Commit Graph

1927 Commits

Author SHA1 Message Date
Benjamin Larsson
22f9b4ebf8 Merge pull request #274 from mnhauke/master
Migrate fineoffset to new data format and add support for TFA 30.3157…
2016-01-23 23:58:25 +01:00
Martin Hauke
320b921391 Migrate fineoffset to new data format and add support for TFA 30.3157 Temperature sensor 2016-01-23 22:16:19 +01:00
Benjamin Larsson
44990b174f Merge pull request #273 from vestom/upstream
Pulse analyzer: Print abolute timing and frequency estimates
2016-01-23 19:32:42 +01:00
Tommy Vestermark
56ebc1cb4d Pulse analyzer: Print OOK carrier frequency 2016-01-23 18:28:32 +01:00
Tommy Vestermark
c7e81ca716 Pulse analyzer: Print FSK frequencies 2016-01-23 18:07:20 +01:00
Tommy Vestermark
ff56d009a5 Pulse analyzer: Print absolute timing 2016-01-23 18:06:33 +01:00
Benjamin Larsson
22330d87f0 Merge pull request #272 from vestom/upstream
Pulse detect: Suppress spurious short gaps
2016-01-23 16:29:03 +01:00
Tommy Vestermark
7e82322ff8 Pulse detect: Minor renaming for consistency 2016-01-23 14:48:32 +01:00
Tommy Vestermark
4c4bfde7e7 Pulse detect: supress spurious short gaps
Requires gaps to be at least PD_MIN_PULSE_SAMPLES before declaring a
true gap. Otherwise rewind and continue with pulse.
Decodes a couple of more files in test corpus:
rtl_433_tests/tests/oil_watchman/05/gfile239.data
rtl_433_tests/tests/waveman/01/gfile025.data (1 extra message)
rtl_433_tests/tests/xc0348/01/gfile003.data
2016-01-23 14:17:03 +01:00
Benjamin Larsson
629798046c Merge pull request #271 from dwmw2/master
Change bit limits in r_device structure to float too
2016-01-23 12:02:44 +01:00
Tommy Vestermark
9a90b1ecb3 Pulse detect: Make low level (noise) estimator even slower
Fixes situations, where noise estimate rise too fast for signal to
exceed threshold.
Tested with test corpus.
2016-01-23 00:16:20 +01:00
David Woodhouse
a7b60db353 Change bit limits in r_device structure to float too
Oops, missed this. In order to decode tests/emontx/01/gfile001.data we do
*really* need emonTx to be able to give the 20.3µs bit period precisely.
2016-01-22 15:37:55 +00:00
Benjamin Larsson
a6f342f3b2 Merge pull request #270 from dwmw2/emontx
Emontx
2016-01-22 15:53:14 +01:00
David Woodhouse
20dc24afb5 Finish implementing emontx driver 2016-01-22 13:14:53 +00:00
David Woodhouse
0df00e0fa4 Add crc16() function 2016-01-22 13:14:09 +00:00
David Woodhouse
4be45997fc Add bitbuffer_extract_bytes() function for accessing unaligned data 2016-01-22 13:13:57 +00:00
Benjamin Larsson
412fe2e6a5 Merge pull request #268 from vestom/experimental
FSK demodulator rewritten
2016-01-22 13:54:45 +01:00
Benjamin Larsson
712bed7d65 Merge pull request #269 from dwmw2/µs_limits
Switch decoder bit timing limits to floating point values in units of µs
2016-01-22 12:45:58 +01:00
David Woodhouse
7ead0882ed current_cost: Fix precision of bit limits
Now I'm just *guessing* that this should have been 250µs instead of 248.
2016-01-22 11:00:35 +00:00
David Woodhouse
dc9c8e841a dsc: Fix precision of bit limits
It says it should be 250µs. Now it's possible to have that, so let's try...
2016-01-22 10:57:07 +00:00
David Woodhouse
32aaece877 valeo: Fix precision of bit limits
It said "probably 26.5", so let's see if it was right, now that it's actually
possible to ask for that.
2016-01-22 10:56:01 +00:00
David Woodhouse
a0388750e1 ec3k: Fix precision of bit limits
It says it should be 50µs even though it could only ask for 48µs or 52µs
before. Take it at its word...
2016-01-22 10:54:45 +00:00
David Woodhouse
9afca9da90 Switch protocol decoder limits to µs instead of "samples at 250000Hz"
Strictly no intentional functional change here. So this patch includes
things like (in ec3k.c):
-       .short_limit    = 12,   // NRZ decoding (it apparently 12.5 which is a problem!!)
+       .short_limit    = 48,   // NRZ decoding (should be 50µs)

These *will* be changed to what the driver presumably intended, but in
separate bisectable commits.
2016-01-22 10:54:10 +00:00
David Woodhouse
59667852f7 rubicson: Fix precision of bit limits
Like the elv decoder, this also ends up with a fractional limit, so change
it now in a separate, bisectable commit.
2016-01-22 10:45:19 +00:00
David Woodhouse
24fc387e65 elv: Fix precision of bit limits
If we're going to describe the timings as 'XXX/4' then now it's a float
we might as well use 'XXX/4.0'. This is a functional change (not that
it causes any difference in results on the packet corpus), so goes in
a separate commit.
2016-01-22 10:37:03 +00:00
David Woodhouse
a06fb730f9 Convert decoder limits to float instead of int, to allow more precision
The (about to be submitted) emonTx decoder really needs to be allowed to
specify a limit of 20.3µs, and neither '50 samples' (20µs) nor
'51 samples' (24µs) will suffice.
2016-01-22 10:13:23 +00:00
Tommy Vestermark
b0f6fa1101 FSK: Fix slow estimation at initial frequency 2016-01-21 01:32:02 +01:00
Tommy Vestermark
9c90920669 FSK: correct comment 2016-01-21 00:23:00 +01:00
Tommy Vestermark
f00e0d806e current_cost: Invert bits to compensate for absolute FSK
Will now decode cc_2542watt.data :-)
2016-01-21 00:17:25 +01:00
Tommy Vestermark
b692da8a35 bitbuffer: Add function to invert bits 2016-01-21 00:03:43 +01:00
Tommy Vestermark
e675131223 FSK: Rework to make pulse polarity absolute (pulse = high frequency)
Necessary to avoid changing polarity for some sensors.
Side-effects to be worked around:
- Some pulse trains may start with a 0 width pulse
- Some device drivers may rely on other polarity
2016-01-21 00:03:15 +01:00
Tommy Vestermark
d6521b64bc Emontx: Correct for sample frequency of 2.5 MHz 2016-01-20 22:34:31 +01:00
Benjamin Larsson
889f3adb50 Merge pull request #263 from zuckschwerdt/feat-tempconv
new command line option -C native | f-to-c | c-to-f
2016-01-19 20:17:56 +01:00
Christian W. Zuckschwerdt
914c865eb2 new command line option -C native|si|customary that hooks into data_acquired_handler and
if `si` and type is DATA_DOUBLE and key is temperature_C then celsius2fahrenheit,
if `customary` and type is DATA_DOUBLE and key is temperature_F then fahrenheit2celsius.
Also if there is a FORMAT the last character "C" / "F" will be switched accordingly.
2016-01-19 08:46:34 +01:00
Tommy Vestermark
b73031a6c9 Merge branch 'master' into experimental 2016-01-18 23:56:59 +01:00
Tommy Vestermark
9e83e66d2b FSK: Fix for handling pulse overflow correctly 2016-01-18 23:53:32 +01:00
Tommy Vestermark
924b668300 FSK: Fix data is not cleared at End Of Package 2016-01-18 23:35:07 +01:00
Benjamin Larsson
c781612a38 Merge pull request #267 from viraptor/digitech_csv
[xc0348] Add csv fields
2016-01-18 09:23:13 +01:00
Stanisław Pitucha
f7b3c01021 [xc0348] Add csv fields 2016-01-18 14:39:32 +11:00
Benjamin Larsson
aa627068ca Merge pull request #266 from viraptor/xc0348_fixes
Xc0348 fixes
2016-01-18 02:39:53 +01:00
Stanisław Pitucha
da4ff03ab0 [xc0348] Add json output 2016-01-18 11:50:04 +11:00
Stanisław Pitucha
414ed77ac9 Include stdio for FILE 2016-01-18 11:25:44 +11:00
Stanisław Pitucha
498f4cb687 [xc0348] Fix temperature (top nibble doesn't belong) 2016-01-18 11:25:40 +11:00
Stanisław Pitucha
1c10402968 [xc0348] Don't filter on device id 2016-01-18 11:25:36 +11:00
Stanisław Pitucha
5edc354030 [xc0348] Add CRC validation 2016-01-18 11:25:28 +11:00
Benjamin Larsson
d7058da208 fix limits for decoder 2016-01-18 00:10:09 +01:00
Benjamin Larsson
0c293af6d8 Merge pull request #265 from vestom/unsigned
Change pulse data from unsigned to int to avoid undefined behaviour
2016-01-17 21:59:28 +01:00
Tommy Vestermark
945b99df70 Change pulse data from unsigned to int to avoid undefined behaviour
Thanks to zuckschwerdt for noticing this!
2016-01-17 21:41:25 +01:00
Tommy Vestermark
edd7fbac28 FSK: Increase initial frequency threshold
Minimum delta seen until now is ~7000. Threshold will be 3000. Makes
initial estimate more robust.
2016-01-17 20:29:22 +01:00
Tommy Vestermark
93571a535a emonTx: Add stub driver
Proof of concept...
2016-01-17 19:49:37 +01:00