FHEM mit CUL + FHT8V-2 ohne FHT80B-3?
Terminar
ich steige gerade in das Home-Automatisierungsthema ein und bin
relativ schnell zu FHEM gekommen.
Anfangs hatte ich vor, von ELV das FHZ1000PC zu holen mit den
Raumreglern ELV FHT80B und an der Heizung dann die FHT8V Stellantrieb
an die Heizung.
Nun habe ich bei FHEM gelesen, das von Busware das CUL (CC1101 USB
light - V3???) auch funktioniert bzw noch mehr unterstützt.
Wäre es möglich nur mit FHEM und dem CUL die FHT8V Stellantriebe
direkt anzusteuern und abzufragen bzw die Temperatur zu setzen oder
benötigte ich zwangsweise das FHT80B-3 Raumsteuerungs-Ding?
Gibt es alternative Geräte von einem anderen Anbieter oder anderem
System, das vielleicht "besser" mit FHEM und dem CUL funktioniert?
Bis denn
Björn
Hodosi Szabolcs
Unfortunately I can't understand a word in german, but it seems you want
to deploy the same architecture as I also want.
I ordered a CUL and I have FHT8V (v3), FHT8R-3 (I cannot do anything
with it...).
As per long discussions with others FHT80B can be used as a temperature
sensor and CUL is able to read the reportings.
If you simply don't associate the FHT8V with the FHT80B-s it won't
control your valve openings, what you can do with the CUL.
You can bind a PID controller algorithm (almost "self-learning") between
an FHT8V + a temperature sensor.
FHT80B as a temperature sensor only is a little bit expensive, you can
order S555TH,S300TH from ELV only for 15 EUR per each.
(as the forum users stated), despite I think it's hard to find such
temperature sensors out of those two (or La-Crosse TX2/TX3, which is
also used to out of stock everywhere).
I hope this method will work..
I'm open to do private email discussions on it, or keep it being tracked
herein the forum if others are also interested...
All the bests,
Szabi
2011.11.14. 22:06 keltez�ssel, Terminar �rta:
> Hallo,
>
> ich steige gerade in das Home-Automatisierungsthema ein und bin
> relativ schnell zu FHEM gekommen.
> Anfangs hatte ich vor, von ELV das FHZ1000PC zu holen mit den
> Raumreglern ELV FHT80B und an der Heizung dann die FHT8V Stellantrieb
> an die Heizung.
>
> Nun habe ich bei FHEM gelesen, das von Busware das CUL (CC1101 USB
> light - V3???) auch funktioniert bzw noch mehr unterst�tzt.
>
> W�re es m�glich nur mit FHEM und dem CUL die FHT8V Stellantriebe
> direkt anzusteuern und abzufragen bzw die Temperatur zu setzen oder
> ben�tigte ich zwangsweise das FHT80B-3 Raumsteuerungs-Ding?
>
> Gibt es alternative Ger�te von einem anderen Anbieter oder anderem
> System, das vielleicht "besser" mit FHEM und dem CUL funktioniert?
>
> Bis denn
> Bj�rn
>
Zrrronggg!
S300TH as temp sensor.
I personally have not done it yet, but all the pieces are there.
I personally use FHT80bs. They have at least the advantage, that you
can adjust the desired temperature in the room without involving FHEM,
very important for the WAF.
Also measuring the temp and opening the valves to a certain degree
needs some additional FHEM-coding, while with the FHT80b you simply
send the desired-temp. Furthermore you have to take care for window
contacts yourself and many other things.
So, the advantage is, that it costs about 15-20 euros less per room,
but I'm not certain that the extra time and effort in codeing the
whole stuff in FHEM is actually worth it.
Only my opinion.
>Wäre es möglich nur mit FHEM und dem CUL die FHT8V Stellantriebe
>direkt anzusteuern
Ja.
>und abzufragen
weiss nicht.
>bzw die Temperatur zu setzen
> oder
>benötigte ich zwangsweise das FHT80B-3 Raumsteuerungs-Ding?
nicht unbedingt.
Die FHT8V kennen keine Temperatur sondern nur Öffnungsgrade. Man kann
mit FHEM und CUL den Teilen sagen, sie sollen das Ventil z.b. zu 50%
offnen.
