Ergomo Pro or Power Tap SL 2.4?

[peterwright]

If both devices are calibrated properly, I WOULDN'T expect the AP for a TT to be the same for both devices. I would expect the PT to record a lower AP. If they are identical, then one or the other device is "off".

My point exactly - At factory calibrated K Factor, the two do read very closely at FTP and greater - to within a few %age points -so FTP would be very close when tested. We then have a situation where the Ergomo reads ~3% high @ FTP and ~12-14% at lower powers. Whatever we attribute it to this does not lessen the fact that the Ergomo is seriously compromised as a training tool as a result of this non linearity.

People who run PTs with SRMs do observe the drivetrain losses....however, I haven't seen any other attempts to quantify it over a wide range of power levels and with a look at old vs. new drivetrains.

I am not referrring to DT losses but to non linearity - this "quirk" seems to be Ergomo specific. In fact we have run compariosons with PT & SRM and found they track very well across all power bands.

 

[Ergoman]

Isn't the purpose of using a powermeter to provide a reliable benchmark against which training can be planned and conducted and against which progress can be gauged? Assuming that a powermeter shows realistic and consistent readings is it really important for it to be in total agreement with any other power meter? Once an individual develops a database of power info with one type of power meter and uses this power info for training does it really matter if another power meter might give slightly different results?

It appears to me that power meter users spend an inordinate amount of time comparing one meter to another, calibrating, testing, and in general obsessing about absolute accuracy, when, in fact, absolute accuracy is unobtainable and not even that important. The things that are important are cost, consistency, reliability, and functionality.

The reason I list temperature sensitivity of my Ergomo installation as one of the major cons is that temperature sensitivity is the one main thing that contributes to inconsistency. I don't care if my powermeter shows my FTP at 280 watts or 320 watts, so long as it shows the same results for the same effort every time and so long as it functions reliably.

 

[peterwright]

Isn't the purpose of using a powermeter to provide a reliable benchmark against which training can be planned and conducted and against which progress can be gauged? Assuming that a powermeter shows realistic and consistent readings is it really important for it to be in total agreement with any other power meter? Once an individual develops a database of power info with one type of power meter and uses this power info for training does it really matter if another power meter might give slightly different results?

It appears to me that power meter users spend an inordinate amount of time comparing one meter to another, calibrating, testing, and in general obsessing about absolute accuracy, when, in fact, absolute accuracy is unobtainable and not even that important. The things that are important are cost, consistency, reliability, and functionality.

The reason I list temperature sensitivity of my Ergomo installation as one of the major cons is that temperature sensitivity is the one main thing that contributes to inconsistency. I don't care if my powermeter shows my FTP at 280 watts or 320 watts, so long as it shows the same results for the same effort every time and so long as it functions reliably.

Ergoman - to a certain degree you are correct. From a training perspective, repeatability is paramount.

There are a few reasons why this is not enough IMO;

1. If I pay this sort of money for a PM then should I not expect it to be accurate and able to be benchmarked ?

2. I have one client who, after a reinstall, has seen the power on his Ergomo drop by 60w

3. This same client rides with a team mate that has an Ergomo - his reads 50w higher...

4. As a coach, I do a lot of inter rider comparison - many of my clients ride the same races and as such an accurate reading is critical

5. Again, from a coaching perspective, the comparison of various power/duration figures is essential in developing a rider.

If this was an Ibike we were talking about then I would maybe live with these issues but a device costing this much... leg imbalance, temp sensitivity, installation issues, non linearity, inability to benchmark against calibrated devices -are these really acceptable compromises ?

 

[Pelotonium]

I'm not sure why non-linearity is a show stopper. Aren't the Ergomo manual k-factor calibrations designed to linearise the power output?

As long as the relationship between "power-at-the-crank" and "power-at-the-hub" is monotonic, it should be fairly easy to map one to the other, shouldn't it?

And I would certainly be shocked if the relationship wasn't monotonic.

