Can't get mqttany to read the gpio #157
Replies: 2 comments 7 replies
-
|
It should be reporting the number of pulses per minute, as you are trying to do. Are you sure that the pin 24 you're connected to is actually the WiringPi pin 24 and not GPIO24 (the default pin mode This configuration will actually report the kWh usage in the last minute: gpio:
pulsecounter:
pin: 24
counter:
interval: 60
interrupt: 'rising'
function: 'count'
divider: 0.06 # 1000 pulses per kWh / 60 minutes = 16.666 pulses per minute -> 1/16.666 = 0.06In the configuration you posted the |
Beta Was this translation helpful? Give feedback.
-
|
WiringPi pin 24 should be the correct as using "gpio read 24" I can seen the state change using bash script and according to "man gpio" WiringPi is the default numbering scheme it uses. Also when setting polling interval to 1 I get following results and MQTT broker shows accordingly So at least something is happening! To me it seems like there is something wrong with interrupts, as clearly MQTTany can see the state changes but somehow the interrupts are not working. MQTTany version is 0.14.3 |
Beta Was this translation helpful? Give feedback.
-
|
So it is actually working there. The error is from the new GPIO interface system in Linux called cdev (character device), so MQTTany has fallen back to the old sysfs system which did not error. It's possible that the account you're running MQTTany as does not have access to the cdev device, it would output some logs right at the beginning of startup to indicate if it doesn't have access, and with debug logged activated it will also indicate what it does have access to. I am back to thinking you shouldn't be using WiringPi pin numbering. I think the |
Beta Was this translation helpful? Give feedback.
-
|
LDR is connected to physical pin 35, which is SoC 19 and WiringPi pin 24. However, your speculation about the kernel module made me think that the problem might be related to kernel. It seems like my kernel version 5.10.103-v7l+ was the problem. After updating the kernel with rpi-update to 6.1.46-v8+, MQTTany started to work just fine with with my config: This config gives the impulses per minute as it should. However, the divider parameter is causing problems. Using divider smaller than 1, for example 0.06, makes the output always be 0.0:
Is this some rounding issue? |
Beta Was this translation helpful? Give feedback.
-
|
It seems to be the problem in counter.py, divider is defined to be int.
Changing this to float solved the problem.
|
Beta Was this translation helpful? Give feedback.
-
|
Ok so that was 2 separate issues. The divider being coerced to an int made it 0, even with cdev throwing an error the sysfs interface was giving pulses correctly but they were being multiplied by 0 and always returning 0. That was my mistake, I checked that the output of the multiplication calculation wasn't being cast to an integer but I didn't think to check the divider option. I'm going to open an issue and change that, I think this kind of use case to allow converting counts to an measurement directly is handy. Glad you got it sorted out in the end! |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for your help! |
Beta Was this translation helpful? Give feedback.
-
I have a LDR based impulse counter connected to Raspberry Pi 4 GPIO pin 24, which I intent to find out real-time power consumption of my house. The utility meter has a led which gives 1000 impulses per used kWh, so based on counting those (per minute, 10 minute or so) would give quite good estimate on real-time power consumption.
I have verified that the LDR setup works as I can seed the impulses using simple infinite bash loop, output being:
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
...
However, I haven't been able to figure out how should I configure the MQTTany to get these impulses to it. No problem with MQTT connection, that works fine, but I just can't get the impulses read.
Here's my config:
What I am missing here?
Beta Was this translation helpful? Give feedback.
All reactions