Dutchy Marc

Blog

Here the registered trainings of Dutchy Marc are shown. If you want to see more, e.g. heartrate, intervals, planning etc, you must register at BikePlanner (free) and ask Dutchy Marc if he/she wants to authorise you to view his/her datas. After registration you may register your own trainings and choose if you want to display them in the public Blog section.

Dutchy Marc:

Marc Breukers,

  • 42 jaar,
  • getrouwd met Linda,
  • woonachtig in Eindhoven.
  • Vader van 2 kinderen. Adne(8) & Imke(6).
  • Werkend voor ASML.
  • Fietsen? ja, als hobby en uitlaatklep.
  • Nog meer hobbies? ja

Special:

power of a powertap versus tacx

 40 weeks and 2 days ago:
Sunday, 29 November 2009


1:10 h
33.1 kms
28.4 km/h

Purpose: A qualification of my PowerTap on my Tacx Fortius.  Analysis is performed in different dimensions.  The cases which are tested:

I.  constant gear, constant cadance varying power (50[Watt]-350[Watt])
II.  constant gear, constant power, varying cadance (60-100[rpm])
III.  constant cadance, constant power, varying gear

The following analysis is done, with figures and tables and conclusions in the text below.  By clicking on the image (s), the extension of the figure is achieved.
First, the total training in a graph is presented as a function of time in figure 1 and figure 2.  Comparing the measured power of the PowerTap w.r.t the Tacx in a graph, the noisy signal of the PowerTap power is filtered (from green to black line in figure 2). More analyses on the measured noise is given in table 1 (see APPENDIX below) and figure 6

The "high frequent" noise on the measured power of the powertap is, anyway, about 10-15% to it's value. To minimize this noise in analysis, extra averaging is recommended to notice the relevance of the numbers. In the numbers of table 1 (see APPENDIX below), noise qualification is listened ("pt std", "pt range"). Figure 5 presents these noise numbers.

This training in total is presented in figure 1 and figure 2. The training can be divided into 6 distinct phases.  Each stage has its own purpose:
1.  Warming up for the systems, means myselfKnipogen  and the Tacx system
2.  Compare the measured power  - between the Tacx and the PowerTap on a wide range of power - with a fixed bike gear (= case I) (figure 3, figure 4, figure 5, figure 6
)
3.  Compare the measured power - between the Tacx and the PowerTap with a fixed set point on the power, a fixed bike gear, but a vary-ing cadance between 60-100[rpm] (= case II) (figure 7 and figure 8
)
4.  Compare the measured power - between the Tacx and the PowerTap, all fixed, cadance, gear and power setpoint. (figure 9)
5.  Compare the measured power - between the Tacx and PowerTap with a fixed set point on the power, a variable gear and a fixed bicycle cadence (= case III) . (figure 10)
6.  Compare the measured power - between Tacx PowerTap with some random and sudden accelerations.  Gear and Cadence are not of interest (figure 11)

 Each objective from above is discussed in more detail below (extra tables and figures)

figure 1: total trainings data unfilteredfigure 1: whole trainings as function of time.

 

figure 2: total trainings data filteredfigure 2: whole trainings as function of time (power filtered).

 Purpose of Phase 2: (case I)
Compare the measured power  - between the Tacx and the PowerTap on a wide range of power - with a fixed bike gear (= case I)
1.  Compare in absolute numbers
2.  Comparable in relative terms
3.  Compare the numbers in the reproduction of the PowerTap

Results of Phase 2 (case I)
•  The PowerTap measured systematically lower values as the Tacx
•  The higher the capacity the greater the delta between the power of the PowerTap and Tacx into absolute perspective.  In a relative perspective, the constant delta is somewhere around 7%
•  The reproduction of the power of the PowerTap is dependant of the absolute power in a absolute sense. In a relative sense it is more or less equal. (see table 1)
 

The most interesting so called AVG-numbers from table 1 (see APPENDIX below) is listened in figure 5 for better overview and oversee some trending. The ratio of the power between the tacx and the powertap is almost constant over all powers (first subplot of figure 5). The quotient is about 93%. This means the Tacx system gives systematically 7% higher values for all powers. To quantify the measured noise of the powertap (means the spread in values in the green lines in figure 1 and figure 2), it is defined in 2 ways. Both definitions are relative definitions with respect to it's mean power measured.

