Hi everyone,
I have been running my own version of 10_ZWave.pm for quite a while because the components I have did not work straight out of the box with Fhem. But I just noticed that the code has been updated a bit since last time I checked.
Upgrade looks good, but there is still a small problem with my temperature readings. Negative values ( below zero ) are not correct. I know that the sensors I have use "2's complement"-encoding of the value in the ZWave-message. I guess it will be so for all temperature sensors, and it would surprise me if this isn't true for all values in the SENSOR_MULTILEVEL -message.
Has anyone else seen the problem in other sensors with this messagetype? I guess temperature is the one type that you will actually see negative values for. Readings like power and humidity etc will all be positive or zero.
Anyway, I have patched the current version of 10_ZWave.pm so that it should work for negative readings, ig anyone wants to try it. At least it works for me. :)
--- 10_ZWave.pm.org 2014-02-08 18:08:55.309988918 +0100
+++ 10_ZWave.pm 2014-02-08 18:11:49.721212671 +0100
@@ -484,9 +484,12 @@ ZWave_ParseMultilevel($$$)
my $pr = (hex($fl)>>5)&0x07; # precision
my $sc = (hex($fl)>>3)&0x03; # scale
+ my $bc = (hex($fl)>>0)&0x07; # bytecount
+ my $msb = (hex($arg)>>8*$bc-1); # most significant bit ( 0 = pos, 1 = neg )
+ my $val = $msb ? -( 2 ** (8 * $bc) - hex($arg) ) : hex($arg); # 2's complement
my $ml = $ml_tbl{$type};
return "UNKNOWN multilevel type: $type fl: $fl arg: $arg" if(!$ml);
- return sprintf("%s:%.*f %s", $ml->{n}, $pr, hex($arg)/(10**$pr),
+ return sprintf("%s:%.*f %s", $ml->{n}, $pr, $val/(10**$pr),
int(@{$ml->{st}}) > $sc ? $ml->{st}->[$sc] : "");
}
regards,
Simen
I've applied your changes.