Stomp Test Results

Discussion in 'Power Training' started by Felt_Rider, May 23, 2010.

  1. Felt_Rider

    Felt_Rider Active Member

    Joined:
    Oct 24, 2004
    Messages:
    3,257
    Likes Received:
    27
    For a number of weeks I have been skeptical of my power output numbers based on a particular route that I am familiar with the results and data in WKO. Like today climbing some short hills and having to come out of the saddle and yet my cpu would barely break 100 watts and then toward the summit would hit maybe 220, but these are short efforts that I know would typically be in the 300's.

    I finally got around to doing a "stomp test"
    If I did everything right in the spreadsheet and on the bike here is the data.
    Spreadsheet link

    Results
    Measured torque (equation) = 366.24 in-lbs
    Displayed torque (cpu) = 310

    Since most of the web writings on a stomp test are a little vague here is what I did.
    Weighed myself = 171.0 lbs
    Put the gear on the 50/16 combination
    I mounted the bike in the trainer to get a stable platform
    Set the cranks at 3 and 9 o'clock position
    Set the cpu in torque mode (holding down the select button until it flashes)
    Applied the rear brakes and stood with all my weight on the left pedal = 310 avg (it bounced a little between 309 to 311 but stablized at 310)
    Switch the pedals around and did the same thing on the right pedal = 310 avg

    Did I do the procedure correctly?
    If I did the error is fairly large
    What is my next step? Send the wheel to Saris?
    Any experiences on how long they keep the wheel and such?

    Thanks
     
    Tags:


  2. RChung

    RChung New Member

    Joined:
    Sep 12, 2006
    Messages:
    306
    Likes Received:
    1
  3. Felt_Rider

    Felt_Rider Active Member

    Joined:
    Oct 24, 2004
    Messages:
    3,257
    Likes Received:
    27
    Thanks for the link. I did not get that thread in a search, but it is right on the money for my issue. Wish I had a picture of Dave's method.

    I will take my bike to the gym in the next few days and use a 45 lb plate.
     
  4. daveryanwyoming

    daveryanwyoming Well-Known Member

    Joined:
    Oct 3, 2006
    Messages:
    3,857
    Likes Received:
    97
  5. acoggan

    acoggan Member

    Joined:
    Jul 4, 2003
    Messages:
    3,047
    Likes Received:
    9
    Perhaps not, but it seems obvious from the OP's description that he didn't have all of his weight on the pedal.
     
  6. Felt_Rider

    Felt_Rider Active Member

    Joined:
    Oct 24, 2004
    Messages:
    3,257
    Likes Received:
    27
    Thanks Dave.....the picture and description are great.

    I was doubtful that I was doing it right. I tried doing some Googling on the subject, but only came up with vague descriptions at best for using body weight. I found it hard to get all of my weight on the pedal and I imagine leaning over to apply the brake would remove some of the body weight. I suppose I did not put enough thought into it.
    _____________________

    Below are rides that I use for comparison
    (all my rides in 2010 are reading lower for NP and I think all my 2010 outdoor data is bad)

    2009 when it seemed to be working okay
    2010 about the same amount of effort
    I was not suspecting a PT issue when I wrote the 2010 blog entry.
     
  7. RChung

    RChung New Member

    Joined:
    Sep 12, 2006
    Messages:
    306
    Likes Received:
    1
    How do you know that? Maybe his PT really *is* off.

    That's the problem with The Coggan Stomp. How would someone who hasn't done it before know that his technique was off rather than the PT being off?

    "I see no point in introducing unnecessary sources of variability if I can avoid it."
     
  8. Felt_Rider

    Felt_Rider Active Member

    Joined:
    Oct 24, 2004
    Messages:
    3,257
    Likes Received:
    27
    It just seems too obvious from my perspective that something is not right and not being a long time user I have done all that I know to do. I have checked the wiring, batteries and some other things. My data in WKO does not show drops. Speed and cadence are consistent with my Garmin data.

