Bug? Time between GPS and RC3 gauges in app is not the same

Hello,

It's best to look at the screen shot to understand. I have 50Hz of RC3 and GPS data in 1 stream being sent to the Android. In RaceChrono I go to the gauges panel and both sets of data are updating correctly. But the time gauge can differ from 1 to 10 hours between the 2 and I get weird dates like 4/28/58. I don't see weird data when I look at the output from my microcontroller or through a terminal program connected to the BT. When I spot check the GPS and RC3 time codes are the same. The session directory also has the device output file included. I assume the difference between the 2 at this level will affect the merging of the data in the export? (have not confirmed this, I will)

Screen image @ https://drive.google.com/open?id=0B_iB0vrN7TEWVzBDbG1ualJnRk0 and https://drive.google.com/open?id=0B_iB0vrN7TEWRVJRSGdkTHc2dFE

Can get screen shots and session directories @ https://drive.google.com/drive/folders/0B_iB0vrN7TEWMlRCZ3ZneHVrdms?usp=sharing

Not sure if it's a bug in your app or an issue in my data logger code. It is reproducible. Using latest app in Play (4.7.1)

Thanks, Jeff

Comments

  • Turn on expert setting "Save device output" and dig the .nmea file from the session folder. You can verify the output like that. Pick a wrong time stamp and search it from the file (with time zone conversion). See if the sentences are in sync there.
  • edited January 2017
    If the output looks ok to you, I can check it too
  • I think the issue is one of bandwidth. I think RC3 plus full GPS content both at 50Hz simply overwhelms the bandwidth available. I need to do some maths to do the calculations to determine what bandwidth I require. If I disable the output of my sensors (no RC3 sent) then I get no errors in the GPS gauges in RC. I also added some error count tracking in my logger where an error is a multi-start string being sent (ie $GPGGA,1235324.23,22342$$GPRMC,2323235 ) and in 100,000 output GPS lines I got < 10 error counts (and my CTS/RTS signals never fire). As soon as I enable my RC3 output then my error count started to increase instantly and quickly and my RTS/CTS pins start to light up regularly between my logger and the BT module.

    Solutions looking at reviewing:
    - lowering the signal rate of GPS and/or RC3
    - separating the GPS and RC3 to 2 separate bluetooth channels and using the RC3 count field instead of timestamp
    - use bluetooth only for GPS to RC and log locally on the logger to SD card

    For now the weird results are probably on my end or results of limitations of the communication channel bandwidth. Going to do some research on the Bluetooth RFCOMM channel bandwidth as well

    Thanks, Jeff
  • edited January 2017
    My advice would be just drop sentences (data points) when you hit the flow control. That way they don't queue up, and you don't get delays. Remember that you don't have fixed data rate over Bluetooth, but it varies depending on radio interference and amount of devices communicating near by.
Sign In or Register to comment.