 

[peterwright]

I'm not sure why non-linearity is a show stopper. Aren't the Ergomo manual k-factor calibrations designed to linearise the power output?

As long as the relationship between "power-at-the-crank" and "power-at-the-hub" is monotonic, it should be fairly easy to map one to the other, shouldn't it?

And I would certainly be shocked if the relationship wasn't monotonic.

Ergomo have said that altering K factor has an affect on linearity (see earlier post)

I have certainly seen this behaviour - I have been unable to achieve any linearity by altering K factor over the last 1000kms

 

[Ergoman]

Ergoman - to a certain degree you are correct. From a training perspective, repeatability is paramount.

There are a few reasons why this is not enough IMO;

1. If I pay this sort of money for a PM then should I not expect it to be accurate and able to be benchmarked ?

2. I have one client who, after a reinstall, has seen the power on his Ergomo drop by 60w

3. This same client rides with a team mate that has an Ergomo - his reads 50w higher...

4. As a coach, I do a lot of inter rider comparison - many of my clients ride the same races and as such an accurate reading is critical

5. Again, from a coaching perspective, the comparison of various power/duration figures is essential in developing a rider.

I can certainly understand how, from a coaching perspective, life would be easier if all your clients' power meters matched precisely; however, I'd guess that in the real world that's never going to happen. Further, from an individual standpoint, I wouldn't be inclined to pay additional bucks (to buy a more expensive power meter or to swap) just to simplify my coach's analytical tasks.

In any case, it's my personal experience that if an Ergomo is properly installed and used it will provide consistent and reliable data for training. I'm self-coached, so keeping a coach happy is not my concern. Finally, if I want a head to head comparison of my real, effective power vs. another rider I just look at my finishing times and/or position. No one ever changed race results by presenting power data that showed they output more than the winner.

 

[Pelotonium]

I can certainly understand how, from a coaching perspective, life would be easier if all your clients' power meters matched precisely; however, I'd guess that in the real world that's never going to happen.

Well, the whole idea of measuring power - or anything, for that matter - is to get it right. There is nothing subjective about measuring power. Either you measure at the bb, the crank, or the rear hub. Each of these measurement points should be single-valued.

Finally, if I want a head to head comparison of my real, effective power vs. another rider I just look at my finishing times and/or position. No one ever changed race results by presenting power data that showed they output more than the winner.

The idea is also to be able to compare set-ups for hills, TTs, etc... Understanding how the power is spent is a welcome addition to the set of training tools you can use. You may finish first in a hill race because you're lighter -- not because you put out more power. You might get smoked on a TT and wonder if you had a bad day. Measuring power correctly will demystify all this.

Successful racing is usually the end-result of a properly-planned, properly-measured training regimen.

 

[peterwright]

I can certainly understand how, from a coaching perspective, life would be easier if all your clients' power meters matched precisely; however, I'd guess that in the real world that's never going to happen. Further, from an individual standpoint, I wouldn't be inclined to pay additional bucks (to buy a more expensive power meter or to swap) just to simplify my coach's analytical tasks.

In any case, it's my personal experience that if an Ergomo is properly installed and used it will provide consistent and reliable data for training. I'm self-coached, so keeping a coach happy is not my concern. Finally, if I want a head to head comparison of my real, effective power vs. another rider I just look at my finishing times and/or position. No one ever changed race results by presenting power data that showed they output more than the winner.

Actually it does happen - any clients with an SRM or PT can calibrate and achieve a decent level of accuracy and repeatability. In the case of a PT they don't pay additional bucks but in fact they save some.

It is not about simplifying the coaches job - it is about making it possible. Maybe you have never wandered what power the guy who won the race is putting out up that selective climb, and how it compares to yours - but I can assure you that this sort of information is of great interest to most of the athletes we coach.

You just look at your finishing position - then what ? Do you not think it would be more effective to know where you were weaker than the winner in actual terms rather than just by feel ? Race results get changed by effective training, and effective training results from accurate data and targeting of weak areas.

