View Single Post
      08-23-2008, 05:11 PM   #9
Lieutenant General
swamp2's Avatar

Drives: E92 M3
Join Date: Sep 2006
Location: San Diego, CA USA

iTrader: (1)

Originally Posted by lucid View Post
Thanks Swamp. This is rather interesting preliminary data. I am looking forward to seeing more measurements taken at a higher sampling rate (300 Hz on the first graph seems to be just the bare minimum for capturing anything meaningful), and a comparison of the same shift, say 2->3, in D and S modes under WOT so that we can isolate any effects of your throttle input during the shift, which might not be constant (I assume that's why we are not seeing a significant drop in acceleration after the 2->3 shift, which should happen due to the change in gearing, or is it just the scale of the y-axis combined with the noise that makes it difficult to observe this?). I realize that is not easy to do on public roads though. Is there any reason why you are not using the 1kHz sampling rate? I believe the accelerometer will do that, right? And, yes mounting the accelerometer rigidly should decrease noise significantly.
  • 300Hz does provide enough resolution to get about 10 sample points across the shorter D3 shift. So it is providing shape and duration. For example even along the largest slope sections there are quite a few samples. Those slopes, as far as I can tell are real.
  • If you look at the data on a longer time scale the steadily decreasing acceleration before and after the shifts is clear. Noise is a problem and the accel may (pure speculation) have a slightly transient sensitivity after the large jerk shift events. However you can clearly see this effect on the other trace, despite the similarly short duration covered.
  • Just did not bother with 1kHz yet. 300 seemed sufficient. I will try it eventually.