1. I get a Can't call method "opened"
message when starting
fhem. Why?
You need perl >=5.6, on Unix with the Device::SerialPort (
http://search.cpan.org/dist/Device-SerialPort/),
on Windows Win32::SerialPort (use ActiveState's ActivePerl)
module.
If this module reports Can't call method "opened" on an undefined
value...
when starting the server, then you either may ignore
this message, or replace the mentioned line with:
$self->{HANDLE}->close if (defined($self->{HANDLE}) &&
$self->{HANDLE}->opened);
2. I get "undefined" messages in the log after upgrading fhem.pl
Stop fhem.pl, delete the previous .save file and restart fhem.pl.
If the problem still exists, send a bugreport.
3. I switched on a FS20 device directly (without the remote), but
the fhem.pl did not noticed it.
Is it a bug?
The protocol used by the FS20 family is quite simple: it is not
encrypted in any way, and there is no feedback to the sender. So if you
push any buttons on a pure receiver, no radio waves will be sent out, and
the FHZ1000 won't notice anything. The FHZ1000PC does not even know if
somebody received its message, it simply sends it out 3 times and hopes
for good luck.
To answer the question: it is not a bug :-)
4. I have some FS20/FHT/HMS devices. How do I know their housecode?
If you already programmed some devices, then just start fhem.pl with one
of the example configuration files, and watch the log. When activating a
device (e.g. with the remote) then it will be logged as an unknown device
with the housecode. Note: the verbose level must be 3 or higher. KS300
devices do not have a proper code, so you can use anything.
By the way, FS20 switches and dimmers are unidirectional. If you change their
states by pressing the connected hand switch, no information will be sent out.
You need to imprint it on a remote control or on fhem to set the housecode and
buttoncode. Please read the manual that came with the device how to do it. In
fhem simply define the device with a housecode and buttoncode at your liking
and switch the device via fhem while the device is in learning mode.
5. I have the code for my devices in the ELV notation, which contains 1,2,3
and 4, but fhem requires a hex code. How should I convert it?
Note: current fhem versions accept FS20 codes in ELV notation.
The code used by ELV is in the "quaternal" (?) system plus one added to each
digit, so you can even use the 4 button remote for programming. To convert,
you have two choices: the program four2hex in the contrib directory, or the
good old unix program bc.
6. I replaced my FHZ1X00PC, now the FHT80b's does not work anymore.
Help me!
The FHT80b's are talking to a single FHZ1XXX device, which has a unique
"FHTcode". You have two choices: if you know the old FHTcode, then you can
set it, or you can tell the FHT80b to forget the old FHZ, and start talking
with the new one.
- Set the FHTcode: The problem with this method is that I don't know how
to read out the value [nothing discovered yet either -AndiM],
we just can change it with
set FHZ FHTcode <hex-code>
(1 byte, range 1 to 99)
- Resync the FHT80b: Press PROG until "Sond" appears, then select "CEnt"
with the wheel, press PROG again, Select "nA" with the wheel, press PROG
7. I can specify an optional ml/raincounter for a KS300. Why do you think
that 255 should be the default?
The manual talks about 0.3l resolution, but I wanted to calibrate my device.
So I filled a plastic bottle with 0.5 liter water from the measuring cup,
drilled a small hole in the bottom of the bottle, let the water flow slowly
in the KS300 rain-cup, and looked at the counter after the bottle was empty.
The raincounter was incremented by 130 ticks. The diameter of my KS300
rain-cup is 13.9 cm, the area ca 151.75 cm2, which
is ca 1/65.9 m2.
The 0.5 liter corresponds to 32.95 liter per m2. 130 ticks
correspond to 32.95 l/m2 -> 1 tick is ca 253 ml. I estimate
the error margin to +/- 2%
You are welcome to do your own experiments, I am interested in the results.
8. The time specification of the builtin at command is not very
flexible. Please add day/month/weekday to it.
Please take a look at the holiday device, and the perl helper variables (both
described in the commandref.html).
9. I defined my FS20STR as an FHT device, but I do not get any data from
it.
The FS20STR is an FS20 device, even if it looks like an FHT80b.
You'll get "only" on-for-timer and off-for-timer events sent.
10. How to convert the FHT80b code seen in its display to the hex code needed
by fhem.pl?
Convert the first 2 digits first from decimal to hex, then the next two.
