NodeManager: plugin for a rapid development of battery-powered sensors

[Robbie_]

Hi User2684,

As promised i took NodeManager for a spin. It functions well but i have an issue with powering a DHT22 from the digital pins. It just does not show any temp/hum. If i use the power from the battery it functions well.

Any clue what this could be?

I used setPowerPins6,7,10000 to be sure the DHT22 has enough time to start up. If i use my multimeter between pins 6,7 i see the voltage going to 3.something volts and to 0 when the 10000 miliseconds are gone.

My config is as following;

// before
void before() {
  // setup the serial port baud rate
  Serial.begin(MY_BAUD_RATE);  
  /*
   * Register below your sensors
  */
  nodeManager.setSleep(SLEEP,1,MINUTES); 
nodeManager.registerSensor(SENSOR_DHT22,3);
nodeManager.setPowerPins(6,7,10000);
  /*
   * Register above your sensors
  */
  nodeManager.before();
}

[bilbolodz]

@Robbie_ I've NOT checked anything but I think that probably better:

1. power up DHT
2. wait a little to "dht startup"
3. register sensor

[Robbie_]

@bilbolodz
So like this?

nodeManager.setPowerPins(6,7,10000);
nodeManager.registerSensor(SENSOR_DHT22,3);
nodeManager.setSleep(SLEEP,1,MINUTES);

[bilbolodz]

@Robbie_ said in NodeManager: plugin for a rapid development of battery-powered sensors:
Try this:

nodeManager.setPowerPins(6,7,10000);
wait(10000);
nodeManager.registerSensor(SENSOR_DHT22,3);```

[Robbie_]

@bilbolodz said in NodeManager: plugin for a rapid development of battery-powered sensors:

wait(10000);

Thanks for the fast reply! Still the same issue;

SEND D=0 I=200 C=0 T=48 S=STARTED I=0 F=0.00
14258 TSF:MSG:SEND,100-100-0-0,s=200,c=1,t=48,pt=0,l=7,sg=0,ft=0,st=OK:STARTED
14395 MCO:BGN:INIT OK,TSP=1
ON P=7
14426 MCO:SLP:MS=10000,SMS=0,I1=255,M1=255,I2=255,M2=255
14493 MCO:SLP:TPD
14512 MCO:SLP:WUP=-1
DHT I=1 T=nan
OFF P=7
SLEEP 60s
SEND D=0 I=200 C=1 T=48 S=SLEEPING I=0 F=0.00

[user2684]

@bilbolodz multiple DS18B20 are supported indeed, when you register a DS18B20 sensor and provide the pin where multiple devices are attached, NodeManager will go through all of them (by calling getDeviceCount() of DallasTemperature) and will register each of them with a different child id. Is this what you were looking for?

[user2684]

Hi, @Robbie_ the order doesn't matter since all of this is happening inside before(), NodeManager just keeps track of all your settings without doing anything at that stage.

Your configuration looks correct to me, the DHT sensor attached to pin 3, pin 6 playing the role of ground, pin 7 the role of vcc and waiting for 10 seconds after powering on, before actually reading the value from the sensor. The log also tells the same story, pin 7 is turned on, then 10s of sleep and then the "nan" read which I cannot really understand (the code is very simple, just executes a digitalWrite(_vcc_pin, HIGH) at that stage).
Would you please ensure pin 7 is HIGH during the 10s sleep?

Not sure if it matters (and those more expert than me with a DHT sensor can confirm) but I've noticed that I call dht.begin() only once (during before()) and during loop() I just call readTemperature(). Does powering off the sensor have an impact on this? e.g. should I call begin() before every read?
Thanks

[gohan]

I think you need to call the begin if you power off the sensor

[bilbolodz]

@user2684 said in NodeManager: plugin for a rapid development of battery-powered sensors:

@bilbolodz multiple DS18B20 are supported indeed, when you register a DS18B20 sensor and provide the pin where multiple devices are attached, NodeManager will go through all of them (by calling getDeviceCount() of DallasTemperature) and will register each of them with a different child id. Is this what you were looking for?

Multiple sensors: great but for me important thing is "fixed" matching exact DS18B20 to particular child id. I'm going to monitor may two function water boiler and check input/output temperature. In case when one of the DS disappear (because of many reasons) after next nodemanager device startup assigning sensors to child id will change and it will be a great problem. That's why I'm asking for exact assigning DS id - child id. Is it clear now?

[user2684]

Thanks @gohan. @Robbie_ if you can then try copying that "_dht->begin();" you have in line 576 of NodeManager.cpp also at the beginning of "SensorDHT::onLoop()" which is just below we can see if this can help (unfortunately I don't have anymore a DHT sensor to test by myself).

Another try would be to disable both MySensors and NodeManager debug: I've seen strange behaviors happening when approaching the dynamic memory limit (including this) and since objects are allocated dynamically, it could be the case (this issue is partially mitigated in the dev branch)

[user2684]

@bilbolodz I understand what you are saying. I've opened this: https://github.com/mysensors/NodeManager/issues/42. I did not consider this use case before because according to DallasTemperature documentation the attached sensors are always returned in the same order but you're right, if one of the sensor for any reason is not detected, it will mess up your logic. Thanks for reporting it.

[bilbolodz]

@user2684 Great, so I'm waiting for version 1.4. Any ideas about timeline?

[vikasjee]

@user2684 This is exactly i had mentioned in my requirement for IOExpanded pins attached to any sensor or virtual buttons... getting the same child_id everytime. Hope its clear now...

[Robbie_]

@user2684 Thanks for your reply. Ive tested the following;

i editted the NodeManager.cpp and added _dht->begin(); to the beginning of "SensorDHT::onLoop()"

I also disabled both debug, still the same issue.

In your other post you mentioned "Would you please ensure pin 7 is HIGH during the 10s sleep?"

My log looks as following;

PWR G=6 V=7
REG I=1 P=3 P=6 T=0
REG I=2 P=3 P=7 T=1
NodeManager v1.30
INT1 M=255
INT2 M=255
RADIO OK
PRES I=200, T=23
PRES I=201, T=30
BATT V=3.26 P=94
SEND D=0 I=201 C=0 T=38 S= I=0 F=3.26
PRES I=1 T=6
PRES I=2 T=7
READY

MY I=100 M=1
SEND D=0 I=200 C=0 T=48 S=STARTED I=0 F=0.00
ON P=7
DHT I=1 T=nan
OFF P=7
SLEEP 60s
SEND D=0 I=200 C=1 T=48 S=SLEEPING I=0 F=0.00

AWAKE
SEND D=0 I=200 C=1 T=48 S=AWAKE I=0 F=0.00
ON P=7
DHT I=1 T=nan
OFF P=7
SLEEP 60s
SEND D=0 I=200 C=1 T=48 S=SLEEPING I=0 F=0.00

Where do you want me to messure if the pin is on?

AWAKE
SEND D=0 I=200 C=1 T=48 S=AWAKE I=0 F=0.00
ON P=7
DHT I=1 T=nan
<------------------------ I messured here and the voltage is 3.something volts
OFF P=7
SLEEP 60s
<--------------------------- HERE?
SEND D=0 I=200 C=1 T=48 S=SLEEPING I=0 F=0.00

[gohan]

use a multimeter and check if pin 7 during the startup of the arduino and see if you get 5v out for 10 seconds

[user2684]

@bilbolodz v1.4 is actually ready, just want to see if the DHT issue discussed here requires a code fix or not before releasing but it is already available in the dev branch (https://github.com/mysensors/NodeManager/tree/8280de908929a117287a70e73cd047f50ab9a5bb). Let me explain the way I have implemented it because may look cumbersome but I didn't want to make exceptions and have a sensor configured in a different way the all the others.
First of all, for any sensor, the child id can be manually assigned by providing a third argument to registerSensor(), e.g. to assign child id 5:

nodeManager.registerSensor(SENSOR_THERMISTOR,A1,5);

This was available also in v1.3 but not very clear from the docs. Of course this approach does not work when you have multiple DS18B20 sensors attached since they are all configured with a single line.
So I have implemented a renameSensor() function which basically assign a different child id, e.g. to give the sensor with child id 1 the child id 5:

nodeManager.renameSensor(1,5);

This is available for all the sensors. So in the case of multiple DS18B20 you need to register them all in the usual way ( registerSensor(SENSOR_DS18B20,pin)) then assign the child id you like just after so in case one is lost, you still have a static mapping. The alternative would have been to provide a special registerSensor only for DS18B20 which I don't like that much. Any other better alternative of course is more than welcome!

Btw, I have also added for DS18B20 getDeviceAddress() - in case you want to assign the child id based on the device address, getResolution() and setResolution()

[user2684]

@Robbie_ even if unfortunately I do not have a spare DHT sensor to test right now, I did some tests with other sensors trying to understand the issue. My educated guess it that the problem is related to the fact that setPowerPins() set both ground and vcc but initialize both to LOW. This means that when you register the sensor (and the DHT library is initialized), the sensor is off. This behavior will be changed in v1.4 (vcc wil be set to high). What you can do with the version you are currently running is to call nodeManager.powerOn() just after the call to setPowerPins() and before registering the sensor so to keep the power on.

If this does not work, try calling nodeManager.setAutoPowerPins(false) after setPowerPins(): this prevent nodeManager to turn the power off which is not what you want but can help in isolating the issue.
Thanks!

[luizrrocha]

Hello, good work !!! I'll be trying your extension very soon....

Quick question: is it compatible with ATMEGA chips flashed with MYSBootloader pre-1.3a ?

[user2684]

@luizrrocha I did not test it by myself but the bootloader used should have no impact on the code (unless you need to call different/specific functions with that bootloader but I guess this is not the case).

[Robbie_]

@user2684 Ive tested this morning a DHT22 without resistor between VIN & DATA, now it works correct. I reverted the changes back and just use the default code (1.3), everything keeps functioning. I don't understand why it works but hey, i got data! :-)

Somehow the following issue pops up with domoticz (also before the DHT22 was working), im not sure if anybody can help me here with it, else il open a threat at the domoticz forum. As soon ase NodeManager presents itself (via ESP8266 WiFi Gateway // MyS v2.1.0) to domoticz, domoticz goes offline. Does somebody else had the same experiance?