Quarq Exogram Power Meter

Listed on this site as $1995 for GXP and $2045 for BB30. I guess that is a bit much for me.

Felt_Rider said:

Listed on this site as $1995 for GXP and $2045 for BB30. I guess that is a bit much for me.

It'll be interesting to see where the real world pricing ends up.

Quoted post said:

Originally Posted by alienator [IMG]/img/forum/go_quote.gif[/IMG]

To the training expert: Alex, your thoughts? How do you rate Quarq versus the other PM's?

As a coach, I am comfortable with the data from my clients using SRM, Quarq or Powertap. Of course we apply common sense to the data to pick up on any issues that can sometimes happen (e.g. someone's not properly setting torque zero, or the slope needs checking).

Quoted post said:

Originally Posted by alienator [IMG]/img/forum/go_quote.gif[/IMG]

To give context, at this the power meter field is:

• SRM
  
• Quarq/SRAM
  
• Power2Max (not yet in the US)
  
• PowerTap
  
• Look/Polar
  
• Garmin Vector
  
• Brim Brothers (????)
  
• Ergomo (on life support and in need of euthanizing)
  
• iBike (not achieving its original promise)
  

Some slides from my recent power meter seminar:

[IMG]cyclingforums.com263Open ↗[/IMG]

[IMG]cyclingforums.com263Open ↗[/IMG]
[IMG]cyclingforums.com263Open ↗[/IMG]
[IMG]cyclingforums.com263Open ↗[/IMG]

Quoted post said:

Originally Posted by Alex Simmons [IMG]/img/forum/go_quote.gif[/IMG]

As a coach, I am comfortable with the data from my clients using SRM, Quarq or Powertap.

Have you ever encountered the abnormally high torque (and hence power) values the Quarq I had use of occasionally threw out?

trainingandracingwithapowermeter.comprediction of muscle fiber type from.htmlOpen ↗

Personally, I would never buy/recommend a Quarq until 1) that mystery was solved, and 2) they came up with an easier way to A) disable the auto-zero function and B) change the slope.

Quoted post said:

Originally Posted by acoggan [IMG]/img/forum/go_quote.gif[/IMG]

Have you ever encountered the abnormally high torque (and hence power) values the Quarq I had use of occasionally threw out?

trainingandracingwithapowermeter.comprediction of muscle fiber type from.htmlOpen ↗

Personally, I would never buy/recommend a Quarq until 1) that mystery was solved, and 2) they came up with an easier way to A) disable the auto-zero function and B) change the slope.

I've never seen torque spikes, BUT I've never gone searching for them so it's possible they are there and I never noticed (though they would surely show up in the VE graphs wouldn't they?). I might have a look when I get home.

Changing the slope on the Quarq couldn't really be any easier, but by "easy" you probably meant a method that doesn't require an iPhone. What's REALLY nice is that despite testing both my Quarq's on multiple occasions and the odd chainring swap (same brand slightly different size), I've never needed to change the slope. It just works.

Quoted post said:

Originally Posted by lanierb [IMG]/img/forum/go_quote.gif[/IMG]

I've never seen torque spikes, BUT I've never gone searching for them so it's possible they are there and I never noticed (though they would surely show up in the VE graphs wouldn't they?). I might have a look when I get home.

Changing the slope on the Quarq couldn't really be any easier, but by "easy" you probably meant a method that doesn't require an iPhone. What's REALLY nice is that despite testing both my Quarq's on multiple occasions and the odd chainring swap (same brand slightly different size), I've never needed to change the slope. It just works.

1. VE wouldn't necessarily show the artifacts I observed. That is, while the power was abnormally high given the recorded cadence, it wasn't high in absolute terms, such that a second-by-second VE calculation would at most have shown only a tiny upward blip.

2. If changing the calibration of a powermeter requires buying/borrowing additional equipment, it ain't "easy", at least not as easy as pushing some buttons on the handlebar computer.

