Road Bike, Cycling Forums banner
1 - 18 of 18 Posts

·
Registered
Joined
·
65 Posts
Discussion Starter · #1 ·
For people who own both a power meter and a magnetic style indoor trainer how closely does your power meter numbers track vs. the speeds you see on your trainer. For example, if at the start of your workout 300 watts is 40 kph do you see the same 40 kph for 300 watts at the end of your workout? Do you see the same watts/speed ratio from one day to the next?

As background:

In his book on training, Graeme Obree goes to great lengths to explain how he uses a simple magnetic style turbo training for training and testing in a way that is similar to the way power meters are sometimes discussed.

For people who have not read the book the setup is as follows:
  • Magnetic style trainer (not fluid or fan based)
  • Flywheel removed from trainer
  • Extreme care used always have the same tire/roller pressure with no changes from week to week.
  • No adjustments to trainer resistance level
  • A good long warm up.

The idea is that when used carefully, the 20 minute time trial type workout on the trainer produces a number expressed as average speed that a measure of athletic performance level. Obree uses this as the core metric to ensure that training is producing a measurable improvement. An improvement in fitness should produce a small increase in the average speed. Obree is not a fan of power meters and in his approach he does not use one.
 

·
Registered
Joined
·
3,935 Posts
Obree uses this as the core metric to ensure that training is producing a measurable improvement. An improvement in fitness should produce a small increase in the average speed. Obree is not a fan of power meters and in his approach he does not use one.
Why is he not a fan of the power meter I wonder, that seems really odd. Speed the way he suggests setting up and using the trainer is just a different approach to assessing power output.
 

·
Registered
Joined
·
201 Posts
I set my trainer up exactly as displayed above, except I left the flywheel on. Reason:
I did not want to buy a $500+ power meter. I have fan positioned same place & temperature in my basement is pretty much same all the time in the winter. Interval workouts at various speeds are done with 15 min warm ups. Works well, except tires have different CRR, so speeds must be recalibrated. Speed changes are linear.
 

·
Registered
Joined
·
2,006 Posts
Obree is something else all the way around. His nuttiness got him to cycling god status, but his training and antics and all are not and should not be something any other person tries to duplicate. He played on an entirely different level than anyone else.

With that said, at a time when SRMs were 2k plus and there were no other options, you did what you had to do.

That is not the case today. It's silly to not buy a powermeter if performance training is your goal. You can spend more on race tires than a used powermeter.
 

·
Pack Fodder.
Joined
·
1,871 Posts
His approach is completely valid, if all of your measured workouts are on a trainer and you limit the variables.

However, power meters, when used with a good analytical software package, provide a lot more data points than just for pacing. As power meter prices drop because of the expanding market, they are a very much a cost-effective training tool. I've bought a couple used wireless DuraAce SRMs that cost me less than $800 each- and that includes having the batteries replaced and calibrated by SRM). If I had've gone wired, I could have found them for $300-$350 easily. Power meters aren't a rich man's plaything anymore.

If all you use it for is an eWang, he's got a solid point. It's an expensive toy and you can use the money better elsewhere.
 

·
Registered
Joined
·
113 Posts
If you're keen on improving your performance like I am then a power meter is essential. Prior to getting my power meter I used virtual power on Trainerroad.com and this works pretty well provided you keep the same bike setup e.g. tyre pressure. However the accuracy is not as good as using a power meter. It's also hard to hit high power targets on short sprint efforts using virtual power. Not so using a power meter.

I'm running a $400 4iiii crank based power meter. I agree with the previous post, power meters aren't a rich man's plaything anymore.

Obree sounds like he's way behind the times and thinks power meters are expensive.
 

·
Cycling Coach
Joined
·
1,734 Posts
Why is he not a fan of the power meter I wonder, that seems really odd. Speed the way he suggests setting up and using the trainer is just a different approach to assessing power output.
While trainer speed is a proxy for power, it's has many more inherit sources of bias and error than a good power meter would.

Trainer speed is a low-fi power proxy. As a result, the applications of such data are confined to those that can be achieved with low-fi data (e.g. broad indicator of intensity while training).

Whether an individual would benefit from being able to use applications that require higher-fi data (e.g. insight derived from use of power duration models, management of training load dose and response, aerodynamics assessment/refinement and so on), well that's a different question and depends on individual circumstances. It's certainly possible to train well with lo-fi data but there are some helpful insights that such a tool can never provide that quality power data can (if you know what you are doing).

Not all power meters provide hi-fi data, some are still what I'd call low to medium-fi. e.g. no unilateral power meter can provide sufficiently hi-fi data for some applications.
 

·
Registered
Joined
·
3,935 Posts
While trainer speed is a proxy for power, it's has many more inherit sources of bias and error than a good power meter would.

