Porting MySensors to work with the RadioHead library

[kolaf]

@hek I tried setting up a second sensor (temperature and power supply voltage) with the updated library. I wiped eeprom, uploaded the sketch, and fired it up.

It requested a node ID from the gateway, which it received, and everything seemed to go smoothly until it had transmitted the first data point. After that it immediately enters a request for a new node ID, which it receives, and things go in an endless loop. Here is the log:

find parent
send: 255-255-255-255 s=255,c=3,t=7,pt=0,l=0,sg=0,st=bc:
read: 0-0-255 s=255,c=3,t=8,pt=1,l=1,sg=0:0
parent=0, d=1
req id
send: 255-255-0-0 s=255,c=3,t=3,pt=0,l=0,sg=0,st=ok:
read: 0-0-255 s=255,c=3,t=4,pt=0,l=1,sg=0:1
send: 1-1-0-0 s=255,c=0,t=17,pt=0,l=5,sg=0,st=ok:1.4.1
send: 1-1-0-0 s=255,c=3,t=6,pt=1,l=1,sg=0,st=ok:0
id=1
send: 1-1-0-0 s=255,c=0,t=17,pt=0,l=5,sg=0,st=ok:1.4.1
send: 1-1-0-0 s=255,c=3,t=6,pt=1,l=1,sg=0,st=ok:0
sensor started, id=1, parent=0, distance=1
send: 1-1-0-0 s=255,c=3,t=11,pt=0,l=20,sg=0,st=ok:Temperature basement
send: 1-1-0-0 s=255,c=3,t=12,pt=0,l=3,sg=0,st=ok:1.0
send: 1-1-0-0 s=1,c=0,t=6,pt=0,l=5,sg=0,st=ok:1.4.1
send: 1-1-0-0 s=2,c=0,t=13,pt=0,l=5,sg=0,st=ok:1.4.1
1023send: 1-1-0-0 s=2,c=1,t=38,pt=7,l=5,sg=0,st=ok:5.4
send: 1-1-0-0 s=255,c=3,t=0,pt=1,l=1,sg=0,st=ok:102
send: 1-1-0-0 s=1,c=1,t=0,pt=7,l=5,sg=0,st=ok:20.6
req id
send: 1-1-0-0 s=255,c=3,t=3,pt=0,l=0,sg=0,st=ok:
read: 0-0-1 s=255,c=3,t=4,pt=0,l=1,sg=0:3
send: 1-1-0-0 s=255,c=0,t=17,pt=0,l=5,sg=0,st=ok:1.4.1
send: 1-1-0-0 s=255,c=3,t=6,pt=1,l=1,sg=0,st=ok:0
sensor started, id=1, parent=0, distance=1
send: 1-1-0-0 s=255,c=3,t=11,pt=0,l=20,sg=0,st=ok:Temperature basement
send: 1-1-0-0 s=255,c=3,t=12,pt=0,l=3,sg=0,st=ok:1.0
send: 1-1-0-0 s=1,c=0,t=6,pt=0,l=5,sg=0,st=ok:1.4.1
send: 1-1-0-0 s=2,c=0,t=13,pt=0,l=5,sg=0,st=ok:1.4.1
955send: 1-1-0-0 s=2,c=1,t=38,pt=7,l=5,sg=0,st=ok:5.0
send: 1-1-0-0 s=255,c=3,t=0,pt=1,l=1,sg=0,st=ok:95
send: 1-1-0-0 s=1,c=1,t=0,pt=7,l=5,sg=0,st=ok:20.7
req id
send: 1-1-0-0 s=255,c=3,t=3,pt=0,l=0,sg=0,st=ok:
read: 0-0-1 s=255,c=3,t=4,pt=0,l=1,sg=0:3
send: 1-1-0-0 s=255,c=0,t=17,pt=0,l=5,sg=0,st=ok:1.4.1
send: 1-1-0-0 s=255,c=3,t=6,pt=1,l=1,sg=0,st=ok:0
sensor started, id=1, parent=0, distance=1
send: 1-1-0-0 s=255,c=3,t=11,pt=0,l=20,sg=0,st=ok:Temperature basement
send: 1-1-0-0 s=255,c=3,t=12,pt=0,l=3,sg=0,st=ok:1.0
send: 1-1-0-0 s=1,c=0,t=6,pt=0,l=5,sg=0,st=ok:1.4.1
send: 1-1-0-0 s=2,c=0,t=13,pt=0,l=5,sg=0,st=ok:1.4.1
956send: 1-1-0-0 s=2,c=1,t=38,pt=7,l=5,sg=0,st=ok:5.0
send: 1-1-0-0 s=255,c=3,t=0,pt=1,l=1,sg=0,st=ok:95
send: 1-1-0-0 s=1,c=1,t=0,pt=7,l=5,sg=0,st=ok:20.7

As far as I can see this should be caused by the node ID still being AUTO, but this shouldn't be the case since it appears to be correctly stored in ROM?