 

1) Std(signal)/mean(signal)

2) (max(signal)-min(signal))/mean(signal)

 

The numbers of both definitions are listened in the table above and figure 5. The relative noise (definition 1) is from above 100[Watt] somewhere constant, about 5-7% (see also the second subplot of figure 5). In definition 2 the same data the noise is from above the 100[Watt] about 20-30% (see also the third subplot of figure 5). Note: my tacx system was not calibrated in this experiment. 

 

Interesting is the distribution of the noise of the powertap (means the spread in values in the green lines in figure 1 and figure 2). What is it's typical distrubution curve? Figure 6 shows 7 different histograms, 1 histogram for every power setpoint (50, 100, .. , 350W). The distribution in figure 6 is not exactly (an approach of) a Gaussian (or -normal-) distribution as one can seen. The vertical red line is the setpoint, the vertical green line is the average of the histogram.

 

  figure 3: case I

 

  figure 4: case I (statistics) 

 

 

From the figure 3 and figure 4 above, the following numbers can be extracted listened in table 1

 

   figure 5: case I statistics (see table 1)

 

figure 6, histogram of the noise at a certain power level

 

Purpose of Phase 3 (case II)
Compare the measured power - between the Tacx and the PowerTap with a fixed set point on the power, a fixed bike gear, but a vary-ing cadance between 60-100[rpm] (= case II)

Results of Phase 3 (case II)
There seems to be some trending, the higher the cadence, the higher the PowerTap measurement.  In the measurement at 150 [Watt], this trend is clear visible.  In the measurement at 200 [Watt] this trend is not seen very clear.
figure 7: case II

 

figure 8: case II

 

Purpose of Phase 4 
Compare the measured power - between the Tacx and the PowerTap, all fixed, cadance, gear and power setpoint.

Results of Phase 4
The trend of the measured power is stable (enough!) As hoped and expected in this case.
figure 9: everything constant

 

Purpose of Phase 5 
Compare the measured power - between the Tacx and PowerTap with a fixed set point on the power, a variable bike gear and a fixed bicycle cadence (= case III).

 

Results of Phase 5
For a relatively large number of gears the measured power remains almost constant.  At very small bike gears, the power decreased.  This is not a very good result, when for example the -Keutenberg or better the Mortirollo- appears!
figure 10: case III

 

Purpose of Phase 6 
Compare the measured power - between Tacx PowerTap with some random and sudden accelerations.  Gear and Cadence are not of interest.

 

Results of Phase 6
There is some correlation between the Tacx and PowerTap measure.  But I have no faith in the Tacx results.  This means I can not explain about the PowerTap in this test.

 figure 11: random accelerations

 

I hope to contribute to anyone with this measurement and analysis of the results who wants to learn more about the powertap measurement w.r.t. the Tacx.  Both The PowerTap and the Tacx is not "holy" in the results (take your own opinion in this), the data show only the delta between the two in relation to different dimensions.

 

 Best regards,
 Marc

 

 =================================================================

 

 APPENDIX TABLE 1: (case I)

 

+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
| nr |P sp |P tacx |P pt  |pt min|ptmax|pt std|pt range|  |pt/tacx |std pt/pt|R pt/pt|
|    |Watt |Watt   |Watt  |Watt  |Watt |Watt  |Watt    |  |   -    |   -     |       |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|  1 |  50 |  52   |  55  |  48  | 88  |  9   | 40     |  | 106    | 15.9    | 73.2  |
|  2 |  50 |  54   |  52  |  42  | 83  |  9   | 41     |  |  97    | 17.5    | 78.8  |
|  3 |  50 |  52   |  45  |  35  | 57  |  5   | 22     |  |  85    | 11.0    | 49.4  |
|  4 |  50 |  53   |  47  |  35  | 55  |  6   | 20     |  |  89    | 12.4    | 42.5  |
|  5 |  50 |  52   |  47  |  41  | 56  |  5   | 15     |  |  90    | 10.0    | 32.0  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|AVG |  50 |  52   |  49  |  40  | 68  |  7   | 28     |  |  94    | 13.4    | 55.2  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+

+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
| nr |P sp |P tacx |P pt  |pt min|ptmax|pt std|pt range|  |pt/tacx |std pt/pt|R pt/pt|
|    |Watt |Watt   |Watt  |Watt  |Watt |Watt  |Watt    |  |   -    |   -     |       |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|  1 | 100 |  99   |  99  |  87  |149  | 13   | 62     |  | 100    | 13.3    | 62.4  |
|  2 | 100 | 100   |  93  |  88  |100  |  3   | 12     |  |  94    |  3.2    | 12.8  |
|  3 | 100 |  99   |  92  |  83  |101  |  4   | 18     |  |  92    |  4.6    | 19.6  |
|  4 | 100 |  99   |  88  |  82  | 93  |  3   | 11     |  |  89    |  3.9    | 12.5  |
|  5 | 100 |  99   |  89  |  79  | 96  |  4   | 17     |  |  89    |  4.2    | 19.2  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|AVG | 100 |  99   |  92  |  84  |108  |  6   | 24     |  |  93    |  5.9    | 25.3  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+

+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
| nr |P sp |P tacx |P pt  |pt min|ptmax|pt std|pt range|  |pt/tacx |std pt/pt|R pt/pt|
|    |Watt |Watt   |Watt  |Watt  |Watt |Watt  |Watt    |  |   -    |   -     |       |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|  1 | 150 | 150   | 142  | 128  |151  |  7   | 23     |  |  95    |  5.2    | 16.2  |
|  2 | 150 | 149   | 138  | 127  |149  |  5   | 22     |  |  93    |  3.7    | 15.9  |
|  3 | 150 | 149   | 135  | 126  |146  |  6   | 20     |  |  91    |  4.1    | 14.8  |
|  4 | 150 | 150   | 132  | 127  |138  |  3   | 11     |  |  88    |  2.5    |  8.3  |
|  5 | 150 | 150   | 133  | 127  |141  |  4   | 14     |  |  89    |  2.7    | 10.5  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|AVG | 150 | 149   | 136  | 127  |145  |  5   | 18     |  |  91    |  3.6    | 13.1  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+

+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
| nr |P sp |P tacx |P pt  |pt min|ptmax|pt std|pt range|  |pt/tacx |std pt/pt|R pt/pt|
|    |Watt |Watt   |Watt  |Watt  |Watt |Watt  |Watt    |  |   -    |   -     |       |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|  1 | 200 | 200   | 188  | 175  |217  | 11   | 42     |  |  94    |  6.0    | 22.3  |
|  2 | 200 | 199   | 186  | 172  |208  | 10   | 36     |  |  93    |  5.5    | 19.3  |
|  3 | 200 | 199   | 184  | 168  |194  |  9   | 26     |  |  92    |  4.7    | 14.2  |
|  4 | 200 | 199   | 179  | 166  |192  |  8   | 26     |  |  90    |  4.5    | 14.6  |
|  5 | 200 | 200   | 179  | 169  |190  |  6   | 21     |  |  89    |  3.5    | 11.7  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|AVG | 200 | 199   | 183  | 170  |200  |  9   | 30     |  |  92    |  4.9    | 16.4  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+

+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
| nr |P sp |P tacx |P pt  |pt min|ptmax|pt std|pt range|  |pt/tacx |std pt/pt|R pt/pt|
|    |Watt |Watt   |Watt  |Watt  |Watt |Watt  |Watt    |  |   -    |   -     |       |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|  1 | 250 | 249   | 237  | 213  |257  | 13   | 44     |  |  95    |  5.6    | 18.5  |
|  2 | 250 | 249   | 234  | 212  |285  | 18   | 73     |  |  94    |  7.6    | 31.2  |
|  3 | 250 | 250   | 228  | 210  |246  | 12   | 36     |  |  91    |  5.0    | 15.8  |
|  4 | 250 | 250   | 226  | 210  |244  | 13   | 34     |  |  90    |  5.6    | 15.1  |
|  5 | 250 | 249   | 229  | 204  |244  | 13   | 40     |  |  92    |  5.5    | 17.4  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|AVG | 250 | 249   | 231  | 210  |255  | 14   | 45     |  |  93    |  5.9    | 19.6  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+

+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
| nr |P sp |P tacx |P pt  |pt min|ptmax|pt std|pt range|  |pt/tacx |std pt/pt|R pt/pt|
|    |Watt |Watt   |Watt  |Watt  |Watt |Watt  |Watt    |  |   -    |   -     |       |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|  1 | 300 | 301   | 285  | 256  |330  | 22   | 74     |  |  95    |  7.7    | 26.0  |
|  2 | 300 | 301   | 278  | 254  |328  | 19   | 74     |  |  92    |  7.0    | 26.6  |
|  3 | 300 | 299   | 272  | 245  |298  | 18   | 53     |  |  91    |  6.4    | 19.5  |
|  4 | 300 | 300   | 276  | 250  |298  | 17   | 48     |  |  92    |  6.2    | 17.4  |
|  5 | 300 | 299   | 278  | 246  |302  | 17   | 56     |  |  93    |  6.2    | 20.2  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|AVG | 300 | 300   | 278  | 250  |311  | 19   | 61     |  |  93    |  6.7    | 21.9  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+

