Results 1 to 2 of 2

Thread: How to orient Phone so Accel X and Accel Y are 0

  1. #1
    Potential Tuner
    Join Date
    Jan 2022
    Posts
    6

    How to orient Phone so Accel X and Accel Y are 0

    Hi There:
    I'm trying to use Accel X and Accel Y, to see if I'm keeping the car at maximum load G in the corner, recognizing that G= SQRT((X*X)+(Y*Y))

    The first challenge is the phone, when mounted to be able to capture Video never shows 0 for either of Accel X and Accel Y, so the numbers captured are not realistic. (First Attachment)

    As well, due to the values captured for each is a 4 decimal places, the display of X and Y is very had to use (Second Attachment)

    What is anyone else using to confirm that you are keeping the car at the limit of it's traction, not over slowing in the corners?

    Thanks for any insights you can provide.
    Attached Images Attached Images

  2. #2
    Potential Tuner
    Join Date
    Mar 2023
    Posts
    3
    I assume Track Addict is just displaying the raw accelerations reported by your phone. If you dont have GPS, there is a good chance that sensor bias will show up in those reported accelerations. This would possibly explain why you are seeing non-zero accelerations while stationary. Normally, GPS is used to correct for these biases. I'm wondering if TrackAddict has insight into what exactly it is displaying from our phones. Also, I'm curious what their definition of the X, Y, and Z axes are. If our phones are actually reporting specific forces, rather than accelerations, we would see the normal force due to gravity in the upward direction. If this is the case, error in the initial leveling could cause this normal force to appear in other axes (X and Y). One way to confirm this is taking the magnitude of all three axes. If the value is 1G, then your problem is likely due to error in the initial estimate of attitude (i.e., where 'down' is). If the value is zero, that would indicate the phone compensates for gravity, thus displaying acceleration. If the value is biased (1.1G) you could infer that your accelerometers are reporting biased data.

    I'm curious how/when TrackAddict performs its initial leveling and if there are ways we could improve that initial attitude estimation to get better acceleration data.