Stomp Test Results

[daveryanwyoming]

....I can take my bike to the gym and use a 45 lb plate and use their scale to weigh the plate. .....

Sounds good, here's some other tips to help in terms of getting reliable test results.

- Hopefully the large hole in an olympic plate will fit right over you pedals (olympic plates slide right over speedplays) so that you can just hang the large plate on your pedal with the bike on level ground. If not you'll need to work out some arrangement to hang the weight from your pedal without letting it hit the ground.

- Wake the hub up with a quick spin, go into torque zero mode and zero the torque before each measurement.

- Slide the weight onto the pedal (assuming it fits) and slowly roll the bike backwards while watching the torque display. It'll take a second or two to settle out at each reading but slowly roll the bike backwards in small steps until you see the max reading and further rolling back makes the readings drop. Roll forward again slowly and double check the max torque reading and write the number down. Then remove the weight and double check the zero reading.

- Try this a couple of times for each gear combo you want to test to make sure your method is giving you repeatable results and try a couple of different gear combinations.

Good luck,
-Dave

 

[Felt_Rider]

I talked to the gym manager this morning about the scale and he has not calibrated the scale everyone uses near the locker rooms, but he took me over to a digital scale they just purchased for a competition they were having and he said they bought it for the accuracy. I suppose I will trust in what he says. If I can pick up a Park Tool digital scale from one of the local shops I will do that as well and compare, otherwise, I will have to order one.

The olympic plates have a 2-inch diameter hole and my Look Keo pedals are 4-inch, but I will use the CB eggbeater pedals off my mt. bike and those should work nicely.

If time allows (lot of work deadlines and training) I would like to do a photo catalog of the process and write up on my blog. Since this is all new to me I may not be the best person to do this so maybe one of you guys could help proof read it. The reason I would like to do this is to take the best tips and compile them in one location. Over the past weeks I have been gathering information for various internet resources and unfortunately most of what I find is on forums. The information is not only scattered around, but one has to sift through many pages of discussions and then try to determine if it is valid or not.

Just since starting this thread there are things you guys have revealed that I have not seen on the other web sites and unfortunately the most obscure of all is the manufacturer. Even the manual that came with this 2.4 SL was incorrect. It said to go to test mode 6 just as the representative said, but on mine it was test mode 3 and is test mode 3 on the online pdf manual.

I can only imagine how many PT users there are out there that are like me and have no clue about all of this. Before searching the web and posting something here I asked a friend that is a long time PT user and he didn't even know what "zeroing" the cpu was and had never done it. I suppose he must be fortunate that his setup has worked fine for all these years or perhaps it may be slightly out and he doesn't realize it.

 

[Felt_Rider]

I suppose if anything this has not hurt my training effort even though I potentially have about 4 months of bad data accumulated.

At least I was pushing harder because the cpu is/was potentially reading lower wattage.

It kind of reminds me of a rumor that back in the 80's the Russians were experimenting with some of their lifters by creating a set of rubberized olympic plates* that were actually heavier than the numbers they posted on the plates. So the lifter were unknowingly lifting heavier weights in training and then when they went to competition the weights were true and they were able to make the lifts easier.

Don't know if that was true or not. There were alway those types of myths of how certain nationalities were training their athletes by using odd concepts.

* - rubberized plates vary in thickness depending on the manufacturer. So one could potentially make a rubberized plate several pounds lighter or heavier and the lifter not be able to sense the difference. Other than think they are have an exceptionally good day or bad day. An experienced lifter can get a better feel of a metal plate because of the size and shape. They may not be able to sense a pound or two, but the since the circumference of an olympic plate is somewhat standard the thickness of metal can give someone experienced and idea of how much it weighs. I remember using those rubberized plates in college and it could take a few minutes to discern between a 20 kg (blue) and 25 kg (red) plate because the numbers and colors were worn off. Imagine loading up a bar with with 25 kg plates when you thought they were 20 kg. If you were lifting those easy you would be much stronger. Maybe have a power meter that reads lower wattage isn't so bad after all. - Just kidding of course

 

[Felt_Rider]

I hope to do the real test tomorrow at the gym, but I am one those kind of impatient type of people and I want to know things. I did a rough test this evening kind of based on Dave's test using the digital scale.

