›

$RC2 synchronization with phone

stez90stez90
in RaceChrono for Android
Hello!
I'm currently testing a simple homemade datalogger that sends out $RC2 sentence like:

$RC2,,0,-0.02,0.02,0.98,,,1.11,0.96,84.46,,,,,*23
$RC2,,1,-0.02,0.01,0.98,,,0.75,1.04,84.43,,,,,*2D
$RC2,,2,-0.02,0.02,0.98,,,0.94,1.19,84.89,,,,,*28
$RC2,,3,-0.02,0.02,0.99,,,0.62,1.01,84.80,,,,,*21
$RC2,,4,-0.02,0.01,0.98,,,-0.03,1.03,85.04,,,,,*1
$RC2,,5,-0.02,0.01,0.98,,,-0.58,1.10,85.25,,,,,*F
$RC2,,6,-0.02,0.01,0.98,,,-1.02,1.17,85.46,,,,,*0
$RC2,,7,-0.02,0.01,0.98,,,-1.30,1.19,85.68,,,,,*2
$RC2,,8,-0.02,0.02,0.98,,,-1.57,1.25,85.89,,,,,*F
$RC2,,9,-0.02,0.01,0.98,,,-1.76,1.26,86.09,,,,,*6
$RC2,,10,-0.02,0.01,0.98,,,-1.84,1.24,86.31,,,,,*3A
$RC2,,11,-0.02,0.02,0.98,,,-1.70,1.05,86.63,,,,,*37
$RC2,,12,-0.02,0.02,0.98,,,-1.82,1.08,86.83,,,,,*3A
$RC2,,13,-0.02,0.02,0.98,,,-1.93,1.10,87.04,,,,,*3C
$RC2,,14,-0.01,0.03,0.98,,,-1.71,0.83,87.50,,,,,*3F

so with only the counter made with an incrementing variable. Data rate is about 3Hz. Gps is the internal one of the phone (tested with two android phones).

How are supposed to work the counter/timestamp fields? I'm experiencing big delays (in the order of 9 or 10 sec in one of the two phones, around 5 in the other one) between GPS data and $RC2 graph, even if the "live" data display is in sync...

Thanks,
Stefano

Comments

  • aol_of_RaceChronoaol_of_RaceChrono
    One that immediatedly catches my eye is the checksum. It needs to have always two digits.
  • stez90stez90
    Didn't notice that in previous tests, quite strange behaviour.. Looks like live display of RC shows those values even with single digit checksum (I'm cross checking looking at phone screen and serial monitor simultaneously)... Nevertheless, I'll try to fix that.

    Anyway, the real "problem" is that I see the live values with correct timing while I drive, but when I check the graphs they are not saved in sync with gps traces.

    For example: https://www.dropbox.com/s/p8gico43tgy0p88/2015-05-27 23.30.31.png?dl=0
    The purple line is the lean angle (not perfect, about -13°deg off) and it is about 9sec ahead of the gps trace (the S turn in the map that is the dip at 1213sec of speed plot).. But while I was driving data was correctly showed in real time..

    Any advice? Should I use timestamp (don't know how..)?

    Thanks
    Stefano
  • aol_of_RaceChronoaol_of_RaceChrono
    edited May 2015
    It should work without real timestamp. But when timestamps are not used the data rate you use should be constant. If it's not constant the internal timing in RaceChrono will get screwed. The problem might be related to having inconstant data rate, combined with the low number of updates...

    RaceChrono tries to mitigate the change in delays in the Bluetooth connection, using the counter field. So it figures out the (constant) update rate the device outputs, and then uses the counter field to estimate the "real" timestamp for each update. So if the update rate changes a lot and the counter field does not match this, it can throw off the timing.

    If you use timestamps it practically means you need to have GPS chip on the device, as the timestamp needs to be GPS timestamp, otherwise it will not be in sync with the other data recorded.
  • stez90stez90
    Ok. At the moment the rate is given by the microcontroller loop duration, so it is quite constant at 3.2Hz. There are small bumps, but so small that RC shows always 3.2Hz (1pixel in the graph).. I'll try to set it constant via timer interrupts and report back the results.
    Thank you!
    Stefano
Sign In or Register to comment.