Quoted post said:

Originally Posted by acoggan [IMG]/img/forum/go_quote.gif[/IMG]

Have you ever encountered the abnormally high torque (and hence power) values the Quarq I had use of occasionally threw out?

trainingandracingwithapowermeter.comprediction of muscle fiber type from.htmlOpen ↗

Personally, I would never buy/recommend a Quarq until 1) that mystery was solved, and 2) they came up with an easier way to A) disable the auto-zero function and B) change the slope.

FWIW, I've had a similar issue (occasionally) with my SRM. Sometimes after coasting, there will be crazy high cadence and power reading for a few seconds. It used to be worse before having the dual reed switch installed.

Quoted post said:

Originally Posted by acoggan [IMG]/img/forum/go_quote.gif[/IMG]

Have you ever encountered the abnormally high torque (and hence power) values the Quarq I had use of occasionally threw out?

trainingandracingwithapowermeter.comprediction of muscle fiber type from.htmlOpen ↗

Personally, I would never buy/recommend a Quarq until 1) that mystery was solved, and 2) they came up with an easier way to A) disable the auto-zero function and B) change the slope.

I can't say it's jumped out at me, but then in most cases maximal F-V testing is not a priority for my Quarq using clients due to the nature of their target events (compared to the way it might be more of a priority for a track racer, and they all use SRMs anyway). I don't see any unusual mean maximal power curves though, although that is not evidence of absence.

As for the slope thing, yeah, same applies to Powertap, can't change that at all. I think at minimum if you can at least validate the slope then that's a good start. As for auto zero, it's my understanding it's not really "auto" in that it still requires the user to perform a task in order to zero. That may create other issues of course.

I had an online chat with a local user of a P2M unit, who tested the variance of torque zero when clipped in and unclipped and he did notice that despite trying to not apply pressure it did sometimes give a different reading of a unit or 2, which in case of a P2M would seem a significant amount. The worry with that unit (aside from the torque zero drift issues) is one does not know when a zero is happening (and hence if it's happening under valid circumstances).

Quoted post said:

Originally Posted by acoggan [IMG]/img/forum/go_quote.gif[/IMG]
1. VE wouldn't necessarily show the artifacts I observed. That is, while the power was abnormally high given the recorded cadence, it wasn't high in absolute terms, such that a second-by-second VE calculation would at most have shown only a tiny upward blip.

Am I doing this right? I just took a look through about 25 Quarq files, including some sprint workouts and crits and races, and I looked at the quadrant analysis charts and looked for high AEPF values. The highest value I could find was about 700N. (I weight about 175 lbs.) Most of my files top out at around 5-600N.

Edit: Looking through a few more files and finally found one that looks totally different and has a few spikes at 1000-1100N and....... turns out the file is from my Powertap not my Quarq. I'm guessing it's most likely an artifact from the virtual cadence though and not a real torque spike.

Quoted post said:

Originally Posted by dkrenik [IMG]/img/forum/go_quote.gif[/IMG]

FWIW, I've had a similar issue (occasionally) with my SRM. Sometimes after coasting, there will be crazy high cadence and power reading for a few seconds. It used to be worse before having the dual reed switch installed.

That's not the same issue. The Quarq I had was grossly overestimating torque, and hence power, when resuming pedaling - cadence was reported correctly. When an SRM double-counts (or more rarely, triple-counts) a pedal revolution, cadence and power are doubled (or tripled), but the torque is correct.

Quoted post said:

Originally Posted by Alex Simmons [IMG]/img/forum/go_quote.gif[/IMG]

I can't say it's jumped out at me, but then in most cases maximal F-V testing is not a priority for my Quarq using clients due to the nature of their target events (compared to the way it might be more of a priority for a track racer, and they all use SRMs anyway). I don't see any unusual mean maximal power curves though, although that is not evidence of absence.

As for the slope thing, yeah, same applies to Powertap, can't change that at all. I think at minimum if you can at least validate the slope then that's a good start. As for auto zero, it's my understanding it's not really "auto" in that it still requires the user to perform a task in order to zero. That may create other issues of course.