Let me ask you another question - what if your Ergomo changes it's power readings as I have seen on a few occasions ?

What if you change Ergomo's or "horror" decide to change to another PM ?

Whatever you might feel - there are plenty of reasons that somebody training and/or racing with a PM would want comparitive and accurate numbers and in a device costing this much they should expect to get it.

 

[acoggan]

Could you define "a slope not significantly different from unity" a little better?

Ask, and ye shall receive...

Here are the raw data from the first Velodyne vs. SRM comparison I did back in 1997 (which Bruce Sargeant posted on the web back in 1998):

Velodyne power (W) SRM power (W)
50 54
100 107
150 159
200 207
250 257
300 305
350 355

slope = 0.9979 (95% CL = 0.9798, 1.0158)
intercept = 6.7 (95% CL = 2.7, 10.7)
R^2 = 0.9998

So, as you might expect, power measured at the crank by the SRM was higher than power measured (set) at the rear wheel by the Velodyne, with the difference being reflected by a higher intercept, with a slope not different from unity.

Here are the raw data from the first Velodyne vs. PowerTap comparison that I did (which I posted to the web back in 1999):

Velodyne (W) PowerTap (W)
50 51
100 101
150 150
200 197
250 245
300 292
350 344
400 396

slope = 0.9771 (95% CL = 0.9617, 0.9926)
intercept = 2.1 (95% CL = -1.8, 6.0)
R^2 = 0.9998

This time, with power measured at the rear wheel using the PowerTap, the intercept barely achieved significance (P=0.03), with the difference in the slope presumably being due to the "true" difference between the devices.

Skipping over the numerous other formal cross-comparisons that I've done, here are the data from the most recent head-to-head comparison that I've done between the PowerTap and SRM (last spring):

PowerTap (W) SRM (W)
54 58
110 114
157 163
211 214
257 265
308 315

Slope = 1.0133 (95% CL = 0.9910, 1.0356)
Intercept = 2.9 (95% CL = -1.6, 7.4)
R^2 = 0.9998

This time, neither the slope nor the intercept were significantly different from one and zero, respectively, but obviously the difference is primarily reflected in the intercept, not the slope.

Unfortunately, I can't tell you the exact state of the chain for any of the above, although I can say that 1) I replace my chains yearly or when wear reaches 0.5%, whichever comes first, and 2) all data were collected w/o shifting gears (and thus are not confounded by chainline issues...including the fact that the sensitivity of the PowerTap and SRM vary with chainline).

The only data set that I do have with a chain that I know was quite new (<100 mi use) come from my track crankset (average of three trials on different days during the summer of '04):

Velodyne power (W) SRM power (W)
50 58
100 109
150 158
200 206
250 259
300 311
350 360

Slope = 1.0078 (95% CL = 0.9934, 1.0223)
Intercept = 7.1 (95% CL = 3.9, 10.1)
R^2 = 0.9998

Obviously, it would be better to compare the SRM against the PowerTap, but the latter won't fit a track frame, so the Velodyne is the only option. In any case, these data are very similar to what I got back in 1997 (using a different SRM and a different Velodyne).

The bottom-line conclusion sthat I draw from these data are:

1) all three devices are extremely linear;

2) taking into consideration power loss in the drivetrain, all three are quite accurate; and

3) when data are collected in this manner are subjected to linear regression analysis, power losses in the drivetrain are primarily reflected in a different intercept, and not a different slope.