+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
| nr |P sp |P tacx |P pt  |pt min|ptmax|pt std|pt range|  |pt/tacx |std pt/pt|R pt/pt|
|    |Watt |Watt   |Watt  |Watt  |Watt |Watt  |Watt    |  |   -    |   -     |       |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|  1 | 350 | 345   | 325  | 168  |365  | 43   |197     |  |  94    | 13.3    | 60.7  |
|  2 | 350 | 348   | 331  | 304  |362  | 19   | 58     |  |  95    |  5.9    | 17.5  |
|  3 | 350 | 348   | 319  | 257  |358  | 24   |101     |  |  92    |  7.6    | 31.5  |
|  4 | 350 | 346   | 328  | 301  |353  | 17   | 52     |  |  95    |  5.1    | 15.8  |
|  5 | 350 | 352   | 322  | 295  |354  | 20   | 59     |  |  91    |  6.3    | 18.3  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+
|AVG | 350 | 348   | 325  | 265  |358  | 25   | 93     |  |  93    |  7.6    | 28.8  |
+----+-----+-------+------+------+-----+------+--------+  +--------+---------+-------+

01.12.2009 14:19

henkV

In hoeverre heb je de calibratie van de tacx betrokken? Heb je afgesteld in een poging op 0 uit te komen of heb je er verder niets aan ingesteld?

01.12.2009 14:23

Dutchy Marc

Mijn tacx is niet gecalibreerd tijdens dit experiment! Aangezien ik een systematische delta zie van 7% kan dit de delta kleiner maken. Wat ik niet weet is of de calibratie van de tacx een -offset- calibratie of een -gain- calibratie of beide is? In geval van -gain- calibratie zou het verschil van 7% kleiner kunnen uitvallen?

01.12.2009 14:36

henkV

Die calibratie kan je zowel helpen, als tegen werken. Ik ben alleen niet in staat daar een zelfde waarde te reproduceren. Afhankelijk van de warmte van het apparaat (= gok) varieert eea. Maar elke keer dat ik calibreer, bv om de 5 minuten zie ik een andere calibratie waarde.

Je zou het nog een keer moeten meten nadat je 10 min hebt warm gereden, dan de calibratie, en dan een vergelijk. Op zich vind ik die 7% niet schokkend, zelfs oninteressant. Wat ik veel interessanter vind is om te weten of het verschil steevast in dezelfde orde van grootte zit. Want wat we excact aan watts produceren is niet zo belangrijk, maar wel weer de constatering of er een toe- of afname is van het geleverde vermogen. Toch?

01.12.2009 14:44

Dutchy Marc

Ik ben het helemaal eens met je. Als de 7% "fout" (beter gezegd, "delta") tussen powertap en tacx systematisch is over de hele range van vermogens is er niks aan het handje. Ik heb zojuist een extra tabel toegevoegd aan dit artikel en dan zie je dat de powertap over ALLE vermogens rond de 93% van de tacx zit. prima! Zeker prima voor een analyze achteraf, eenmaal thuis. Komende zomer zal ik eens een steile helling langzaam beklimmen en aan de hand van de gemeten hoogtemeters/tijd en mijn totale gewicht de waarden van de powertap nog eens valideren. Ik ben benieuwd dan?

Wel zie je de powertap flink ruizen in vermogen (zie de groene lijn figuur 1 en figuur 2). Dat betekent dat ik een 'getrained' oog moet zien te ontwikkelen voor de waarden op de display, hoe juist te interpreteren? Het goede nieuws in deze is wel dat als ik straks met mijn Garmin compatible ben er extra middelingen mogelijk zijn over de tijd, waardoor de ruisgevoeligheid op het display flink afneemt. Maar nog zal er tijdens het fietsen goed gekeken moeten worden (enkele keren) naar het display voor juiste interpretatie ...

01.12.2009 20:35

litespeed

Heel erg interessant, Marc. Mooi werk. Ik heb nog geen tijd gevonden/genomen om het goed te lezen. In ieder geval bevalt die laatste conclusie, dat de Tacx een min of meer constante afwijking heeft, mij prima. Lachen Dank.

Add a reply... 

Web link to this entry: