rf12_canSend() seems not to work

(16 posts) (4 voices)

Tags:

No tags yet.

jan

Hello everybody,
maybee someone can give me a hint on whats wrong here:

void loop() {
readdata();
rf12recvDone();
if (rf12canSend()){ // wait until sending is allowed

    rf12sendStart(0, &tempC, sizeof tempC );

    Serial.print("temp: ");

    Serial.println(tempC);

  }

}


what i get is only one printing of the tempC on the Serial console. then nothing more. when i reset the arduino, i always get

one reading.. seems strange, any ideas?
seems as if rf12canSend() is always false..
thx a lot in advance,

Posted 13/09/2009 17:57:34

jcw

I think the problem is timing - you need to add some delay logic, i.e. send new packets at least a few milliseconds apart. See the timer in "pollSensors()" and the logic w.r.t. "txrequest" in the rooms sketch:
http://code.google.com/p/jeelabs/source/browse/trunk/rooms/rooms.pde

IOW, set a "want to send flag" once a second or so. Send when possible and also clear that flag. Rinse and repeat.

Posted 14/09/2009 09:19:43
jan

thanks for the hint to the demo code!
the thing is i already tried that, there is a delay(1000); in the readdata()
function. should be the same as a flag or what do you think?
since i am at work right now i cant try this code.

Posted 14/09/2009 11:50:36
jcw

The delay() is not enough, there is a subtle but important difference with timers: you need to keep calling rf12recvDone() regularly - maybe use a loop of 250x calling "delay(4); rf12recvDone();" instead?

Posted 14/09/2009 12:02:05
jan
hello,
i tried your sugestion, but i am afraid it's not working.. here is my complete code:

#include <RF12.h>
float tempC;
void setup() {
Serial.begin(9600);
rf12initialize(0, RF12433MHZ);
}
void readdata(){
tempC = analogRead(0); //read the value from the sensor
tempC = (5.0 tempC 100.0)/1024.0; //convert the analog data to temperature
}
void loop() {
readdata();
for(int i=0; i<250; i++){
```
    rf12recvDone();

    delay(4);

  }

  if(rf12canSend()) {	// wait until sending is allowed

    rf12sendStart(0, &tempC, sizeof tempC);

    Serial.print("temp: ");

    Serial.println(tempC);

  }

}


everytime i restart the µc i get one reading, then nothing.
don't know what to try next. Anyone any advice?
```
thx!
Posted 14/09/2009 18:14:53
jcw

Don't initialize the RF12 driver to node id 0 - that's used for pseudo-OOK transmit mode.

Posted 14/09/2009 19:33:06
jan

Hello Jean-Claude,

thank you very very much for reading my code!!!! this was the error. i changed initialization to rf12config()
and know it prints out temp on the console 1/sec! Great!
Only that my second arduino died so i cannot check reception... same thing happened yesterday but now no more reserve chips.. seems as if the rf12 module killed the avr (or its bootloader, cant check because my isp programmer is out of hause, argh), keep getting out of sync errors when trying to upload a sketch.
hered of anything like this before?

Posted 14/09/2009 21:09:03
skobalczyk

Actually I have similar issue with the JeeNode now. But unpluging the FTDI cable from PC and connecting it back seems to help.

Typically it shows this:
avrdude: stk500getsync(): not in sync: resp=0x00 avrdude: stk500disable(): protocol error, expect=0x14, resp=0x00

Posted 14/09/2009 22:22:46
jan

today i got my isp programmer back. i veryfied the bootloader of the atmega and yes, it was corrupted! on two chips! i then burned fresh bootloaders on the avr's and could upload any sketches on them. after some testing with the rf12 modules i got this error again... same procedure brought the atmega back to life.
maybee the problem is linked to the default LED on pin 13 of the original arduino design. i also noticed that the reset blink code of the LED is different with a connected rf12 module.
any ideas?

Posted 15/09/2009 19:34:30
seaton

I don't know if the LED pin would be causing it as I think the LED is going to one of the pins on the RFM12, but you could try these bootloaders as I've changed LED pins to suit my Widgetboard (PB0 and PB1) and should be compatible with the JeeNode. (these are a modified lilypad bootloader)

http://code.google.com/p/strobit/source/browse/#svn/trunk/arduino/hardware/bootloaders/rfm12widget

Currently compiled for only 8 and 10MHZ, if you need any other speeds let me know and I can re-compile and add to SVN.

If boot loader is becoming corrupt then something may be overwriting a portion of it due to a bug in the main application code somewhere? invalid pointer? array out of bounds?

From memory there is a boot-loader protection fuse setting in the AtMega168/328 so you may want to investigate this when next programming the boot-loader. I don't know but it might protect the boot-loader from corruption, but if a code error is there then your application will probably lock-up but at least your boot-loader will be intact (hopefully)

Regards,

Stephen...

Posted 16/09/2009 04:48:12
jan

what exactly do you mean when you say you changed led pins? i would need the rf12 module not to connect to d13 of the arduino (PB5 of atmega). this would make the rf12 librairy fully arduino compliant, as for now the led is useless and/or causing trouble.

i checked the fuses when i programmed the atmel, everything was fine, bootloader protection fuse was still set correctly.

unfortunately i destroyed one of my two rf modules while testing and have to wait for my new ones to arrive to continue testing..

regards,
Jan

Posted 16/09/2009 10:20:34
jcw

Are you using RFM12B's? These only work up to 3.8V. Only the older RFM12's can be used up to 5V.

I don't know what is causing these problems - you can't easily change the D13 pin to something else because it is the SPI clock pin, which is fixed as PB5 on the ATmega. But then again, I don't think having a led on there can cause any trouble since the SPI clock is simply a fast-switching output signal.

Posted 16/09/2009 12:24:37
seaton

@Jan
I've changed the bootloader code so the leds are mapped to the same pins as my onboard LEDS, so are not using PB5(pin D13), that is used as the SCK pin on the SPI. I don't think this would be causing your problem though.

I agree with JCW. Do you have 3.3V on the RFM12b? If the RFM12B are dying then I think it would be something along those lines, to high voltage, mind you I've run some RFM12B at 5v in the past for some time (days) and not killed them (yet)..

Posted 16/09/2009 14:31:50
jan

the thing with the led pin is that the default arduino config has a led on PB5 and it would be neat to let it blink every once in a while when a packet is send of or something like that. this is not possible at the moment.
please could you post how/where to change this pin?
and for the supply voltage this is not the problem. i mounted the modules on tiny pieces of breadboard with pins, so i could easily work with them. while putting it back an the board some of the pins broke through and scratched the copper lead of the pcb off... they are now burried under the black plastic thing...
good example of murphys law...

Posted 16/09/2009 18:19:55
jcw

Let me rephrase...

You can't change the use of PB5 for the SPI clock needed for the RFM12B, it's hardwired in the ATmega. So w.r.t. SPI interfacing, the Arduino team made a somewhat unfortunate design choice :(

You'll have to connect another LED to another pin for the kind of signaling you're after.

Posted 16/09/2009 23:47:59
seaton

@Jan,

If you are using a standard arduino then the only way to change the LED pin to something else will be to cut and solder a jumper to another I/O port that is not conflicting. I can't remember if they have a jumper for this LED to enable/disable it, if so then use this instead of cutting the track. I'm using PB0 and PB1 for two onboard leds.

I don't know why they used PB5 for the onboard LED as this is the SCK for the hardware SPI.

Check the source of the bootloader and you can easily see where the LED is defined. I have it flashing 3 times when the bootloader is entered, once the bootloader has finished I use the leds for whatever.

Posted 16/09/2009 23:52:10