Thanks. It could very well have just been the demo unit Quarq lent me, but they claimed to have given it the once-over before sending it out to be sure it was functioning correctly.

(Since mine was from the 1st couple of years of production, I suppose another explanation is that they changed something in the hardware or firmware along the way w/o telling anyone, but we'll presumably never know.)

As for the auto-zero, it kicks in whenever you backpedal "enough" (not 4 revolutions...that's the average the auto-zero is based upon, but they don't have to be consecutive revolutions.)

Quoted post said:

Originally Posted by lanierb [IMG]/img/forum/go_quote.gif[/IMG]

Am I doing this right? I just took a look through about 25 Quarq files, including some sprint workouts and crits and races, and I looked at the quadrant analysis charts and looked for high AEPF values. The highest value I could find was about 700N. (I weight about 175 lbs.) Most of my files top out at around 5-600N.

Sounds normal. So, either your unit doesn't produce the same artifacts, or none of the files you examined had situations where cadence increased from a very low value to a normal value quite quickly.

(It would be interesting to do a F-V test using your crank to see if you get a straight line, as you should, or a curved line, as the Quarq I had produced.)

Quoted post said:

Originally Posted by acoggan [IMG]/img/forum/go_quote.gif[/IMG]

Sounds normal. So, either your unit doesn't produce the same artifacts, or none of the files you examined had situations where cadence increased from a very low value to a normal value quite quickly.

(It would be interesting to do a F-V test using your crank to see if you get a straight line, as you should, or a curved line, as the Quarq I had produced.)

I'm happy to give it a try. The biggest problem is the head unit I think. I normally use a Garmin 500, but that only records a data point once per second. Does anyone happen to know an iPhone app that will record Ant+ data at high resolution? (I have the Wahoo Ant+ receiver.)

Quoted post said:

Originally Posted by lanierb [IMG]/img/forum/go_quote.gif[/IMG]

I'm happy to give it a try. The biggest problem is the head unit I think. I normally use a Garmin 500, but that only records a data point once per second. Does anyone happen to know an iPhone app that will record Ant+ data at high resolution? (I have the Wahoo Ant+ receiver.)

(Sorry, upon review I realize that I misinterpreted what you meant.)

1 s recording is less than ideal, but if combine a few efforts you should be able reconstruct your F-V curve well enough to see if it is linear.

Quoted post said:

Originally Posted by acoggan [IMG]/img/forum/go_quote.gif[/IMG]

That's not the same issue. The Quarq I had was grossly overestimating torque, and hence power, when resuming pedaling - cadence was reported correctly. When an SRM double-counts (or more rarely, triple-counts) a pedal revolution, cadence and power are doubled (or tripled), but the torque is correct.

Yeah, after I re-read your post I realized the same. Sounds like this issue with Quarqs might have been addressed?

Quoted post said:

Originally Posted by dkrenik [IMG]/img/forum/go_quote.gif[/IMG]

Sounds like this issue with Quarqs might have been addressed?

That is what I am attempting to ascertain. It seems, though, that others either haven't looked closely enough at their data to really say, and/or they haven't collected data under the conditions in which I encountered obvious artifacts (i.e., when cadence is rising rapidly from a very low value).

Alex Simmons said:

Some slides from my recent power meter seminar:
 
[IMG]cyclingforums.com263Open ↗[/IMG]

I looked up Laser Spoke. As an optics guy, I can see how Laser Spoke is a possibly useful approach. I helped develop/test a system similar used to measure flatness in things needed to be very flat (with a surface error less than about 10nm/meter), and that system was sensitive enough that under the right conditions it could measure the angular separation of local gravity vectors 1.5m apart. Anyway, the problem that I see with Laser Spoke is that it is open to the air so that dirt and liquids can end up on the prism surface, potentially causing enough laser beam deviation to substantially change the power measurement. It will also have to be calibrated for each different wheel configuration, as well as for when a spoke is changed, and possibly even for changes in tire air pressure.