Example:
% bc
obase=16
<first two digits>
<last two digits>
E.g The FHT80b code 1121 is 0b15 for the fhem.pl
11. I'd like to use this sunrise/sunset stuff, can you help me?
Look for the geographic coordinates of your home, e.g with a GPS
receiver or with googleearth. Compute the latitude/longitude as needed, and
set the latitude/longitude global attributes.
{ sunrise() } now will return the time of the next sunrise,
in a HH:MM:SS format. Also note faq20
Note: 99_SUNRISE_EL.pm is the ExtraLight version of the original
99_SUNRISE.pm, which needs the DateTime::Event::Sunrise perl module, which
in turn is usually difficult to install.
12. I'd like to switch on the ventilator if the FHT tells me its too hot. How to tell fhem to do that?
To get used to the methods, we do it step by step. Telnet to fhem, and type
the following:
# Check simply the value. It is the same as seen in "list"
fhem> {Value(myfht)}
measured-temp: 23.8 (Celsius)
# Get the second word, so we can compare it.
fhem> { my @a = split(" ", Value(myfht));; $a[1] }
23.8
# Set the ventilator on now, if its too hot.
fhem> { my @a = split(" ", Value(myfht));; fhem("set ventilator on") if($a[1] > 25.0) }
# Now do this regularly
fhem> define chilldown at +*00:30:00 { my @a = split(" ", Value(myfht));; fhem("set ventilator on") if($a[1] > 25.0) }
# An alternative:
fhem> define chilldown at +*00:30:00 { fhem("set ventilator on") if(Value(myfht) gt "measured-temp: 25.0") }
13. I'd like to see directly in the Web frontend if a window is open.
How to do that? (by fwolff)
Just define a new (pseudo) FS20 device called "window" and use the following
notify settings:
define window_open notify bad.*warnings.*Window.*open.* set window on
define window_close notify bad.*warnings.*none.* set window off
14. In the summer I get a lot of "actuator:lime-protection
messages from my FHT80b. How to switch back to the actuator:0% messages?
(Thanks for Martin Fischer for researching)
The problem happens if at the weekly lime-protection time (usually saturday
at 11:00 AM) the desired temperature is lower than the measured temperature.
I think this is an FHT80b firmware bug, a workaround is to set the desired
temperature for a short time higher than the measured temperature.
You can automate it with the following notify:
define lime_reset notify .*lime-protection {\
$d = ReadingsVal("@", "desired-temp", 0);;\
$m = ReadingsVal("@", "measured-temp", 0);;\
if($m gt $d) {\
fhem("set @ desired-temp 29");;\
fhem("set @ desired-temp $d");;\
}\
}
15. My notify/filelog does not work, help me!
First learn what are the trigger messages: telnet to your fhem (e.g. telnet
localhost 7072) and issue the command "inform on". If there is an
FS20/HMS/etc event, then you'll see a line of the form
"TYPE DEVICENAME EVENT". Some devicetypes generate more than one trigger
message for one event, e.g. HMS or KS300.
Example: you have an FS20 device named piri, which sends a command
"on-for-timer 60". In this case you'll see: FS20 piri on-for-timer 60
Now define your notify/filelog. The regexp specified will be tested
against the DEVICENAME and against the DEVICENAME:EVENT strings. If they
match, then the action will be executed. For our example this would be:
define piri_ntfy notify piri:on.* some_command
You can test the notify/filelog with the trigger command: issue "trigger
DEVICENAME EVENT" in the telnet prompt, and check if your notify/filelog
works as designed. Example: trigger piri on-for-timer 60
Tip: you can test a new setup without a connected FHZ1x00. Create a new
config file, set the FHZ1x00 device to none, and generate events via
trigger.
16. Setting auto-mode from manual for the FHT won't change the desired temperature!
Thanks to Waldemar Porsche the following notify will take care of that:
define ExtendedAutoMode notify .*mode.*auto {\
my $olddesired = ReadingsValue("@", "desired-temp", 0);;\
return if (!defined($olddesired));;\
my @@day = ("sun", "mon", "tue", "wed", "thu", "fri", "sat");;\
my $now = sprintf("%%02d:%%02d", $hour, $min);;\
my $wd = $day[$wday];;\
my $desired;;\
if (ReadingsVal("@", "warnings", "none") eq "Window open") {\
$desired = ReadingsVal("@", "windowopen-temp", 0);;\
} elsif (ReadingsVal("@", "$wd-from1", "") le $now && ReadingsVal("@", "$wd-to1", "") ge $now) {\
$desired = ReadingsVal("@", "day-temp", 0);;\
} elsif (ReadingsVal("@", "$wd-from2", "") le $now && ReadingsVal("@", "$wd-to2", "") ge $now) {\
$desired = ReadingsVal("@", "day-temp", 0);;\
} else {\
$desired = ReadingsVal("@", "night-temp", 0);;\
}\
fhem ("set @ desired-temp $desired") if (defined($desired) && $desired ne $olddesired);;\
}
attr ExtendedAutoMode comment Setzt fuer den Auto mode die desired-temp in Abhaengigkeit von Temperaturprofil und Fensterkontakt
17. I have problem receiving data, I get "Bad CRC" or "Oversized
message."
