Unsigned current Power Value

[SirkoVZ]

According to the SML protocol
https://www.vde.com/resource/blob/951000/252eb3cdf1c7f6cdea10847be399da0d/fnn-lastenheft-edl-1-0-2010-01-13-data.pdf
page 31 the direction (consumption or delivery) is marked using flag.
It seems the SMLlib is not honoring the flag, as it always shows positive values, even when delivering power.
At least in the EMZ ED300L no negative values are shown when delivering power
The (un)sign seems to be addressed already in Tasmota:
arendst/Tasmota#8001 (comment)
and finally in
arendst/Tasmota@61b9735
Maybe we can also add this in the SML lib...

[r00t-]

can you attach or link binary dumps of sml transmissions with both variants?
(to enable developing/testing an implementation without owning an affected meter.)

[SirkoVZ]

How do I do that?
dd if=/dev/ttyAMA0 of=/tmp/dumpfile.bin bs=1 count=4096 iflag=fullblock
creates an empty file, while vzlogger works properly using /dev/ttyAMA0

[r00t-]

@SirkoVZ:
in theory that should work,
bute note that using dd is never necessary, and kind of silly on a character device.
just cat /dev/ttyAMA0 >logfile
if in doubt also try xxd </dev/ttyAMA0 (a hexdump also works, as it can be converted back. and actually it can simply be posted here as text.)

[r00t-]

oh, look, it's all here:
https://github.com/devZer0/libsml-testing
you could just submit your dumps right there

[r00t-]

yay, data!

first results look good:
i added this:

[~/libsml]$ git diff examples/
diff --git a/examples/sml_server.c b/examples/sml_server.c
index 1de47bc..83a775b 100644
--- a/examples/sml_server.c
+++ b/examples/sml_server.c
@@ -131,6 +131,12 @@ void transport_receiver(unsigned char *buffer, size_t buffer_len) {
                                        if (prec < 0)
                                                prec = 0;
                                        value = value * pow(10, scaler);
+                                       if (entry->status)
+                                               printf(
+                                                       "status: %x (type %x)\n",
+                                                       *(entry->status->data.status16),
+                                                       entry->status->type & SML_TYPE_FIELD
+                                               );
                                        printf("%d-%d:%d.%d.%d*%d#%.*f#",
                                                entry->obj_name->str[0], entry->obj_name->str[1],
                                                entry->obj_name->str[2], entry->obj_name->str[3],

and get:

[~/libsml]$ ./examples/sml_server -v - <../libsml-testing/EMH-ED300L_consumption.bin | grep -A1 status
status: 182 (type 60)
1-0:1.8.0*255#17243368.3#Wh
[~/libsml]$ ./examples/sml_server -v - <../libsml-testing/EMH-ED300L_delivery.bin  | grep -A1 status
status: 1b0 (type 60)
1-0:1.8.0*255#17243675.6#Wh
--
status: 1b0 (type 60)
1-0:1.8.0*255#17243675.6#Wh

consumption = 0x182 = 0b110000010
delivery: 0x1B0 = 0b110110000

(the values differ from those in arendst/Tasmota#8001 (comment) because it's a bit-field, each bit has a separate meaning.)

according to the docs
https://www.vde.com/resource/blob/951000/252eb3cdf1c7f6cdea10847be399da0d/fnn-lastenheft-edl-1-0-2010-01-13-data.pdf

bit 5 Energierichtung 0 <=> +A; 1 <=> -A.

so indeed in the delivery case, bit 5 is 1!

the code in tasmota is using & 0x20 = 0b100000, also bit 5:
arendst/Tasmota@61b9735#diff-97c9ae02f894436fc4188ce195998aa3d258fed1e78d21239c11c0c8539d01e5R1053

(it's a bit annoying that the status field is attached to the energy (Wh) reading, and not to the power (W) reading,
so one can not trivially apply that bit as a sign bit to the value of the (power) reading.)

[r00t-]

@SirkoVZ:
are you using vzlogger or some other utility using libsml?

i'm not sure if this should be implemented in libsml or in it's user.
it seems to me that at the sml level, libsml extracts the data just fine,
the meaning of the bits in the status field is not defined by sml itself,
but in that vde document ("Ausführungsvariante SML").
(just like the meaning of the obis numbers i guess)
so it doesn't seem that there's any need to change anything in libsml itself.

[r00t-]

@SirkoVZ:
volkszaehler/vzlogger#476 has an experimental implementation for vzlogger that works with the data in your dumps

[SirkoVZ]

I can confirm that it is working... more details later,

thanks so far :-)

[r00t-]

closing this, as it's out of the scope of libsml, see volkszaehler/vzlogger#431