    The links above show a large difference for about the same effort on the same course and a friend of mine's data for the same course, mileage and speed also reflects the same as my 2009 data link above. His NP was almost the exact same and yet my data for 2010 is low. If I push hard it will go up, but for hills that I feel as if they would be 300 watts + show like 200 watts.

    Yesterday I doubted on several climbs where I was out of the saddle and yet the data would be around 100 watts. I was on a 3 mile stretch of 2% incline and felt like I was approaching my threshold and yet the power meter was showing 50 watts and sometimes less. PT cpu speed reading was consistent with the Garmin and there were no data drops. I think if my my hub was out a little I may go many weeks without knowing that the data is bad, but this just seems too obvious. :confused:

    I think I am going to go ahead and call Saris today and see about sending my wheel to them. I believe even if I pursued a better torque test I think I am just delaying getting back to tracking my outdoor training with confidence.

    Any other thoughts or things to consider before I call?
     
  9. acoggan

    acoggan Member

    Joined:
    Jul 4, 2003
    Messages:
    3,047
    Likes Received:
    9
    I don't. That's why I wrote "seems" instead of "is". (Oh, wait, I forgot: English isn't your native language.)

    Did I say that it wasn't?

    The same can be said of anyone calibratng any scientific instrument.

    The point in this case is that your body mass is 1) high, and 2) always readily available. To successfully use it, though, you do have to possess reasonable balance (and understand a little something about simple trig and basic physics). If you don't, you are better off using weight plates.
     
  10. RChung

    RChung New Member

    Joined:
    Sep 12, 2006
    Messages:
    306
    Likes Received:
    1
    They'll probably ask you what the stored torque value in the PT head is. It should be in the ballpark of 512 or so.
     
  11. RChung

    RChung New Member

    Joined:
    Sep 12, 2006
    Messages:
    306
    Likes Received:
    1
    Perhaps not but my response to Felt Rider was both on point and helpful while yours, um, wasn't. I guess I'll just have to seek solace in that.
     
  12. Felt_Rider

    Felt_Rider Active Member

    Joined:
    Oct 24, 2004
    Messages:
    3,257
    Likes Received:
    27
    I did the test mode and the value is 519 and I passed that on to the contact at Saris.
    Thanks again for the help
     
  13. swampy1970

    swampy1970 Well-Known Member

    Joined:
    Feb 3, 2008
    Messages:
    10,057
    Likes Received:
    185
    Just out of curiosity, how important is the choice of gearing for the stomp test? Is it more accurate in 29x23 than it is in 53x12?

    Has anyone compared?
     
  14. acoggan

    acoggan Member

    Joined:
    Jul 4, 2003
    Messages:
    3,047
    Likes Received:
    9
    Back-to-front: yes.

    Regardless of whether you use your body mass or some other mass, varying the gear ratio does two things: it changes the relationship between the torque applied to the crank and the torque applied to the hub, and it changes the angle at which the chain pulls on the cassette.

    Based on test using a chainwhip instead, at least with my wife's PowerTap the effects of the latter are below the limit of detection. I therefore use the 39 x 23 so as to maximize the torque applied to the hub, so as to calibrate it over the full range of torques it experiences in actual use.
     
  15. RChung

    RChung New Member

    Joined:
    Sep 12, 2006
    Messages:
    306
    Likes Received:
    1
    First, it's not at all important for the SRM or Quarq. It only matters if you have a PT.

    Second, the result you get for the PT will be more precise in 39x23 than 53x12; however, the largest source of poor precision (and often, poor accuracy, too) is likely to be in using the Coggan Stomp rather than using a fixed and known mass.

    Third, if you're using a PT it's good to check the measured torque over several combinations of chainwheel x cog to make sure that the response is consistent -- sometimes you'll get different results from the innner cog than the outer cog.
     
  16. acoggan

    acoggan Member

    Joined:
    Jul 4, 2003
    Messages:
    3,047
    Likes Received:
    9
    In my experience, the SRM (and probably the Quarq, although I haven't tested it) is more impacted by variations in chain angle than the PowerTap. IOW, while technically the gear ratio doesn't matter, the particular combination of chainring and cog that you use does have a small, but measurable, effect.
     
  17. acoggan

    acoggan Member

    Joined:
    Jul 4, 2003
    Messages:
    3,047
    Likes Received:
    9
    As amusing as your attempts to smear my name might be, I feel compelled to point out that my "stomp test" covers the use of inanimate objects as well (cf. the first-ever public description of how to statically check the calibration of a PowerTap, which was really just an extention/application of a procedure I first developed in 1977).
     
  18. Felt_Rider

    Felt_Rider Active Member

    Joined:
    Oct 24, 2004
    Messages:
    3,257
    Likes Received:
    27
    I have sent in my number per the request of the Saris contact, but have not heard back from them yet.

    None of the manuals give much explanation other than 512 is the zero point for the hub. My cpu reading was 519 and from what I have gathered off the web that the number can vary a little bit. How far off of 512 would it need to be to be considered "in need of calibration?" I not stumbled across good in depth discussion to these details.

    So while I am waiting for official word as what to do next, I am really curious as to what the number indicates or does it give some sort of diagnosis as to why my wattage data seems to be lower than what I would expect and by comparison to similar efforts on the same course.

    At the moment the folks at Saris have not requested me doing a test as you guys are describing. They just want to know if the number is at 512.

    It is a little hard to sift through all that is out there on the web.
    One day when I get past all of this I will have learned a good bit, but for now I just want to get back to training with confidence that my data is reliable and that all in WKO reflects the right information.
     
  19. daveryanwyoming

    daveryanwyoming Well-Known Member

    Joined:
    Oct 3, 2006
    Messages:
    3,857
    Likes Received:
    97
    Strain gauges like those used in PT hubs generate some output signal even when they're unloaded. The 512 (or thereabouts) number is the unloaded output level of the strain gauge, in other words it's unloaded bias point. Basically Saris has sampled a bunch of 'good' PT hubs off the assembly line and for known working hubs the average unloaded bias point measurement is 512 and based on the statistics of their samples and the standard deviation of good units off the assembly line they feel confident that a number near 512 is representative of a 'good' unit.

    It's a simple quick test and lot's of manufactured electronic devices use similar simple tests to show that a unit is within expected norms. It does not however definitively assure you that the loaded measurement while actually riding is correct. It's like having an old style mechanical bathroom scale with a stretched spring, you can adjust it so that the zero reading lines up but if the unit isn't functioning properly it's hard to know for certain that the loaded measurements you care about are accurate.

    Static torque tests with known weights give you a loaded data point but Saris' official stance seems to be to dismiss static torque measurements and won't give credence to customer complaints based on 'stomp test' results. SRM takes a very different stance and actually provides instructions on how to perform static torque tests in their printed manual which of course is the standard 'stomp test' using known weights and several repeated measurements to reduce error.

    Anyway, that 512 number is a measure of the unloaded strain gauge bias point, it shows that your unit is very similar to known good units coming off the assembly line but it doesn't guarantee that your unit reports torque (and hence power) accurately at working loads. A carefully performed static torque test can tell you that but you do have to perform the test carefully, ideally with accurate weights. Weight plates at gyms can vary by a bit, but the error in heavier weights tends to be a smaller percentage of the weight which helps. FWIW, I use a 45 pound olympic plate that's marked in both pounds and kg that I've weighed on both postal and accurate digital scales and it's a bit heavier than marked. It comes in at 20.64 kg vs. 20.4 kg expected but that's still within ~ 1% of marked weight and the PT's stated accuracy is ~+/-1.5%. Lighter weights will tend to have more error in percentage terms.

    -Dave
     
  20. Felt_Rider

    Felt_Rider Active Member

    Joined:
    Oct 24, 2004
    Messages:
    3,257
    Likes Received:
    27
    Great post Dave.....thanks a bunch

    I can take my bike to the gym and use a 45 lb plate and use their scale to weigh the plate. I will ask the owner if their scale has been calibrated lately.

    I also plan to order a digital scale (been meaning to get one anyway) and will try your method at a later date as noted in the picture you provided.
     
Loading...
Loading...