Um eine Temperatur einzustellen braucht man noch ein Termometer, z.b.
das günstige S300TH. Allerdings muss man dann eine Menge Logik selber
machen, wie die Temperatur überwachen und die FHT8v selbst
entsprechend anpassen, incl. so Dinge wie Schwingen der Temperatur
dämpfen und Nachabsenkung selber Programmieren, Türsensoren soweit
vorhanden selbst überwachen, eine Bedienoberfläche bauen, weil man ja
nicht im Zimmer direkt steuern kann etc. etc. Der dazu notwendige
FHEM Code ist vielleicht nicht gerade was für Anfänger.
Die FHT80b erledigen das alles für einen, hier muss man nur die
Zieltemperatur eingeben, fertig. Ferner bietet das dort
einprogramierbare Wochenprogramm eine prima Fallbacklösung, sollte
FHEM mal ausfallen UND man kann die Temperatur im Raum an einem Regler
einstellen... anstatt eine Weboberfläche von FHEM benutzen zu MÜSSEN.
Aus mindestens diesen Gründen ist eine FHT80b-lose FHEM-
Direktsteuerung der FHT8v zwar möglich aber meines Eindrucks nach auch
in den Kreisen dieser Google Group nicht sehr verbreitet.
Rudolf Koenig
> additional FHEM-coding
I agree with all your arguments, but this one should be handled by the PID
module.
Hodosi Szabolcs
You don't need to code anything since the PID module of FHEM does this
job for you.
If you want to use the FHT80B's own internal algorithm of controlling,
then you should count with that you cannot modify the logic of it.
FYI, the FHT8R-3 (the one-directional version of FHT80B, very similar to
it) does a very bad logic of controlling.
In my flat when only one room needs to be heated it starts heating, the
water is 70 C degree, but the radiator valve opened only for 8% - 10% if
1 C degree is a the difference between the desired and current
tempeature. On the one hand it's not effective is gas consume viewpoint,
on the other hand a lot of MJ of heat disappear by self-cooling of the
valves, boiler etc...
PID (or my one I've started develop as a private home project)
controlling is much more effective and recommended.
All the "self learning"/"self-tuning" thermostats are PID based.
I hardly believe FHT80B implements PID, but who knows...
Can anybody confirm the FHT80B implements PID or not?
Cheers,
Szabi
Terminar
wow, awesome help, thank you very much for the informative answers!
> If you want to use the FHT80B's own internal algorithm of controlling,
> then you should count with that you cannot modify the logic of it.
> FYI, the FHT8R-3 (the one-directional version of FHT80B, very similar to
> it) does a very bad logic of controlling.
AF) is nice but that
doesn't really count. It is more interesting that the main point is
the temperature sensor
itself is in the FHT80B and Zrrronggg! mentioned the correct question,
i am not sure if
it really counts to have 15-20 euro less per room...
> All the "self learning"/"self-tuning" thermostats are PID based.
relation to a temperature sensor?
Hm, both possibilities are great, using fhem+cul directly with the
FHT8V and using the FHT80B but
I am not sure if that is really needed for us, it seems just to be a
matter of taste.
We own iphones and i saw that there are really nice gui apps for fhem
in the app store. And the second possibilty
is using my dreambox (set top box) to just control the fhem, would be
nice to implement (programming is
not a problem for me). Or just webinterface...
I think i have to install fhem first without any devices to see how it
works and how i can use it... Maybe this
will be a help for my decision.
Bye
Björn
Hodosi Szabolcs
Web UI / Android UI / iPhone app doesn't really matter though.
You will build the heat controlling system for the purpose: not care
about it.
Once it's setup and configured for your taste you won't monitor it all
the time.
The PID is a general controlling mechanism used in
automation/controlling industrial area.
Theories shows that for heating the most effective if the heating is
controlled by such a device which has PID implemented.
I'm not a professional on industrial controlling but let me explain a
little it.
Boiler warms the water in the heating system when a heating demand
triggered (by the thermostat).
The pump is circulating the water for effective heating until the
temperature of the water in the heating system falls under a given value.
Every modern boiler has a built-in electronic for that.
An usual setting used to be: heat the water to 50 C degree, and don't
fire gas until the temperature of it fall under 30 C degree.
Water transfer the heat in the heating system, so the water temperature
will fall constantly, so when it goes under 30 C heating starts again,
water is warming etc...
Until there is no heating demand..
Controlling as industrial area comes into the picture at this particular
point:
Water will circulate in the system after the desired temperature
reached, so the rooms will be overheated, which costs money.
We can basically save this money knowing that heating in general has two
parameters comparing to for example a electrical controlling
(Phase-controllers for example):
1., Water has a heat "stock", depends on its own actual temperature,
electricity doesn't. When you switch off the light there will be dark
instantly. If you switch off the heating, some heat will leave through
the radiators for another half an hour or so.
2., Either you command to the heating, the reaction time is big, the
result of your command will appear with an approx 10 minute shift.
Controlling is for fix these two behavior.
Level 0: Stop-Start Thermostats (which you most probably have now):Works
as I described upon.
Level 1: P-Controlling (P-Regler): If I need 2 more celsius, then
calculate how much heat needs for it and open the valve only to that
percentage to needs: Overshots (~ overheating) always realized, and you
cannot eliminate it. This works like for example in an 1 hour heating
period: open the valve for 25%, keep it there for awhile then check how
many heat is needed, change the value to for example 5%, then finally to
zero.... It reacts to changes too late, if you want 23 C degree, it will
always be reached, but more than needed in every case (for example
+5-10%). The valve openings defines an F(x) function, which is
approximately linear.
Level 2: PD-Controlling: D is the derivate in mathematical manner, it
nominate the F(x) function with the derivate, which can radically
enhance the speed (in heating time, and in statistically the overall
heating time need to reach the desired value), there are overshots, but
can be eliminated faster than we saw in P-Controller.
Level 2.5: PI-Controlling: Theoretically possible, but not realized, no
use for that.
Level 3: PID: P = Phase, I = Integral, D = Derivate... Integral
addition of the F(x) function responsible for decrease the overshots,
the side effect is there will be undershots always, but in time less and
less...
Lim(FOs(x)) = 0 (Os is overshot, F(x) is the function). What does it
mean practically:
First of all: No more energy (money) is consumed than needed, a little
bit less always, so if you want 23C you first will get 22,7 then 22,8
then 22,9... etc....
The F(x) starts being logarithmical...
Details: http://en.wikipedia.org/wiki/PID_controller
If we are so clever, why we don't calculate exactly the heating need,
why we approximate?
Because heating needs depends on such a large number of parameters that
makes electronical too complicated to implement AND unable to
mass-manufacture for end users...
A few of them:
- Difference between the desired temperature and actual temperature
- Outdoor temperature
- Area of windows in the room
- Material of the windows
- Is it sunny weather or windy and couldy
- The wall area looks for outside
- Material of the wall
- If there is a heating from top and bottom (for flats)
- Temperature of the water in the heating system
- Count and effectivity of the radiators
- Effectivity of the boiler
- temperature of other inner side of the walls... etc etc etc...
Something like that.. I'm having dinner now :)
2011.11.15. 15:40 keltez�ssel, Terminar �rta:
> Bj�rn
>
Terminar
> Once it's setup and configured for your taste you won't monitor it all
> the time.
beautiful :)
> Something like that.. I'm having dinner now :)
about that as deep as you explained.
I found two failures when i thought about how i would heat up the
room. I thought "ok, push it to 100%,
wait until it is ~2° under the target temperature, then put it to 30
or 50 until the temperature matches
around the target temperature." (or something like that). Maybe that
is not really clever...
The PID does the job better alone.
The second failure: I forgot the outside temperature, sun heating,
water temperature and more.
I talked to my wife (and the WAF isn't a problem because she is a
geek) but (!) she said it will be better to
have a room control to enter the temperature instead of only having
web/phone/remote stuff. :) Interesting...
Can anybody tell me if i can use RFM12 devices with FHEM?
I haven't found any info on that devices but i have some power
switches which can be controlled via RFM12....
It will be nice if i can use them in the new planned system with fhem.
Bye
Bjoern
Hodosi Szabolcs
> Hi Hodosi,
>
>
>> Once it's setup and configured for your taste you won't monitor it all
>> the time.
> True - and sometimes play a bit with the system because of working so
> beautiful :)
Hey, okay, it's a hobby too :)
>> Something like that.. I'm having dinner now :)
> Thank you very much for the detailed information. I haven't thought
> about that as deep as you explained.
> I found two failures when i thought about how i would heat up the
> room. I thought "ok, push it to 100%,
> wait until it is ~2� under the target temperature, then put it to 30
> or 50 until the temperature matches
> around the target temperature." (or something like that). Maybe that
> is not really clever...
> The PID does the job better alone.
> The second failure: I forgot the outside temperature, sun heating,
> water temperature and more.
Behavior like you mentioned is more or less a P-Controller, that's my
plan too at home, with some AI enhancements
(based on historical data figures out when - at which percentage - the
falling point will be, or should be, if for example I want the desired
temperature in 10 mins, 1 hour.
Time is also a factor what PID don't cover, but approxing to it by
mathematical way.
One more thing I forgot to describe is that PID makes the heating slower
in time, so it's fine when you have long time periods, like: whole day,
but when the period is shorter I'm not sure in that
it will serve an optimal F(x) function.
Moreover, I read that since complete calculation of the PID F(x) is
costy, takes a lot CPU time, all the algorithm (like Alexander's in FHEM
repository) needs good starting points.
A good choice of P,I,D variables can short and make the algorithm more
effective, however bad values makes is worse than a P-Controller!
So it needs to be handled with big experience and extra care...
"Learning thermostats", like the american "Nest" is trying to tune these
P,I,D variables in a right way with numerical approximation.
FHEM - as I investigated the perl source code - doesn't have learning
features now.
(I'm on develop it, but in Java, so out of FHEM world. My code will use
FHEM in cooperative way, but not exclusively).
So controlling is a profession, and I'm not a professional, just a
hobbyst :)
>
> I talked to my wife (and the WAF isn't a problem because she is a
> geek) but (!) she said it will be better to
> have a room control to enter the temperature instead of only having
> web/phone/remote stuff. :) Interesting...
I would agree with her. But as you may know, then FHT80B cannot be
controlled, it's only reporting information to higher level (FHZ1300PC,
CUL).
It would be nice to know how FHT80B firmware is working exactly, which
algorithm does it using.
For example in FHZ1000 (not the -PC version, but the standalone) I found
menu item "H-Regler", and this is the menu where I can bind FHT80B to
this device.
I assume all the FHT80B, FHT8R, FHZ1000 is using H-Regler, H-Controller.
I found some wikipedia references only, but nothing detailed description.
Either way, today it opened the FHT8V to 1%, the radiator was ice-cold,
and the boiler worked as a mad...Frustrating..
If this is the H-Controller, I don't want it.
I personally have never heaord of RFM12, but I'm a niewbie in the FHEM
world... maybe the wise guys knows much..
Szabi
Terminar
hm, it seems that google groups has eaten my answer... Once again...
On 16 Nov., 00:26, Hodosi Szabolcs <hodosiszabo...@gmail.com> wrote:
> I would agree with her. But as you may know, then FHT80B cannot be
> controlled, it's only reporting information to higher level (FHZ1300PC,
> CUL).
that the FHT80B
can be paired with fhem due to the fact that it is sort of
bidirectional communication.
When paired it sends the temperature to the central unit (every 15
minutes) and the
FHEM can send the new target temperature or week plan to the FHT80B.
That's enough for me at the moment. But i have also read that the
communication
with the FHT devices is really problematic due to the 1% rule where
sending and receiving
data is only allowed in 1% of one hour and the bidirectional part
needs some messages
like acknowledge answers to proceed the commands...
Is the use of the FHT80B paired to FHEM sort of stable or useless
unstable? I only want to control
a maximum of 5 rooms (FHT80B) with FHEM.
If this is not a good idea, are there other heater control systems
working with fhem which are
working better?
> I personally have never heaord of RFM12, but I'm a niewbie in the FHEM
> world... maybe the wise guys knows much..
switches operating
at 433MHz. Some companies which uses RFM12 are Tevion,
Intertechno, ...
I played a bit with some power control switches and the AVR-NET-IO
device from pollin (flashed with ethers*x firmware) some time ago.
With a HopeRF RFM12 module (also available at pollin) , these devices
can be controlled. Some information about that can be
found at http://old.ethersex.de/index.php/RFM12_ASK
I read that fhem can also send ECMD, that is the control language for
ethers*x, maybe the RFM12 devices can be controlled with fhem then?
But it seems that i have to try it myself :)
Bye
Björn