[kolaf]

Please disregard my previous "analysis"/guess. It turns out that the repeating behaviour that was caused by a program crash. The trigger for this crash is MyHwATMega328::sleep(ms) . I'm running a RFM69HW radio on a anarduino board, which should be using the ATMega 328 chip, same as the moteino.

By replacing the call to this function in MySensor::sleep(unsigned long ms) with a simple delay(ms), the crashes no longer occur.

[kolaf]

My current guess is that the watchdog timer resets the chip after the time period instead of throwing an interrupt to wake the system...?

[kolaf]

I narrowed it down even further. It appears that there is no appropriate interrupt and present to handle when the watchdog fires. I have no idea why this is the case, and I do not know if this fix is correct, but by a placing this code somewhere in MySensors.cpp everything works nicely:

ISR( WDT_vect ) {
    /* dummy */
}

[hek]

@kolaf

Ok, strange. Would be nice to know why this fixes the problem,

It should probably be placed here if it's atmega328 related (and only need to be called once):
https://github.com/mysensors/Arduino/blob/development/libraries/MySensors/MyHwATMega328.h#L22

[kolaf]

@hek
As far as I know the function is the interrupt routine that is called by the watchdog when the time expires. I guess it doesn't matter where it is defined so it should be fine to move it. It is not called explicitly anywhere in the code, but by the watchdog. As such, I don't think it should be defined as is done in the header file for the other functions.

My question is why is it not working for me without this code, I assume it has been working for everyone else?

[hek]

Ah.. doh..

Moved parts of the low-power library into MyHwATMega328 and forgot the ISR-thingie.

https://github.com/mysensors/Arduino/blob/master/libraries/MySensors/utility/LowPower.cpp#L850

Good catch.

[kolaf]

Glad I could help :-). I guess my prize will be that you patch this so that I do not have to create a pull request ;)

[kolaf]

I might as well continue while I'm at it :-)

Another weird issue I'm having is that every node requests a new ID whenever it starts, but it does not switch to the new ID it receives. It just continues with the old ID it already had.

When I wipe the ROM it requests a new ID as expected, and this time it is saved and used for subsequent communication. Why is it requesting a new ID when it boots, when it already has one?

[hek]

@kolaf

Should not happen. Will look at it later tonight.

[hek]

Ok, finally got the thumbs out and actually tested the RFM69 code using a couple of Moteinos. They seem to communicate just fine. :)

Tested with SerialGateway<->RelayActuator-Example.

[reddy11]

I can also report success and things look pretty good. Had to change the line:

//MyTransportRFM69 transport;
MyTransportRFM69 transport(RFM69_FREQUENCY, RFM69_NETWORKID, RF69_SPI_CS, RF69_IRQ_PIN, true, RF69_IRQ_NUM);

So that the rfm69HW could be supported.

Now, on to see if I can get the MQTT working.

[kolaf]

@reddy11 I had to do the same, although I did the change in the default parameters for the function definition since all my radios are HW :-).

I'm not using the MQTT gateway, I'm using the serial gateway together with a Perl script someone here made which functions as a MQTT gateway, and which I modified to handle node ID assignments.

[kolaf]

@hek You didn't see any extra node ID requests when the sensor booted?

[tekka]

@kolaf are you using the OTA bootloader?

[kolaf]

@tekka said:

@kolaf are you using the OTA bootloader?

No, I haven't figured out that part yet. I'm not even sure if I have available flash memory for this, nor if this is expected to work with the rf69 radios.

[tekka]

@kolaf ok, i was asking because there is an issue wIth the ID assignment and the OTA bl in the dev branch resulting in the scenario you described.

[A Former User]

Thanks everybody for the nice work! My initial tests also work out of the box with a serial gateway and humidity node. I will let it run over the weekend to see if there are any issues coming up.

[lafleur]

I have serial gateway running on a MoteinoMega and a node running on a Moteino with RFM69HW radios... Using the 2April2015 build of the development arm... with the addition of the :

//MyTransportRFM69 transport;
MyTransportRFM69 transport(RFM69_FREQUENCY, RFM69_NETWORKID, RF69_SPI_CS, RF69_IRQ_PIN, true, RF69_IRQ_NUM);

and the correct pin-out for the radios and LED for these boards...

I needed to mod the MsTimer2.cpp file to add 'AVR_ATmega1284P ' to each of the #ifdef line to support the Mega

As I do not yet have a controller running, I see the gateway output as below when the node TX's:

0;0;3;0;9;read: 255-255-0 s=255,c=3,t=3,pt=0,l=0,sg=0:255;255;3;0;3;

[lafleur]

I need a simple controller that can plot data that can be viewed on a web server... Mac/Pi or ??
want hr/day/week/mo/yr plots..

what do you recommend??