Results 1 to 5 of 5

Thread: Trackman and the center of mass

  1. #1
    Join Date
    Jan 2012
    Posts
    4,096

    Trackman and the center of mass

    Even among experienced Trackman users, there exists the misperception that Trackman tracks the "center of mass" for determining its reported clubhead data, such as path and clubhead speed. However, radar can't possibly weigh and balance the clubhead to determine the COG, so Trackman must be doing something else. According to the most recent Trackman newsletter, it purports to track the geometric center of the clubhead, which Trackman considers a reasonable approximation of the clubhead's center of mass (btw, center of mass and center of gravity, or COG, mean the same thing):


    Name:  geometric center.PNG
Views: 1455
Size:  188.2 KB



    Of course, analyzing the shape of an object traveling in excess of a 100 miles-per-hour must be quite a technological challenge, and our research shows that Trackman's reported clubhead speed is typically more consistent with a point located well towards the heel and not at the clubhead's geometric center:


    Name:  chs no smash.PNG
Views: 1356
Size:  17.3 KB



    A notable exception was the Callaway FT-i: Trackman's reported clubhead speed was more consistent with the Phantom-determined clubhead speed at the middle of the clubface. A look at the clubhead suggests why that was the case: it's big, square and flat, with relatively sharp edges.


    Name:  fti 3.PNG
Views: 1329
Size:  124.4 KB



    Here is the clubhead speed data for the FT-i; in this case, Trackman seems to think the geometric center of the clubhead sits to the toe-side of the center of the face:


    Name:  ft-i chs.PNG
Views: 1369
Size:  34.8 KB



    More typical is what we saw with the Krank Black Rage driver, swung by long-drivers Geoff Olohan and Tyler Kellett (ironically, the clubhead is white!). Shaped like a stealth bomber, it is harder for radar to see.



    Name:  krank 1.PNG
Views: 1342
Size:  275.0 KBName:  krank 2.PNG
Views: 1386
Size:  352.5 KB
    Name:  krank 5.PNG
Views: 1306
Size:  223.7 KB



    Here is Geoff's clubhead speed data:


    Name:  geoff CHS no smash.PNG
Views: 1654
Size:  45.1 KB



    Looking at the Phantom v. Trackman clubhead speed data, it appears that Trackman has guessed that the clubhead's geometric center is somewhere between the center and the heel tracking dots, favoring the heel. The heel and center tracking dots are 1.5 inches apart, so let's say Trackman believes the club's geometric center is an inch towards the heel. Geoff hit one in that location and below is the post-impact photograph as well as a screen cap of the ProAnalyst tracking. Does that look like where the clubhead's geometric center should be or where the center of mass belongs? No. But that's what Trackman thinks!


    Name:  go185l vp6.jpg
Views: 1332
Size:  121.9 KB


    Name:  go185l.PNG
Views: 1351
Size:  96.3 KB





    Jeff

  2. #2
    Join Date
    Jan 2012
    Posts
    4,096
    FYI, the clubhead speed is calculated by ProAnalyst on a frame-by-frame basis (24,000 frames per second), then filtered with both a low-pass Butterworth filter and a mean-median filter. This is what the processed data looks like when graphed; zero represents initial ball/club contact. The impact interval is shaded in gray.





    The filtering causes the tracking dot speeds to turn down before impact, which, of course, doesn't actually happen. However, that effect is difficult to avoid when analyzing discrete data. In the data tables we used the tracking dot speed from the frame immediately prior to 0.0, so we may have understated the speeds a bit in that presentation.




    Jeff

  3. #3
    Join Date
    Aug 2012
    Location
    Pacific Palisades, California
    Posts
    73
    Quote Originally Posted by Jeff Martin View Post
    FYI, the clubhead speed is calculated by ProAnalyst on a frame-by-frame basis (24,000 frames per second), then filtered with both a low-pass Butterworth filter and a mean-median filter. This is what the processed data looks like when graphed; zero represents initial ball/club contact. The impact interval is shaded in gray.