Trainer speed is a low-fi power proxy. As a result, the applications of such data are confined to those that can be achieved with low-fi data (e.g. broad indicator of intensity while training).

Whether an individual would benefit from being able to use applications that require higher-fi data (e.g. insight derived from use of power duration models, management of training load dose and response, aerodynamics assessment/refinement and so on), well that's a different question and depends on individual circumstances. It's certainly possible to train well with lo-fi data but there are some helpful insights that such a tool can never provide that quality power data can (if you know what you are doing).
Yep, Agreed - I was just pointing out the approach he's recommending is a crude approximation of power monitoring which he says he is opposed to, there's an obvious conflict with his suggestion.
 

·
Cycling Coach
Joined
·
1,734 Posts
Yep, Agreed - I was just pointing out the approach he's recommending is a crude approximation of power monitoring which he says he is opposed to, there's an obvious conflict with his suggestion.
I didn't intended my comments as being specifically in response to yours about Obree (which make sense, although the sands of time and limited info I'd be surprised if he was truely anti-power meter, probably just more pro-training and ways to do so cost effectively).

I must have hit quote on your post unintentionally when I meant to simply respond to the OP and the thread in general.
 

·
Registered
Joined
·
65 Posts
Discussion Starter · #10 ·
While trainer speed is a proxy for power, it's has many more inherit sources of bias and error than a good power meter would.
I have mixed feelings about this. In particular I am skeptical about crank based power meters. Left side crank in particular does not seem like a good approach. In addition I think a choppy and inefficient pedal stroke may report through the strain gauges as "power" even though it is simply wasted effort. The Power Tap hub based approach seems like a good way to measure but they only claim 1.5% accuracy. For me the appeal of a trainer based design beyond the cost is that the system *might* be more accurate vs. a crank or hub based approach simply because the technology may be "too dumb to lie".

The way I see error points in the trainer from one week to the next is as follows:
1) temp (the training room will be inside so this should be minor)
2) tire (provided the setup of the bike on the trainer did not change this should be quite small since the air pressure will always be checked)
3) magnetic unit (again temp in the room does not change and as such the system resistance should not change).
4) aerodynamics of the moving rear wheel (air conditioning should reduce the impact of humidity but I still need to worry about high pressure vs. low pressure days).

I am not sure if the trainer will be more or less precise vs. power meters.

So basically to simplify my question, what drift do you see between your magnetic trainer speed from one week to the next week vs. power meter readings.

The issue may resolve it's self because I decided to put a bid on a used power tap via ebay.

I'd be surprised if he was truely anti-power meter, probably just more pro-training and ways to do so cost effectively).
Obree's book really good and his approach is very scientific. I found the book and it's writing style enjoyable reading especially because I am a big fan of his riding career.

He does not come across as opposed to the idea of power but rather it simply does not fit in his approach to training. His approach is compelling in it's simplicity and the ability to measure the training effect. For this year I have decided to try to bring some elements of his training into my program.

For those who have not or will not read the book the approach can be simplified as follows:

1) Measure yourself via a 20 minute session on the trainer, record distance
2) Include some other workouts from his plan in the training cycle (7 to 14 days)
3) After recovery, repeat step 1 to ensure you have improved (or at least not become slower)

The fitness change for a well trained athlete from one week to the next is going to be quite small. A 1.5% change in fitness would be huge. Hopefully as an analog system the trainer will be accurate enough for my needs. If my ebay search goes well I plan to obtain a power tap hub to help audit the trainer but for cases where the drift between power tap, trainer, and the prior week is under 2% change it will be hard to say if my fitness changed or if I am inside the margin of error.

The core of his approach is to focus on the feedback loop between training, rest, and the measured turbo workout.
 

·
Pack Fodder.
Joined
·
1,871 Posts
In particular I am skeptical about crank based power meters. Left side crank in particular does not seem like a good approach. In addition I think a choppy and inefficient pedal stroke may report through the strain gauges as "power" even though it is simply wasted effort. The Power Tap hub based approach seems like a good way to measure but they only claim 1.5% accuracy. For me the appeal of a trainer based design beyond the cost is that the system *might* be more accurate vs. a crank or hub based approach simply because the technology may be "too dumb to lie".
All power measurement systems have their strong and weak points. I favor crank spider systems over crank-arm systems (did not have an exceptional experience with my Stages). I favor them over wheel-based systems because I'm not limited to one wheel for power measurement (I use several sets). You pick what fits your specific situation.

At the end of the day, you're looking for a standard to compare your efforts against. It needs to be stable and repeatable. Whether or not it directly compares to another standard is pointless if you don't train by both standards. The exact numbers you hit don't mean anything, except in relationship to other numbers you achieved with a similar standard.

Since I was mixing power meters, I took the time to compare how each one responded to different input. The SRMs and Quarqs seem to track very close to the Powertap wheel I compared it to. The Stages I owned tracked less closely, in terms of indicated power and response curves.

If all I was using was a Stages, that would be the standard and it would have been perfectly fine (works for Sky). Since I was using components from multiple manufacturers, I felt the need to have them track relatively closely.

I have a crank-based power meter on each of my road bikes (2 SRM, 2 Quarq), because I got tired of swapping them around every time I wanted to ride a different bike. That way I would get a more complete picture of stress and fatigue, as I am prone to blowing all of my fitness needlessly instead of building towards specific, targeted goals. I tend to track eBay and other sites and pick up deals when they appear, so I've paid less for my four power meters than a single SRM would cost retail. Calibration/validation by the manufacturer after the purchase provides a little piece of mind when dealing with used equipment, but none of mine have shifted significantly before/after the manufacturer looked at them.

I wouldn't say a power meter is required to improve. I wouldn't say it's invaluable. There are plenty of ways to measure performance that are cheaper. I would say, used correctly, it provides data points you won't get otherwise. If you don't make effective use of those data points, you don't need a power meter and you're wasting your money. The reason I can find them on eBay or wherever so cheap is there are a lot of people out there wasting their money.

If Obree's point is a push back at the perception that the only way to get faster is a power meter, I would agree with him. For the vast, vast majority of riders, a simple structured training plan is the most effective way to get faster. A power meter just helps you quantify it.
 

·
Registered
Joined
·
163 Posts
To answer your specific question, on both my magnetic and my fluid trainer, I notice that my speed has to increase after about 40 minutes to maintain the same power. I use zwift virtual power on one bike and a power meter on another that I use on the trainer. In my experience, as the tire heats up and the drum on the trainer heats up, there is a slight difference in resistance.

That being said, when I used virtual power I trainer and got stronger, I just didn't have any data from outside. The power meter lets me see what's happening when I'm on the road.
 

·
Cycling Coach
Joined
·
1,734 Posts
I have mixed feelings about this. In particular I am skeptical about crank based power meters. Left side crank in particular does not seem like a good approach.
I said good power meters.

Keep in mind that some good power meters have been subject to scientific validation and you can review the science. For example, SRM power meters are regularly used in science study of cycling performance because of their known level of measurement capability.

In addition I think a choppy and inefficient pedal stroke may report through the strain gauges as "power" even though it is simply wasted effort.
You'd be wrong though. A power meter measures mechanical output, not cycling efficiency. For that you'd also need a metabolic cart and quality lab measurement processes.

Interestingly, pedal stroke type has very little to do with measured efficiency of cyclists, indeed attempts to even out the pedal stroke typically result in lowering gross efficiency (ref studies by Korff et al).

The Power Tap hub based approach seems like a good way to measure but they only claim 1.5% accuracy.
In general the Powertap is a good meter. It has some specific limitations and quirks and isn't the highest-fi meter but it's pretty good. e.g. it has a limitation when measuring short duration maximal power, and that a function of it's method of measuring power via fixed time sampling of hub torque.

1.5% for a instrument of this nature is actually pretty good however accuracy statements of power meters are generally a bit incomplete and are mostly marketing statement, not an engineering performance statement. They really should state under what conditions and type of measurement the accuracy claim holds.

For me the appeal of a trainer based design beyond the cost is that the system *might* be more accurate vs. a crank or hub based approach simply because the technology may be "too dumb to lie".
Typical trainers are very well known to have a quite variable power demand while being ridden at the same wheel speed. There are many reasons for that, different tyres, different press on forces, different tyre pressures, different heat response in the tyre-roller interface, different reactions of the resistance unit to heat build up and dissipation and so on, and of course such trainers have little to no hope of doing anything other than approximate a power reading during steady state conditions - their ability to account for non-steady state cycling is virtually zero.

e.g. stop pedalling and the wheel still coasts for a while. Power = zero but such a trainer will still report a power value. Accelerate and the reported power from such trainers will not reflect the energy demand of accelerating the flywheel.

The way I see error points in the trainer from one week to the next is as follows:
1) temp (the training room will be inside so this should be minor)
2) tire (provided the setup of the bike on the trainer did not change this should be quite small since the air pressure will always be checked)
3) magnetic unit (again temp in the room does not change and as such the system resistance should not change).
4) aerodynamics of the moving rear wheel (air conditioning should reduce the impact of humidity but I still need to worry about high pressure vs. low pressure days).

I am not sure if the trainer will be more or less precise vs. power meters.
It will be less precise than a good power meter. It will also be less accurate.


So basically to simplify my question, what drift do you see between your magnetic trainer speed from one week to the next week vs. power meter readings.
When I had one (long time ago now), it was all over the place like easily a 30-50% drift in power demand at same speed/gear/setting (not that it mattered as I had power meters), but honestly that was the least of the issue with a mag resistance unit, which IMO are just horrible to ride and I was unable to maintain anything like the power output on one than I could sustain outdoors. Once you use a decent indoor trainer, and with adequate cooling then you'll begin to approach outdoor power capability.