Now let's turn to the data from the Ergomo Pro that I tested:

10/14/2005

SRM power (W) Ergomo power (W)
61 84
111 132
160 170
210 218
258 262
306 298

10/15/2005

SRM power (W) Ergomo power (W)
61 95
111 143
161 191
210 212
259 257
309 312
258 260
208 229
158 164
109 128
59 89

Clearly, the response of the Ergomo is non-linear (as reported by Peter Wright), with the Ergomo greatly overestimating power at lower power outputs, but underestimating power at higher power outputs (also as reported by Peter). Having observed this, I decided to test whether it was due to an imbalance in terms of how I apply power, and did several tests using only my left leg - representative data are shown below:

SRM power (W) 0.5 x Ergomo power (W)
69 81
108 122
150 166

If you plot these data along with the other numbers above, you'll find that they are quite consistent, i.e., the non-linearity of the Ergomo was apparent regardless of whether I pedaled with one leg or two, thus ruling out a R/L imbalance as the primary cause.

One final comment: as you might guess from the date, all of the Ergomo data above were collected before they modified their installation instructions (in fact, data such as those above are likely what led to their modification). Specifically, while the b.b. was oriented such that the internal circuit board was close to vertical and everything was tightened to specifications, I did not face the b.b. shell or chase the threads. Now that this is a stated requirement, I've been tempted to do some more experiments, but have given away the Ergomo, and don't have the heart to ask for it back (nor do I really have the time these days for such things, as there are bigger fish to fry, so to speak). Moreover, the fact that Peter has obtained similar results despite (presumably) strictly adhering to the new installation instructions leads me to believe that it probably wouldn't make a difference.

Bottom line: I don't think discrepancies between an Ergomo and a PowerTap can be blamed on drivetrain friction (esp. since similar discrepancies have been observed in comparison to an SRM).

 

[acoggan]

What if you change Ergomo's or "horror" decide to change to another PM ?

I think this is one of the most frequently overlooked reasons for insisting on accuracy (and not just precision). The fact of the matter is that is extremely unlikely that anyone will continue to use the exact same powermeter for a long period of time, which means that if you hope to be able to track long-term trends, the data must be correct.

 

[TamMan]

What if you change Ergomo's or "horror" decide to change to another PM ?

Good point - I have a PT and the Ergomo. When I had to re-install my Ergomo earlier this year due to a bike swap, the values came back much lower. I ended up having to reset my FTP by ~40w. My PT reads about ~30w higher.

One of the things that I am actually really concerned about is that a change in the offset has a dramatic effect on the power readings. I have achieved excellent consistency between the PT and the Ergomo - but only at the same offset. I haven't done enough testing to find out how temparature sensitive my specific Ergomo is.

 

[peterwright]

Good point - I have a PT and the Ergomo. When I had to re-install my Ergomo earlier this year due to a bike swap, the values came back much lower. I ended up having to reset my FTP by ~40w. My PT reads about ~30w higher.

Very strange - that is exactly what has just happened to a client of ours. A simple and careful reinstall has led to FTP dropping from 330w to 270w. He has since reinstalled again and no difference.

 

[vadiver]

I can certainly understand how, from a coaching perspective, life would be easier if all your clients' power meters matched precisely; however, I'd guess that in the real world that's never going to happen. Further, from an individual standpoint, I wouldn't be inclined to pay additional bucks (to buy a more expensive power meter or to swap) just to simplify my coach's analytical tasks.

In any case, it's my personal experience that if an Ergomo is properly installed and used it will provide consistent and reliable data for training. I'm self-coached, so keeping a coach happy is not my concern. Finally, if I want a head to head comparison of my real, effective power vs. another rider I just look at my finishing times and/or position. No one ever changed race results by presenting power data that showed they output more than the winner.

I think this is more than keeping a coach happy. Even if self coached how do you deal with the 1W per degree F inaccuracy? It is not unusual for my rides to have a 20-30 degree temp swing. Just looking at some of my files I have two out of seven that had 40 degree swings.

Granted I have not been riding with power yet but I would think a 40W difference in power would be significant. Particularly if it was 40W low and I was trying to maintain a set power level. I would blow up. If it was 40W high, I would not be getting the workout I wanted.

If set up is so important, what if you regularly ride two different bikes? What is the likelihood of getting the same power readings on the two bikes?

 

[Ergoman]

I think this is more than keeping a coach happy. Even if self coached how do you deal with the 1W per degree F inaccuracy? It is not unusual for my rides to have a 20-30 degree temp swing. Just looking at some of my files I have two out of seven that had 40 degree swings.

Granted I have not been riding with power yet but I would think a 40W difference in power would be significant. Particularly if it was 40W low and I was trying to maintain a set power level. I would blow up. If it was 40W high, I would not be getting the workout I wanted.

If set up is so important, what if you regularly ride two different bikes? What is the likelihood of getting the same power readings on the two bikes?

You've hit upon the one issue that causes me to have doubts about the Ergomo. Temperature sensitivity or whatever it is that causes changes in offset during a ride is a big concern. To this point, I've been able to manage by carefully checking offsets, but as the weather heats up, that may not be enough. I suspect that my BB shell may not be properly faced and that that may be part of the sensitivity of my particular unit. One day when and/or if things get out of control, I'll have the BB shell re-faced. Then, if there's still a temperature sensitivity problem, you'll see my Ergomo listed on eBay.

 

[peterwright]

Then, if there's still a temperature sensitivity problem, you'll see my Ergomo listed on eBay.

Do I detect a name change coming on

 

[acoggan]

Here are the raw data from the first Velodyne vs. PowerTap comparison that I did (which I posted to the web back in 1999):

Velodyne (W) PowerTap (W)
50 51
100 101
150 150
200 197
250 245
300 292
350 344
400 396

slope = 0.9771 (95% CL = 0.9617, 0.9926)
intercept = 2.1 (95% CL = -1.8, 6.0)
R^2 = 0.9998

This time, with power measured at the rear wheel using the PowerTap, the intercept barely achieved significance (P=0.03), with the difference in the slope presumably being due to the "true" difference between the devices.

EDIT: I wrote the last sentence above before discovering that I'd failed to include the 400, 396 data pair in the regression. The slope and intercept are correct, but as should be clear from the confidence limits, it is the slope and not the intercept that is significant.

 

[Tom Anhalt]

Ask, and ye shall receive...

OK...I guess I can't say I haven't seen the data anymore...

Unfortunately, I can't tell you the exact state of the chain for any of the above, although I can say that 1) I replace my chains yearly or when wear reaches 0.5%, whichever comes first, and 2) all data were collected w/o shifting gears (and thus are not confounded by chainline issues...including the fact that the sensitivity of the PowerTap and SRM vary with chainline).

Interesting...I typically change my chains every 6 months or 0.5%, whichever comes first. BTW, the "old" chain in the results I listed above had less than 6 months of use and did not have any detectable "stretch" (i.e. it was much less than 0.5%).

So...all of your testing listed was done on the Velodyne in a single gear? Was it the same gear for each test? Knowing that the PT has chainline, or "gear to gear" sensitivities, doesn't that call into question the applicability of the assumptions on correlation you are describing?

On the SRM, are you saying that chainline effects the static calibration numbers, or are you determining that in some other way?...just curious.

Bottom line: I don't think discrepancies between an Ergomo and a PowerTap can be blamed on drivetrain friction (esp. since similar discrepancies have been observed in comparison to an SRM).

I concur with the above with the addition of the words "all of the" in between the words "think" and "discrepancies".

 

[acoggan]

So...all of your testing listed was done on the Velodyne in a single gear? Was it the same gear for each test?

Not necessarily, in that I used whatever cassette happened to be on my bike at the time. IOW, while I was always in the 4th cog in (and the big ring), that cog could have been either a 14 or a 15.

Knowing that the PT has chainline, or "gear to gear" sensitivities, doesn't that call into question the applicability of the assumptions on correlation you are describing?

I'm not sure I follow...are you saying that you can't, e.g., use these results to correct data collected in the field (unless all of it happened to be collected using the same gear)? If so, I only halfway agree with you, i.e., I view that objection as more of a theoretical one (since the effects of chainline on SRM or PowerTap accuracy are very small indeed...unlike, e.g., with your Polar).