Some Linux distibutions have problems with the settings of the serial USB
lines. Sometimes the 8.th bit is stripped, sometimes software flowcontrol
is enabled. Fhem tries to clear the serial line, but it is not always
successfull. Following command fixes the problem in some cases:
stty < /dev/elv_fhz1300pc\
-parenb -parodd cs8 hupcl -cstopb -cread clocal\
-crtscts -ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr\
-igncr -icrnl -ixon -ixoff -iuclc -ixany -imaxbel -iutf8 -opost\
-olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0\
bs0 vt0 ff0 -isig -icanon -iexten -echo -echoe -echok -echonl\
-noflsh -xcase -tostop -echoprt -echoctl -echoke
18. My FHZ1X00PC gets all time automatically disconnected and
reconnected. My log shows "USB device /dev/elv_fhz1300pc disconnected, waiting to reappear".
What is wrong with my USB?
Several fhem users are plagued by spontaneous disconnects/reconnects of USB
devices. This is not an issue of fhem but a problem of hardware and/or USB drivers.
See USB compendium for help.
19. How do I set the dim100% from the at command?
As the % and @ characters are special, they have to be escaped. See the
following list (copied here from the commandref.html, notify command
description).
- The character
%
will be replaced with the received event,
e.g. with on
or off
or measured-temp: 21.7
(Celsius)
It is advisable to put the %
into double
quotes, else the shell may get a syntax error.
- The character
@
will be replaced with the device
name.
- To use % or @ in the text itself, use the double mode (%% or @@).
- Instead of
%
and @
, the parameters
%EVENT
(same as %
), %NAME
(same as @
) and %TYPE
(contains the device
type, e.g. FHT
) can be used. A single %
looses its special meaning if any of these parameters appears in the
definition.
20. Why do my Sunrise/Sunset times differ from the ones on webseite XXX?
There are different ways of defining sunrise/sunset. To quote "perldoc
DateTime::Event::Sunrise" :
- 0 degrees
Center of Sun's disk touches a mathematical horizon
- -0.25 degrees
Sun's upper limb touches a mathematical horizon
- -0.583 degrees
Center of Sun's disk touches the horizon; atmospheric refraction
accounted for
- -0.833 degrees
Sun's supper limb touches the horizon; atmospheric refraction
accounted for
- -6 degrees
Civil twilight (one can no longer read outside without artificial
illumination)
- -12 degrees
Nautical twilight (navigation using a sea horizon no longer possible)
- -15 degrees
Amateur astronomical twilight (the sky is dark enough for most
astronomical observations)
- -18 degrees
Astronomical twilight (the sky is completely dark)
The default in fhem is -6, as we most often deal with lamps & reading. Most
websites use -0.833, the difference is normally about 30 minutes. You can
change it in 99_SUNRISE.pm by setting the $altit variable.
21. What is "unknown message: EOB" from a CUL device?
If too many messages for FHT devices are queued in CUL, the fht buffer
subsystem of CUL overflows. You get EOB (end of buffer) messages and
likely LOVF (limit overflow) messages, too. Either you tried to send too much
commands to the FHT80b, or (more probably) the communication to the FHT80b is
disturbed or the FHT80b is not paired with the CUL.
Since the FHT buffer is not cleaned automatically, you have to do it manually
with the set CUL raw T01abcd
(abcd= FHT house code) command, or
by unpluggin/replugging the CUL.
22. My FHT80b device does not send temperature messages to my FHZ/CUL/CUN (any more).
What can I do?
Please pair your FHT80b device withe FHZ/CUL/CUN as described in the
device manual.
It is advisable to remind the FHT80b devices from time to time that there is
a FHZ/CUL/CUN device to report to. Adding the following line to your fhem config
file helps in most cases:
define myReminder at *00:40:00 set TYPE=FHT time