Arduino softwareserial clashing with SPI? -

i have arduino project using nrf24l01+ radio module on spi (using library: http://tmrh20.github.io/rf24/) , rfid reader on softwareserial. sleeping arduino , having them wake via interrupt when message has been received or rfid tag ready read. rfid on pins 4 , 5, while nrf covers pins 9 - 13 number 2 interrupt.

both of these modules work fine sleep , interrupt code separately, when combined in single sketch, arduino wake due rfid tag, read it, try send on radio , hang, waiting library call write() return.

i have delved bit 2 libraries, can't make heads or tails of softwareserial library. seems maybe using same isr behind scenes nrf module, don't see why should big problem, , don't understand why should cause radio hang.

i know may long shot, have idea might going on? maybe knows these libraries? thoughts on work around? thanks.

i experiencing same symptoms, , problem turned out buffer overrun in code. overrun due softwareserial dropping bytes because rf24 library interring interrupt handling.

the code in question reads gps receiver using softwareserial, parses nmea sentences , extracts latitude , longitude information sending on radio using rf24. goes this:

if (gpsserial.available()) {   int c = gpsserial.read();   if (c == '\r') {     buf[bdx] = '\0';     // process buf here, contains null terminated nmea sentence     // , send via rf24     bdx = 0;   else if (c != '\n' && bdx < buflen) buf[bdx++] = c; } 

where buflen sized larger single nmea sentence.

the experienced reader pick on problematic line:

buf[bdx] = '\0';

which writes buf without range check. in normal operation, uninterrupted stream of characters gps module, code works fine, because encounter \r before running out of space in buf. sending information on rf24 causes enough delays characters dropped or corrupted softwareserial, , assumption no longer holds.

so happens this:

1. a \r missed, , previous nmea sentence not discarded
2. the next nmea sentence read buf
3. bdx advances until stopped range check in else
4. \r of next nmea sentence countered
5. buf[bdx] = '\0'; writes past end of buf

at point arduino stops responding, , looks write blocking.

the fix line added before if:

if (bdx >= buflen) bdx = 0; 

with no other modification line, code has been running on 5 hours without issue, whereas not last more 30 seconds before write "blocks".