The Cycling Addiction: Powertap Torque Test - Rough Initial Test

Let me know if you see any errors in my "rough test" keeping in mind that it is my rough test.

 

[daveryanwyoming]

Looks to me like you did a good job with your test with the most obvious source of error being the digital bathroom scale used to measure the bucket of grout, but I doubt the scale was off by 22%.

The only thing I might change is to set the cranks slightly below level with the carpenter's level. Then load the weight on the pedal and slowly rotate the rear wheel backwards to double check that you've actually recorded the highest torque. It looks fine from the photos, but some crankarms have a slight taper that could throw you off a bit. Realistically though you'd have to be off by a whole lot to get a reading 22% low based on crankarm angle as it takes approximately 8 degrees of angle error to come in 1% low.

-Dave

 

[Felt_Rider]

I do hope to get the bike set up at the gym tomorrow and do the test based on your directions. I will use the better gym scale, measured olympic 45 plate and roll the bike back slowly and forward slowly as you described.

If all goes well I will try to get pictures and log everything and maybe others will find the information useful.

I was at the bike shop today and though I did not get the wheel and hub from them the shop owner helps me out a lot and said he would help me get it to Saris if need be. He stocks Cycleops/Saris products and may have better luck dealing with them.

 

[Felt_Rider]

Here is the result from my test at the gym this morning.
(bummer that the battery on my camera went dead after the first picture, but I completed the test and hope to return tomorrow with bike and camera battery charged)

I used the gym scale to measure one of the olympic 45 pound plates and the scale measured the plate at 45 pounds. The scale is not a digital, but is like those used in a doctor's office. The slide counter balance was aligned directly with the 45 mark.

I had installed the Crank Brothers egg beater pedals last night so that I could slip the plate on the pedal arm.

I zeroed the PT cpu
I logged that I was in the 34 x 22 gear first
Switched the cpu into torque mode
Placed the plate on the crank side pedal arm and held the bike up right and secure
I slowly inched the bike back and watched the reading
I slowly inched the bike forward and watched the reading
I continued the back and forward motion to get the max torque reading (edit - added for clarity)
I logged the displayed torque for that gear

I repeated the same described process for other gears

The results

34 x 28 - Displayed Torque 208 | Measured Torque (equation) 248.03 | Difference 0.16 or 16%
34 x 22 - Displayed Torque 164 | Measured Torque (equation) 194.88 | Difference 0.158 or 16%
34 x 17 - Displayed Torque 117 | Measured Torque (equation) 150.59 | Difference 0.223 or 22%
34 x 11 - Displayed Torque 61 | Measured Torque (equation) 97.44 | Difference 0.37 or 37%


The 34 x 17 ratio on this test is about the same error that I got on my rough test last night using the bike stand and measured weight with the wheel locked.

I like the method that I used this morning as provided to me by Dave. By setting the olympic plate on the pedal arm and inching the bike forward and back you can see the max torque reading as the weight and crank arm move, whereas, the bike stand it is difficult to know with the bike motionless that you have the crank arm in the most precise position. And believe me just a slightest fraction of movement makes a difference in what you will want to log as the reading.

I would like to go back to the gym and get photos so I can post this all to my blog so that it may be helpful to others. Dave, would it be alright to use your photo for your digital bike scale as another picture source and to point out alternative methods. In the meantime those who want to see my "rough" test in photo format I provided a link on my last post.

What happened to my wheel I am not sure. The bad data is evident since January this year and I am very careful with my equipment so I don't know what happened to it or what is going on just yet, but it does seem that wattage reading is in error.

 

[acoggan]

I like the method that I used this morning as provided to me by Dave. By setting the olympic plate on the pedal arm and inching the bike forward and back you can see the max torque reading as the weight and crank arm move, whereas, the bike stand it is difficult to know with the bike motionless that you have the crank arm in the most precise position. And believe me just a slightest fraction of movement makes a difference in what you will want to log as the reading.

If the "slightest fraction of movement" makes a measurable difference, then something strange is going on with your hub, or you're not making the measurements when truly motionless. You can actually have the cranks a fair off of level and it does not have a significant effect on the applied torqe (as simple trig will demonstrate).

(Note that the above is especially true with the PowerTap since the resolution of the display is ~10x less than with an SRM.)

 

[Felt_Rider]

If the "slightest fraction of movement" makes a measurable difference, then something strange is going on with your hub, or you're not making the measurements when truly motionless. You can actually have the cranks a fair off of level and it does not have a significant effect on the applied torqe (as simple trig will demonstrate).

(Note that the above is especially true with the PowerTap since the resolution of the display is ~10x less than with an SRM.)

My description may have implied too much difference. I think I got too caught up in how little the crank arm tilted up and down during the process and yet the PT did sense the motion.

To be more precise so you guys can help me, it would move up a watt or two with about 2 inches of rolling back and down a watt or two by rolling forward a couple of inches.

 

[RChung]

The error pattern across the cassette could be consistent with a bad bearing.

Remember way up thread when Andy said "it seems obvious" your Coggan Stomp technique was off? Ironic, ain't it? This is why I said that the problem with the Coggan Stomp isn't that it can't be done well, it's that you can't know when it's being done poorly.

 

[acoggan]

it would move up a watt or two with about 2 inches of rolling back and down a watt or two by rolling forward a couple of inches.

You mean in-lbs, not watts...but in any case, it sounds as if you may not have waited long enough for the PowerTap to stabilize before taking a reading. After all, the crank only needs to be w/in +/- 5 deg of level for the applied torque to be w/in +/- 0.4% of maximum. A difference of that magnitude would have been below the limit of the display's resolution in two of the four gear ratios that you tested, and would barely be detectable in the other two.

 

[acoggan]

The error pattern across the cassette could be consistent with a bad bearing.

It isn't an error across the cassette, it is a case of non-linearity.

Remember way up thread when Andy said "it seems obvious" your Coggan Stomp technique was off?

And indeed, it was: Felt Rider subsequently confirmed that he was holding onto the brake lever (rather than, e.g., strapping it down as I've always recommended), thus making it difficult to apply all of his weight to the pedal.

the problem with the Coggan Stomp

As the very first person to ever publically describe how to statically calibrate a PowerTap, I will remind you that the "Coggan Stomp Test^TM" covers the use of inanimate objects as well as body mass.

 

[Felt_Rider]

You mean in-lbs, not watts...but in any case, it sounds as if you may not have waited long enough for the PowerTap to stabilize before taking a reading. After all, the crank only needs to be w/in +/- 5 deg of level for the applied torque to be w/in +/- 0.4% of maximum. A difference of that magnitude would have been below the limit of the display's resolution in two of the four gear ratios that you tested, and would barely be detectable in the other two.

Oops!!

Bottom line as an end user I just want to get back to having confidence in my equipment so that I can be focused on training and not the equipment.

Apart from the test and in my simple mind I do not see how in 2008-2009 I can be on a particular course that I use for a regular 2 hour low L3 training ride and always hit near 174 NP with all the same speed and time splits and then in 2010 on the same course with the same effort end up with NP of 120 on each event. A friend of mine does the exact same course with the same speed and time and also ends with a 174 +/- NP. All of my other rides for 2010 are off as well. No data drops and fresh batteries at all times on a wired unit.

I am willing to pay for the repair. I have tried to do the test to prove to the manufacturer that the wiring harness is okay, the batteries have been replaced, there are no transmission issues that I can tell. No one else is authorized to work on the wheel. What else do I have to do to get back to just focusing on training?

 

[acoggan]

Oops!!

Bottom line as an end user I just want to get back to having confidence in my equipment so that I can be focused on training and not the equipment.

Apart from the test and in my simple mind I do not see how in 2008-2009 I can be on a particular course that I use for a regular 2 hour low L3 training ride and always hit near 174 NP with all the same speed and time splits and then in 2010 on the same course with the same effort end up with NP of 120 on each event. A friend of mine does the exact same course with the same speed and time and also ends with a 174 +/- NP. All of my other rides for 2010 are off as well. No data drops and fresh batteries at all times on a wired unit.

