that temperature when the corresponding time slot begins.
Indeed I noticed the over and understeer. But I ask myself what good that learning algorithm will do if the thermostat just turns the heat onn at the given time it already should be warm…
I mailed the manufacturer ELV my questions also, but apparently they couldn’t be bothered answering my mail…
Ah well, I am thinking about writing something myself to control all my thermostats. Something which has a weekly/monthly program, and starts to heat the room in advance as to achieve the correct temperature at the given time. Nothing too fancy though. Perhaps a simple formula which considers the number of degrees a room heats per hour when the heating is on. And then, emperically, incorporate the effects of the outside temperature. I must find a temperature sensor first for that which can be used through my FHZ 1300PC.
The thermostat temperature sensor doens’t seem to be very accurate. If I place another thermometer next to the thermostat, the thermostat reads halve a degree or a bit more too hot.
Would be a nice project to play some with IPS. Perhaps a c#.net windows program (and a windows service on my server) which can talk to IPS through the web service interface
That would be the HMS100T or - if you also want the humidity - the HMS100TF. However, if the room in question is already wired for 1-wire, a 1-wire temperature sensor will definitely come cheaper.
FS20 is my first step in Home Automation so I don’t have anything from other providers (yet). I guess a wired solution isn’t something I want because I would have to install the wires.
Perhaps the HMS100FT will do fine if I can find it in a local webshop here. I wonder if I perhaps should look at zwave.
Outdated or not, they are still sold and may be the best (read easiest and maybe even cheapest) option if an FHZ is already in use. Sure, they are neither bi-directional nor would I trust the temperature to be any more exact than that of an FHT, but for an empirical measurement it’s working fine for me.
To compensate for more exactness you could use a static offset or a deviation table as is used to correctly interpret ship’s compass readings. Maybe even a deviation function could be found and applied.
From my point of view there isn’t really an alternative for the use with the FHZ. You could use one out of several wired systems or you have to buy an HomeMatic or EnOcean System.
But on the other hand… As long as the HMS is obtainable… :rolleyes:
Well it does not seem to be obtainable here in the netherlands but I could get it from Germany I guess.
I saw that HomeSeer has a temperature/motion sensor for zwave. Since IP-Symcon also supports zwave that could perhaps be an alternative? Of perhaps something from xbee although I haven’t got any knowledge of that yet
HomeMatic is much too expensive since you have to buy the central component of 400-500 euro. If I were to spend that much money, Xcomfort would be a far better professional choice anyway.
I am just reading this now and it reminded me a very old post from our FHT and IPS Heating Control Script specialist GGGss. After some deep research on the Forum, I found the post again. GGGss explained in a long post the entire science behind the FHT system. I dont know exactly how he found that out, probably by logging the temperatures, or the PID principle is something well known out there. This post convinced me at that time that there is a logic behind this system which was a condition for me to use it.
I have read the article but it does seem to talk more about a P.I.D. controller in general than it does about the FHT 80B. In fact, an FHT isn’s mentioned at all.
I’m sure the FHT uses thhe P.I.D. method to control the temperature. Although not very efficient I might add. When I look at the actual temperature, it can sometimes differ more than half a degree from the set temperature. Which is quite substantial in my opinion.
Also the FHT does indeed not seem to have any form of anticipation concerning the time schedule. When the temperature is set for 20 degrees at 18:00, the FHT start to heat at 18:00. And nog for example at 17:00 to reach the desired temperature AT the desired time. My Honeywell thermostat used to do this quite wel.
I think I will incorporate a bit of such logic in the FHT controller I’m writing myself.
It is indeed about the P.I.D. method in general but if I remember well was written to explain how the FHT works. I wouldn’t be surprised if the method is not efficiently implemented in the FHT but I don’t think I had differences of up to 0.5 degrees. I am wondering if the new Homematic FHTs are more efficient.
I had implemented a very basic empirical logic in my FHT heater scripts (adapted from GGGss script) with heating starting before the scheduled temperature setting of an amount of time directly proportional to the difference between the current and the target temperature. This required some adjustements for the different rooms. The temperature increase seemed more or less linear but of course the slope depends on size of the room and number, position of heaters.
Well. I find the FHT not very accurate in the temperature read out. I have to set the temperature half a degree to high to get the right actual temperature. I guess I have to incorporate this offet in my program too.
The heating of a room will also depend on the temperature outside. If it is colder then it takes langer to heat up and will get colder sooner. I want to incorporate that in my design too. But I won’t be using php scripts. I not familiar with it. I’m going to build a windows service in C# I think.
