THZ / LWZ Tecalor Stiebel Eltron Heizung

[micomat]

wiki updated

[immi]

Hi Markus, Dear all
in the wiki there is still a hint on the connect-disconnect problem.
I do not see it anymore, so much has been done to fix it.
Actually my system is running so stable that I do not need to periodically restart.
I would like a short feedback on stability and on possible strange events that you find in logfile.

immi

[houseowner123]

I still see it frequently in the logs and I do not use the serial but the straight forward USB connection. Could this be a voltage issue of the THZ circuit board that the powering has small hickups and goes below 5 V for a second or so?

Tom

[micomat]

i saw it until i updated today to 0.180. will keep you informed if 0.180 fixed it for me.
yesterday i've had some disconnect entrys.

[willybauss]

I saw it just once: yesterday - 1 day after having updated to 0.108.
Couldn't find it in log file of whole June. But I don't use ser2net, I use a USB2RS232 adapter instead.

I do also not see any further THZ related strange messages in any logs.

I check fhem's status (and restart fhem if necessary) on hourly base via a cronjob plus shell script. I avoid intentionally to do so inside fhem, since this automatism should fix issues like crashed or hanging processes. Therefore it doesn't make sense to execute the healing process inside the crashed one ... Works perfect so far. As soon as I manage to kill fhem from remote I have to wait maximum an hour until it works again.

[immi]

Dear all
when you update, you have to restart FHEM ASAP.

Dear Tom
in a direct connection to the heatpump, everything could happen.
EMC (EMV in german) is a very bad beast.
Maybe can you correlate the hour of the event and a booster or compressor activation...
immi

[houseowner123]

I am able to graphically display when the disappear occurs.
Problems: I have no idea what is measured. And whether it is correlative or causative.
But the overlay is striking and I believe that everybody else can test it.
Measure is called: mythz.*    then select column 4 , plotstyle lines. These lines are pretty regular and line up somehow with daily wrg energy. However, exactly when disappear occurs, this parameter shoots over the wrg line.

If this is too little, tell me what else is needed. I will post details later tonight. I am in a hurry.

Tom

[immi]

I am able to graphically display when the disappear occurs.
Problems: I have no idea what is measured. And whether it is correlative or causative.
But the overlay is striking and I believe that everybody else can test it.
Measure is called: mythz.*    then select column 4 , plotstyle lines. These lines are pretty regular and line up somehow with daily wrg energy. However, exactly when disappear occurs, this parameter shoots over the wrg line.

If this is too little, tell me what else is needed. I will post details later tonight. I am in a hurry.

Tom

Try to behave like a Geologist o a big-data-expert, for whom a correletion is enough to prove a deterministic demonstration
joke at a side
pay attention with what you see: mythz.* colum 4 will give you several parameters in one plot.
you have to better filter out... I suggested to filter  compressor on or booster on... but could be something else .. e.g. a valve which is switching.
I cannot reproduce your test because my fhem is running 10 meters away: it is connected with a long serial cable and I forward it with ser2net.
I do not have a connect-disconnet from more than 50 days.
immi

[micomat]

just found:

Code Auswählen Erweitern

2014.07.03 03:50:35 1: 192.168.178.50:2003 disconnected, waiting to reappear (Mythz)
2014.07.03 03:50:35 1: 192.168.178.50:2003 reappeared (Mythz)
2014.07.03 03:50:36 1: 192.168.178.50:2003 disconnected, waiting to reappear (Mythz)
2014.07.03 03:50:36 1: 192.168.178.50:2003 reappeared (Mythz)

[houseowner123]

Hi Immi,

I did what you suggested and looked deeper into correlating events. I can find three categories. The disconnect is associated with
1. turning off the heat pump after WW activity (account for 60%)
2. return of power (my energy supplier turns the power off for 2 hours every day) (account for 30%)
3. nothing I can find. (account for 10%)

I conclude from this that it is an electrical or better shielding problem. Whenever the power turns off or on it may induce a small current in the USB cable/connector which makes FHEM believe the connection is lost. This also would explain that connection is almost immediately restored.

I am still fascinated that I graphically display this all, I just have no idea what it is. However the disappearing events look to me like a strange current. I still want to say that the occurrence i still random and not coupled to anything. I had 17 events in June; 2 already in July.

Here are some screen shots, each representing the above mentioned events in that order.

1) (http://img5.fotos-hochladen.net/uploads/heatpumpendspje1w8vtb.jpg)


2) (http://img5.fotos-hochladen.net/uploads/generalpowero8idbznutow.jpg)


3) (http://img5.fotos-hochladen.net/uploads/unrelateda7trb3cwpy.jpg)

I'd be happy if someone could tell me what the green mythz.* parameter is that is so nice an regular.

Tom

[willybauss]

did you ever try to use a different power outlet for your fhem server's power supply, being far away from the heatpump? You might need to use a extension cable to try this.

[micomat]

mine is placed upon the heatpump but i also tried to move around the power supply. didn't work.

[houseowner123]

I can confirm this. The power lines are completely separate but of course in the same room. EDIT: Sorry for the big images, I make them smaller next time.

Tom

[willybauss]

I'd like you to share the code for the diagrams. Some of them look interresting.

[micomat]

why using code for the diagrams? you can choos the values of the diagrams in the plot-gui. thats easier.