I am willing to pay for the repair. I have tried to do the test to prove to the manufacturer that the wiring harness is okay, the batteries have been replaced, there are no transmission issues that I can tell. No one else is authorized to work on the wheel. What else do I have to do to get back to just focusing on training?

Sorry, I did not mean to imply that I did not believe that your PowerTap is malfunctioning. Regardless of precisely how you did or did not perform your measurements, I think it is clear that it is. The problem, however, will be in convincing Saris that this is true (although it sounds like your LBS may be of assistance there).

 

[Felt_Rider]

Sorry, I did not mean to imply that I did not believe that your PowerTap is malfunctioning. Regardless of precisely how you did or did not perform your measurements, I think it is clear that it is. The problem, however, will be in convincing Saris that this is true (although it sounds like your LBS may be of assistance there).

No offense to me. My frustrations are starting to show through from having to prove to the company why and what is going on.

 

[Felt_Rider]

This thread started because I was suspecting something going on with my powertap hub and/or other hardware.

It took me a while to figure out that I potentially had an issue and I became more suspicious when I went out to do a ride where the ride is solo, I use the same time limit and try to use the same speeds. Because the group rides have so many variables I did not recognize any issues, but would wonder why I would have such a low NP for such long spirited group rides.

You guys helped me with understanding my question about a weighted torque test and I attempted the test, but perhaps did not do it as carefully.

I contacted the manufacturer and with great difficulty finally was able to speak to a representative who asked for a number from the test mode. I gave them my number at 519 and the representative told me that my hub was fine and that I should try changing the batteries and check to may sure I am getting good transmission. Also make sure that I zero the cpu while the bike is still.

I went today to test the hub as directed by the representative. I put new batteries in even though the other batteries seemed to be okay, I zeroed the cpu while the bike was motionless and I set out trying to mimic as best possible an effort I did in September 13, 2009. Almost every effort for 2009 on this route with a 1 hour out and back and I typically end up with a ~170 NP, but not this year.

I posted a comparison of the two rides on my blog. Blog Link
Could someone who has knowledge of these things please take a look and see.
Does it look as if the transmission data is good throughout the ride?
Is there anything that stands out about the data posted?
Thanks

Edit: one more thing I wanted to add for clarity is that I seem to get the same poor result no matter which bike I use the PT wheel. Each bike has it's own wiring harness for the wired 2.4SL hub.

 

[Felt_Rider]

I sent the wheel in last Friday and received it today so that is a pretty good turn around for service considering much of that was shipping time.

I tried the rough test using my bike stand, wheel locked and a roughly measured weight (~53 lbs) applied to the pedal.

Before sending it in it was reading about 22% error based on the rough test.
After service it is reading 2% based on the same rough test.

Hopefully I will have my first good data in WKO for outdoor training this year.

 

[RChung]

Before sending it in it was reading about 22% error based on the rough test.
After service it is reading 2% based on the same rough test.

Very good. Being able to check the accuracy of your hub is a great relief, isn't it? I would never recommend a PM that couldn't be checked for accuracy.

Best of luck in your training.

 

[Felt_Rider]

Very good. Being able to check the accuracy of your hub is a great relief, isn't it? I would never recommend a PM that couldn't be checked for accuracy.

Best of luck in your training.

A big thanks to all of you guys that helped and to those who may stumble across this thread in the future the lesson I learned is:

1. When asked by the manufacturer if your hub torque test number is 512 and you number happens to be close to that number it doesn't necessarily mean that your hub is okay. Mine was close at 519, but still out of calibration. If you believe it is out of calibration and you did all the basic checks like, data transmission, batteries and such just call and ask for a RA number and send it to the manufacturer rather than delay. (If it is out of warranty they charge a flat fee and completely rebuild the hub)

2. As RChung and Dave mentioned previously that every device that is used for precise measurement needs to checked at regular intervals. Fortunately for me mine was out so bad that I eventually picked up on it using the same course on a solo ride. (Group rides are hard to pick up the error) That course with the same environment and conditions allowed me to see that there could be a potential issue and there was. From there I used a weighted torque test to help understand more clearly. If the error had been smaller I could have gone many more months not knowing that it was out. Now I know to check more often to see if the calibration is drifting.

3. Zero the cpu before every ride or training event.