The fitness change for a well trained athlete from one week to the next is going to be quite small. A 1.5% change in fitness would be huge. Hopefully as an analog system the trainer will be accurate enough for my needs. If my ebay search goes well I plan to obtain a power tap hub to help audit the trainer but for cases where the drift between power tap, trainer, and the prior week is under 2% change it will be hard to say if my fitness changed or if I am inside the margin of error.
No mag trainer (or any regular wheel on roller trainer) will be able to provide that level of repeatable resolution in calculated power.

Keep in mind that normal physiological variability in day to day performance is easily of that order of magnitude.

The core of his approach is to focus on the feedback loop between training, rest, and the measured turbo workout.
The principle is sound, it's just using a measurement method that is reliable.
 

·
Cycling Coach
Joined
·
1,734 Posts
If all I was using was a Stages, that would be the standard and it would have been perfectly fine (works for Sky).
It works for anyone who is only using it for applications that are still mostly OK with low-fi data. Sky have access and resources to manage their hi-fi data needs, e.g. other performance testing methods, other power meters (they still have SRMs) wind tunnels and so on. And Stages also supplies them with dual sided models which are not commercially available.
 

·
Cycling Coach
Joined
·
1,734 Posts
At the end of the day, you're looking for a standard to compare your efforts against. It needs to be stable and repeatable. Whether or not it directly compares to another standard is pointless if you don't train by both standards. The exact numbers you hit don't mean anything, except in relationship to other numbers you achieved with a similar standard.
The trick is what happens when you need to change your measurement device. Make life way easier when the devices are accurate to begin with.

BTW - just commenting on a couple of things - basically agree with your post
 

·
Pack Fodder.
Joined
·
1,871 Posts
The trick is what happens when you need to change your measurement device. Make life way easier when the devices are accurate to begin with.

BTW - just commenting on a couple of things - basically agree with your post
I agree about the lo-fi nature of the Stages units. I was an early-adopter with Stages, and they've made great strides with firmware upgrades over the time, but I could not achieve the same sort of consistent results I got with the Quarqs and SRMs. I originally bought it for my 'cross bike so I could get data from those races. I figured it was cheaper than a Quarq, so damaging it would be less painful, but my first-gen battery cover allowed water in and killed it for a while. By the time I replaced the cover and brought it back to life, I no longer raced 'cross anymore. I tried it on the TT bike, but found the difference in power readings to be demoralizing, even though I knew the adjusted targets.

My Quarqs have been very reliable/stable over the years. I'm hoping the SRMs will be equally as reliable/stable now that they have fresh batteries.

My main point is that even a lo-fi unit such as the Stages is a more reliable standard than relying on indicated speed on a wheel-on trainer. Even a wheel-off trainer (such as the LeMond) that eliminates many of the variables can fluctuate a bit. The bonus with the power meter is that you can take that same standard out on the road and apply it. "Trainer miles aren't the same as road miles", so your perceived exertion and resulting speed can vary greatly, effectively creating multiple standards.

Then again, it's all about goals, priorities, and resources.
 

·
Registered
Joined
·
1,449 Posts
Back to one of OP questions, I have a quarq elsa and an old blackburn mag trainer (mid-90s). The trainer has five different resistance levels that are selected with a switch thing on the magnetic unit.

Over relatively short intervals (6-8 minutes), I have to increase my speed from 1 to 2 mph to maintain the same power. As the workout gets longer (mag trainer gets up to temperature), speed has to increase in rate of change and magnitude to maintain power level. IOW, at a given resistance level the trainer does not actually maintain the same resistance over time.
 

·
Registered
Joined
·
65 Posts
Discussion Starter · #18 ·
As an update, I now have three weeks of these Obree style trainer workouts under my belt. The trainer is setup the same each time and the results seem consistent. With the first workout as baseline the second workout was 4% faster. The third workout added another 2%. The results are consistent with my expectations. It's been a good couple of weeks of training so I am sure that I have improved a bit as well. The big question is how the numbers will track as the rates of week to week improvement gets smaller. I expect in a couple of weeks my body will reach a plateau such that changes will be in the order of .25% over the 20 minute workout.

That said, I did end up buying a used Power Tap hub from Ebay. The problem is that although both the hub and the joule computer came I seem to have a defective computer because it reboots after showing fairly erratic watts and speed for between 30 and 60 seconds. Hopefully I will get it resolved in the next couple of weeks.

In any case, for people that are following the thread so far it looks like the a magnetic trainer produces consistent results for measuring fitness.
 
1 - 18 of 18 Posts
Top