On the SRM, are you saying that chainline effects the static calibration numbers

Yes, that's what I have observed. Again, though, the difference is quite small (smaller than the differences you see based on crank position).

 

[Tom Anhalt]

I'm not sure I follow...are you saying that you can't, e.g., use these results to correct data collected in the field (unless all of it happened to be collected using the same gear)? If so, I only halfway agree with you, i.e., I view that objection as more of a theoretical one (since the effects of chainline on SRM or PowerTap accuracy are very small indeed...unlike, e.g., with your Polar).

What I'm saying is that even with moderate power levels (i.e. 200W @75 rpm) drivetrain efficiencies can vary from 91-96% depending on the chosen gear. This is shown (based on a Shimano MTB groupo) in fig. 11 and table 2 of Kyle and Berto's report in the IHPVA journal:

http://www.ihpva.org/HParchive/PDF/hp52-2001.pdf

So...isn't that going to affect your linear fits of "device vs. device?"...especially if you're analyzing a file that uses a wide range of gear choices?

BTW, aside from "on the trainer" data (which is a known bugaboo for the Polar system...but not necessarily for reasons of chainline) have you seen anything with "on the road" data that shows there to be a "chainline" type variation with the Polar? I'm only asking because I'd be interested in seeing that.

 

[Ergoman]

Do I detect a name change coming on

Not quite yet.

Here are results I got yesterday comparing my Ergomo to a Kurt Kinetic Road Machine calibrated as described here: http://www.planet-x-bikes.com/triat...pagemaster&PAGE_user_op=view_page&PAGE_id=189

K=factor on the Ergomo was set to factory spec, and offset was checked and saved after about a 15 minute warmup at room temperature. Ergomo power readings were averaged over a period of time ranging from a minute to about 20 seconds at the higher readings (yesterday was supposed to be a rest day). I didn't mess with power less than 125 watts, because I don't do much training at that level other than recovery. I didn't go over 550 watts because I can't sustain that level of power for any meaningful period of time. Kurt Kinetic power was calculated based on Ergomo speed and the formula provided on the Kurt Kinetic website. My Ergomo speedometer was pretty carefully calibrated earlier this winter over a 10 mile outdoor measured course.

Ergomo power/Kurt Kinetic calculated power
125/126
142/143
162/162
186/183
211/206
234/231
264/265
291/295
313/318
360/364
400/394
435/430
531/539

To be completely honest, I'm stunned at how close the Ergomo compared to the Kurt Kinetic in this test. It's so close that it appears that the data is fudged, but I assure you it's not. If I were going to fudge the data I'd throw in a little disparity for realism. Moreover, I wrote the Ergomo power numbers and speed down before I calculated what the KK power should be, so I wasn't tempted to nudge the readings one way or another.

I realize that a fluid trainer is nowhere as sophisticated as a Velodyne and that the actual calculated power is subject to a lot of variables and is probably far less accurate than a Velodyne. BUT what this test does demonstrate is that the power curve of a properly installed and set up Ergomo corresponds very nicely to the predicted power of a very commonly used and well accepted training device.

Finally, day before yesterday I did a little outdoor ride on a day when the temperature was a steady 50 degrees. Since my bike came from an indoor temperature of 70 degrees I was able to confirm that offset changes about one unit for 5 degrees of temperature change. On the trainer yesterday I was also able to confirm that one unit of offset error is equal to about 5 watts of power error. My conclusion being that for my specific Ergomo installation, one degree of temperature change without a corresponding offset change equals one watt of error in power readings. If I went for a two hour ride and during the course of the ride the temperature changed ten degrees, the maximum power error I would have would be 10 watts, and the average would be about 5. I train at about 250 watts, so the average error would be about 2%.

To put this error in perspective, the Ergomo is specified to be accurate to about 1.5%, the PT is also spec'd at 1.5%, while the SRM Amateur is spec'd at 5%, and the pro to 2%. All of these devices are also temperature sensitive.