THZ / LWZ Tecalor Stiebel Eltron Heizung

[mlb]

First of all - thank you for this great thread. I have been user of Stiebel LWZ403 (+Solar), for 3rd year now. Until now
I've been using excellent heatpumpMonitor from Robert, then I found FHEM and I really liked the idea. I installed a template configuration from Willy and it was rather  easy to get it working (Wille - thank you!  ). However - I found two issues with FHEM where heatpumpMonitor still is better at:

1. Charts generation speed. Yes I know Robert generated pictures  in the background and so they could be displayed instantly, but also Robert used database to store measurements - while in FHEM as I understand, each time a log file is being parsed and that makes it working really slow, even with a decent CPU. So my question would be - have you tried to solve this already? I have seen DB option for FHEM, but frankly I am too new to this subject to try it without having examples, and my German is rather poor to use other threads at this forum...

2. Display of negative values (like temperatures when chart is expected to display a filled area: the shaded part is always from the bottom of the chart, not from zero axis... and so it looks rather strange and confusing.

--
greetings to all,
Marek

[immi]

Hi Marek
happy that you like thz. The idea behing fhemTHZis not to make a better monitoring, but to actively integrate the heatpump in a complete home automatization like fhem.
I have several hints for you.

• use a more powerfull fhem-server than a raspi
• help to improve the SVG module; actually if you update to the last version you should find some improvents: rudolf and    justme1968 are very active on the module, which has potential 
• keep logfiles small; store in logfile only what you need, not so frequently, and only if you have a variable has a change   
• test the dblog 
• generate periodically a plot as batch-job, look in the wiki, there was an example

The first hint is the most efficient: with less money you achieve a better result investing no time.
With the other suggestions, you spend lot's of time, but you learn something or use your mind, which is a nice side-effect.
My preferred way, is to keep logfiles as small I can, storing only what I need.
immi

[micomat]

Overclocking the RasPi to 1000MHz brought some more comfort to my system
Since Overclocking is no longer killing your guarantee on the RasPi it could be worth a try.

Markus

[mlb]

• use a more powerfull fhem-server than a raspi
• help to improve the SVG module; actually if you update to the last version you should find some improvents: rudolf and    justme1968 are very active on the module, which has potential 
• keep logfiles small; store in logfile only what you need, not so frequently, and only if you have a variable has a change   
• test the dblog 
• generate periodically a plot as batch-job, look in the wiki, there was an example

Hi immi, thanks a lot for response and explanation.

Actually for my FHEM test instance I already use a regular server (this is my home server Ubuntu 14.10 with 4 disks as RAID , Asterisk, and couple of other services), having Intel CPU, (Haswell G1820,2*2.7 GHz, a nice processor). I planned to move FHMI to raspi after initial tests (as I already use raspi for other purposes) - but yes, I will try other hints you suggested and will let you know.

Before I found FHEM- I tried to use software from openremote.org, but Java is somehow too heavy for my purposes, and FHEM is just a direction I like better   I am not a developer, still - I can do simple things in Linux environment and from that perspective FHEM is more user-friendly.

Just as a question about my second point and drawing of negative temperature values - have you seen this before? Or maybe there might be  something wrong with my setup? I will try to attach a screenshot:

(http://i58.tinypic.com/o6d9c3.jpg)

... so the blue area that represents negative outside temperature - should rather be between 0 and temperature value, and not between the value and -10

[willybauss]

Hi Marek,

Welcome to the forum! A few comments based on my experience: For me it looks like the bigger part of execution time is related to number of graphs shown in your charts. In the last few months I accidentally created THZ log files having >25MB per month. After having fixed the issue it is down to less than 10MB, but the time needed to create the charts doesn't decrease a lot. "Logging only if a variable has changed" might or might not work. If you log sGlobal only, you're logging lots of values in 1 single variable. So the chance that none of them changed is limited...

I've tested dblog months ago based on the assumption that getting the data out of a database should be faster than reading tons of log files. In fact it is not really faster. Additionally there where hurdles to overcome with the way 00_THZ represents the reading results: e.g. sGlobal shows multiple values in one single reading; so it doesn't make sense to put it into one single database entry, but you need to split them into single readings before logging them to the db.

If you succeed in "generate periodically a plot as batch-job" please let me/us know. I'm really interrested it it. It's not always necessary to have the last 5 minutes in the graphs, but a fast plot every hour could be sufficient in most cases.

For the negative values the solution could be to replace the Y axis from °C to Kelvin  ... For me this never had been an issue, so I didn't even notive it up to now.

I assume you got my configuration example from a file I attached to a post months ago. In this case please check your fhem.log file. I assume there are tons of error messages caused by the failure notifies, e.g.

Code Auswählen Erweitern

define Mythz.alarm_failure notify Mythz { if (((split ' ',ReadingsVal("Mythz","sLast10errors","ERROR"))[1]) > ReadingsVal("Mythz","number_of_faults_old","ERROR")) { DebianMail('MYEMAILADDRESS@t-online.de','Tecalor Mythz Alarm - ERROR','Fehlermeldung: '. $EVENT,'/mnt/fhem/Tecalor-Fehlermeldungen.txt');; fhem("setreading Mythz number_of_faults_old ". ((split ' ',ReadingsVal("Mythz","sLast10errors","ERROR"))[1]));; } }
attr Mythz.alarm_failure room Heizung

In my case the log file increased to >19MB within one day due to this topic. You need to replace all of the "ERROR" entries by 0, since fhem no more accepts them after one of the latest updates. So it should look like

Code Auswählen Erweitern

define Mythz.alarm_failure notify Mythz { if (((split ' ',ReadingsVal("Mythz","sLast10errors",0))[1]) > ReadingsVal("Mythz","number_of_faults_old",0)) { DebianMail('MYEMAILADDRESS@t-online.de','Tecalor Mythz Alarm - ERROR','Fehlermeldung: '. $EVENT,'/mnt/fhem/Tecalor-Fehlermeldungen.txt');; fhem("setreading Mythz number_of_faults_old ". ((split ' ',ReadingsVal("Mythz","sLast10errors",0))[1]));; } }
attr Mythz.alarm_failure room Heizung

(same for all the other failure notifies of course)

Do you still know when I posted this configuration file? I'd like to update it there in order to not create trouble to new users.

[mlb]

Thanks Willie, I will make necessary adjustments.

I downloaded your configuration from post you published on May 14: http://forum.fhem.de/index.php/topic,13132.msg168721.html#msg168721

It was not working initially due to missing global definitions, but once I copied them from default fhem.cfg - it started working I had small issue with remote control for Samsung TV and also your Weather tab gives errors, but I have not investigated reasons for it yet, most likely I will just remove it.

I'm glad to inform my graphs are working blazing fast now! I simply removed all existing log files, applied your advice to log only changed values - and it is so much better now - a big thank you

I will look at background plots generation in a couple of days, after I have more data available in logs.

[willybauss]

In the last few days I dealed with the charting frontend which looks pretty nice. Meanwhile it can handle file logs as well, so you can simply use it additionally to the "classic" web frontend without the need of additional db logging.

Advantages:

   
• hover mouse over the graphs to get measurement details (date + time, value); see screen dump #1
• zoom in for more detailed analysis ; see screen dump #2
• easily usable UI, much easier to use as the classic SVG plots

Disadvantages:

   
• graphs take even more time to show up than the SVG ones
• currently there's a bug inside, which needs to be worked around manually: http://forum.fhem.de/index.php/topic,10439.msg226324.html#msg226324
• in "our THZ module a value named "flowTemp" exists twice: in sGlobal and in sHC1; up to now I don't know how to select the desired one; I work around this issue using a userReading, which fills up the log file a bit

I decided to stay with the "normal" UI normally, but use the charting frontend for deeper analysis occasionally. Since it doesn't need additional effort/data (beside the userReading) this seems to be a valuable solution.

[peter456]

Hallo an alle FHEM Nutzer,

seit ca. 1 Monat versuche ich die Anbindung meiner LWZ303i Wärmepumpe an FHEM bzw. an Heatpumpmonitor einzurichten.
Mein Problem ist die alte Firmware meiner LWZ (V2.06). Und soweit ich das im Fhemwiki gelesen habe, ist dieser Firmwarestand nicht mit der Konfiguration "00_THZ.pm" kompatibel. Mit meiner LWZ sieht der ausgelesene Datensatz wie folgt aus (hinter einigen Hex-Zahlen habe ich die Dezimalwerte angegeben):

001BFB Wahrscheinlich Header
00C5 Kollektor 19,7°C
0016 Aussen 2,2°C
0142 Vorlauf 32,2°C
0115 Rücklauf27,7°C
02D4 Heissgas 72,4°C
0162 Warmwasser 35,4°C
0000 Temp. Heizkreis 2 (nicht vorhanden)
00E5 Kompressor 22,9°C
FFC9 Verdampfer -5,5°C
0142 Verflüssiger 32,2°C
19
1C
00
00
99
CE
00
00
11 Drehzahl Fortlüfter 17
00
0014 ?Aussentemp. gedämpft 2,0°C
00
00
00
00
00
00
00
01
00
1003 footer

Hat den vielleicht schon jemand eine ältere LWZ/THZ mit V2.06 mit FHEM im Betrieb?
Es wäre für mich kein Problem, wenn jemand lieber in Englisch antworten möchte.
Vielen Dank.

[willybauss]

Du solltest die ganze Antwort der LWZ hier einstellen. Idealerweise hängst Du das Mythz.log File an einen Beitrag an.

[peter456]

Hier ein Beispiel aus der LOG-Datei:

2014.12.05 15:09:43 2: Mythz/RAW: 1003
2014.12.05 15:11:37 2: Mythz/RAW: 0073FB00C5
2014.12.05 15:11:37 2: Mythz/RAW: 001F0102
2014.12.05 15:11:37 2: Mythz/RAW: 010401
2014.12.05 15:11:37 2: Mythz/RAW: 2D01C100
2014.12.05 15:11:37 2: Mythz/RAW: 0000E300
2014.12.05 15:11:37 2: Mythz/RAW: B400FD
2014.12.05 15:11:37 2: Mythz/RAW: 091C00
2014.12.05 15:11:37 2: Mythz/RAW: 0000C8
2014.12.05 15:11:37 2: Mythz/RAW: 000000
2014.12.05 15:11:37 2: Mythz/RAW: 00001A
2014.12.05 15:11:37 2: Mythz/RAW: 000000
2014.12.05 15:11:37 2: Mythz/RAW: 000000
2014.12.05 15:11:37 2: Mythz/RAW: 000000
2014.12.05 15:11:37 2: Mythz/RAW: 1003
2014.12.05 15:12:25 2: Mythz/RAW: 00FAF4001EFE
2014.12.05 15:12:25 2: Mythz/RAW: 3F0104FFFF
2014.12.05 15:12:25 2: Mythz/RAW: EA010201
2014.12.05 15:12:25 2: Mythz/RAW: 1B010300
2014.12.05 15:12:25 2: Mythz/RAW: 010001
2014.12.05 15:12:25 2: Mythz/RAW: 011C00
2014.12.05 15:12:25 2: Mythz/RAW: 640100
2014.12.05 15:12:25 2: Mythz/RAW: 000000DC01
2014.12.05 15:12:25 2: Mythz/RAW: 670000
2014.12.05 15:12:25 2: Mythz/RAW: 00DD0000
2014.12.05 15:12:25 2: Mythz/RAW: 0000F41003
2014.12.05 15:14:43 2: Mythz/RAW: 0063FB00C5001E01
2014.12.05 15:14:43 2: Mythz/RAW: 020103012A
2014.12.05 15:14:43 2: Mythz/RAW: 01C000
2014.12.05 15:14:43 2: Mythz/RAW: 0000E300
2014.12.05 15:14:43 2: Mythz/RAW: B400FC
2014.12.05 15:14:43 2: Mythz/RAW: 001C00
2014.12.05 15:14:43 2: Mythz/RAW: 0000C800
2014.12.05 15:14:43 2: Mythz/RAW: 000000
2014.12.05 15:14:43 2: Mythz/RAW: 001A000000
2014.12.05 15:14:43 2: Mythz/RAW: 00000000
2014.12.05 15:14:43 2: Mythz/RAW: 000010
2014.12.05 15:14:43 2: Mythz/RAW: 03
2014.12.05 15:16:37 2: Mythz/RAW: 006AFB00
2014.12.05 15:16:37 2: Mythz/RAW: C5001E0102
2014.12.05 15:16:37 2: Mythz/RAW: 0101012A
2014.12.05 15:16:37 2: Mythz/RAW: 01C000
2014.12.05 15:16:37 2: Mythz/RAW: 0000E300
2014.12.05 15:16:37 2: Mythz/RAW: B500FB08
2014.12.05 15:16:37 2: Mythz/RAW: 1C000000
2014.12.05 15:16:37 2: Mythz/RAW: C80000
2014.12.05 15:16:37 2: Mythz/RAW: 0000001B
2014.12.05 15:16:37 2: Mythz/RAW: 000000
2014.12.05 15:16:37 2: Mythz/RAW: 0000000000
2014.12.05 15:16:37 2: Mythz/RAW: 001003

Um 15:11:37 wird ein Datensatz ausgelesen (14 Zeilen). Die einzelnen Datensätze kann man daran erkennen, dass ein Datensatz immer mit "1003" endet. Nach genau 5 Minuten (Um 15:16:37) wird gemäß eingestelltem Zyklus der nächste Datensatz eingelesen. Und die Inhalte der Datensätze passen soweit auch zu den Anzeigen meiner LWZ (also sind es wohl keine fehlerhaften Daten).

[immi]

Peter
it is a good news. It works very well with your very old firmware.
Nevertheless some small (or big) fix has to be done to the THZ module in order to support yuour old firmware.
please use verbose 5.
In this way, all debug info will be logged.
Then post the logfile as file, because it will be huge.
best regards
immi

[peter456]

Hallo.
So wie ich das verstehe, muss schon einiges im THZ-Modul geändert werden, da zum Einen der Datensatz eine andere Länge hat und zum Anderen der Aufbau des Datensatzes sich komplett unterscheidet. Im HeatpumpMonitor von Robert Penz gibt es deswegen auch verschiede Module.
Im Anhang findet ihr den aktuellen Auszug aus der LOG-Datei nach der Umschaltung auf "verbose 5".
Vielen Dank im Voraus.
Peter

[immi]

Hi Peter
there are several things to do
at first I want to get rid of
2014.12.14 11:43:08 2: Mythz/RAW: 0000000000
2014.12.14 11:43:08 5: Mythz/RAW: 16000000
2014.12.14 11:43:08 2: Mythz/RAW: 16000000
2014.12.14 11:43:08 5: Mythz/RAW: 000000
2014.12.14 11:43:08 2: Mythz/RAW: 000000

these are comunication problems because your interface is very slow

try to uncomment line 679 from
#select(undef,undef,undef,0.010);
to
select(undef,undef,undef,0.010);

and restart fhem, then test and post again the logfile as file

immi

[peter456]

Hallo Immi,

habe sofort deinen Tipp mit "select(undef,undef,undef,0.010);" ausprobiert.
Nach dem "update 00_THZ.pm" und "shutdown restart" im FHEM habe ich mit Erstaunen festgestellt,
dass zum ersten Mal die Temperaturwerte meiner LWZ angezeigt wurden.
WOW! Dachte ich zuerst. Nach einer längeren Zeit hat sich dann herausgestellt, dass diese Werte nicht mehr aktualisiert werden.
Ein Reboot meines Rasppi hat auch keine Besserung gebracht.
Dann habe ich die Kommunikation im FHEM von Serial2Net auf /dev/ttyUSB0 umgestellt. Man muss hier noch dazu sagen, dass meine alte LWZ nur mit 9600kBps kommuniziert. Im THZ-Modul habe ich deswegen überall die Baudrate auf 9600 geändert.
Den aktuellen Auszug der LOG-Datei und noch zwei Bilder aus FHEM habe ich im Anhang beigefügt.

Gruß
Peter

[willybauss]

Probiere mal, was sich tut, wenn Du im ersten Bild bei FanSelectorDay auf eine der Zahlen tippst. Eigentlich sollte dann die Anzahl der ausgemalten Flügel im Symbol sich ändern, und die Lüfterdrehzahl ebenfalls. Wenn das klappt, hast Du eine bidirektionale Verbindung, die prinzipiell funktioniert.

Lass mich raten: Du hast kein 2. Display im Wohnzimmer? Deshalb meldet die Anlage 0% Luftfeuchtigkeit (das Display in der Anlage hat keinen Feuchtesensor eingebaut), die über das UserReading auf 11,5% korrigiert werden. Mit der Innentemperatur dürfte es sich ähnlich verhalten.

Die restlichen Werte im 1. Bild sehen vernünftig aus, bis auf FanSelectorPassiveCooling. Den liefert die Anlage vermutlich nicht mit.

Bei näherer Betrachtung des Logfiles frage ich mich allerdings, woher fhem die angezeigten Werte überhaupt nimmt. Und die Kommunikation scheint auch noch instabil zu sein. Dazu muss immi sich äußern. Er ist der Spezialist für sowas.