    The filtering causes the tracking dot speeds to turn down before impact, which, of course, doesn't actually happen. However, that effect is difficult to avoid when analyzing discrete data. In the data tables we used the tracking dot speed from the frame immediately prior to 0.0, so we may have understated the speeds a bit in that presentation.




    Jeff
    Dear Jeff,

    I am appreciative of the effort and resources you and your team are expending to provide the very interesting results you are posting, thanks.

    From the data above, I 'see' some thought provoking possibilities, for instance the measured relative movement between the heel and toe of the club head before, during and after impact. I am cautious about using elementary science here but am willing to 'go first' in raising some questions/comments.

    97 test events is a rich statistical data base and it looks like about 10 swings per golfer/session were used, also a solid data base to draw meaningful statistical conclusions, but first a few physics basics.

    The velocity as noted for the heel is VERY close to the hosel (at least for some first order analyses), and during the pre-impact time frame for what I believe is swing 26R above, is noted to be 139.5 mph. For a small interval of time before impact, the velocity of the shaft, V=r x omega, the radius of rotation, times the angular velocity in radians per second. (I certainly believe this is basic to a Princeton graduate, but for this post I want to be clear to all the readers to elicit their opinions too).

    So the higher velocity of the toe as shown by your data reduction efforts have to come from (1), the difference in length/radius to the toe and (2) additional rotation of the club head around the hosel, or 'gamma' rotation as defined by Dr. Steve Nesbit in his research.

    While the 'instantaneous' center of rotation at impact can be controversial, IMO it is , for a driver approximately the lead wrist joint, and the additional length from this point to the toe of the driver is 'about' 2 inches more. So for (1) above, the 'delta velocity' of the toe over the heel should be the percentage difference of this radius or about 2/50 or 4 percent higher than 139.5 mph, or about 145 mph versus the 150.4 mph you measured/calculated from the Phantom data base.

    That leaves 5 mph to have been created by rotation of the club head AROUND the hosel from 'gamma' torque/rotation, to me a fascinating look INTO the dynamics even before impact, IF the measurement and data manipulation errors are small in comparison. In my previous career, this was referred to as an accuracy/precision determination, and sometimes, the passion for precision can overtake the value of the realistic accuracy.

    Do you have some thoughts and plans in studying this area, and are you interested in others comments ??

    I hope so, and look forward to a yes, and the possibility of having access to a players swing by swing data package and promise to expand on the appearance of the heel/hosel standard deviations in velocity for multiple swings providing insight into my favorite subject of dynamic balance and stability.

    Appreciatively,
    Art

  4. #4
    Join Date
    Jan 2012
    Posts
    4,096
    art-

    As I've posted before, we plan on making the data available. I'll keep you posted.


    Jeff

  5. #5
    Join Date
    Aug 2012
    Location
    Pacific Palisades, California
    Posts
    73
    Quote Originally Posted by Jeff Martin View Post
    art-

    As I've posted before, we plan on making the data available. I'll keep you posted.


    Jeff
    Dear Jeff,

    Thanks again for your commitment of resources and time in this fascinating research project.

    I go to this site several times a day hoping to find any updates regarding the research you are doing.
    For me, and others, without doubt, it is a new level of seeing and hopefully understanding the conditions at ball to driver impact.

    In my admittedly limited world of interest in 'dynamic stability' of the lower body, upper body and lead arm, a 'growing' hypotheses from the data you have posted is that the 'dispersion/scatter' of the impact points for each person tested could be the result of dynamic instabilities AND involuntary reflex reactions.

    As a result, I have a significant interest in seeing as many of these club face 'scatter' pictures, impact point location calculations and Trackman performance matrices as have already been processed.

    I certainly will post my findings from the data provided.

    Appreciatively,
    Art

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •