Dynamic-boom vs. Fixed-boom Recumbent Bicycle Power Production

[Jim Parker]

Are you able to produce more power when your ride on the road, or when you are on the trainer?

If you have a power meter, trainer, and heart rate monitor, consider trying to do what I did and share your results on this thread. Here's what I did, and my results.

I compared peak power while performing the H.A. & A.C. Power Profile Test (as published by TrainerRoad.com, protocol and described in detail below) once with my V20 locked in a trainer (Wahoo Kickr smart trainer) where the front triangle (boom, chainstay, and bottom bracket) cannot rotate. In this position, the V20 operates like any fixed-boom (RWD) recumbent bicycle, with no ability of tugging on the handlebars to move the BB side-to-side. Next, I performed the same test on the road, where I intentionally engaged my upper body during the test segments.

Effort level during the tests was a maximum effort for the duration of each test interval, which were 5 minutes, 1 minute, and 15 seconds. Active recovery (several minutes of easy pedaling) were interspersed between the test segments. Power and heart rate data were collected and uploaded to the TrainingPeaks.com application for analysis. Peak power at each interval was automatically calculated by the TrainingPeaks application based on data collected from a Garmin heart rate monitor, and power meters 1) in the Wahoo Kickr trainer, and 2) a PowerTap hub-based power meter in the front wheel of the V20. Power meters were calibrated prior to each test.

Results:
Peak Power at every tested interval was significantly higher with the dynamic-boom when compared to the fixed-boom. Peak heart rate was not significantly different.

Table 1. Peak Power and Peak Heart Rate (W = Watts, HR = heart rate)


Figure 1. Graph generated by TrainingPeaks.com application






Discussion:
I recently published a report showing that, on a Cruzbike V20 bicycle, when a forward force is applied to the left pedal when the left crank is in the upright (or near-upright) position, a tug on the left-end of the handlebar will significantly increase the torque in the left crank. The torque increased from 13 Newton-meters by about 30%. This increase in torque in the left crank occurred when the force on the pedal was applied by the human-leg, a 25 lbs. dumbbell, or a bungee cord. It follows, therefore, that more power may be generated during actual cycling when the upper body is able to rotate the bottom-bracket (BB). This pilot study tends to support that theory.

Power Profile Tests, such as the one used in this experiment, have been shown to correlate with actual cycling performance in time trial events. The Hunter Allen and Andrew Coggan Power Profile test requires a warm-up phase, and then contains three 1-minute tests, a single 5-minutes effort, and two 15-second efforts.

Limitations of the experiment:

1) Efforts longer than 5-minutes were not tested.

2) Only one subject was tested, and that subject (me, the author) is also an owner of Cruzbike, Inc., and therefore may have consciously or unconsciously biased the intensity of his efforts. The fact that heart rate monitoring showed no significant difference between the trials would tend to support the claim that an equal effort was made. I gave both trials my best effort. I am in a hard-training phase right now and both trials were done without any taper.

3) The power meters were different. Even though they were both calibrated immediately before the trials. An error range of 1-3% is common with power meters, so this is unlikely to be the cause of the large differences recorded.

4) Heat build-up has been described as a limitation of peak performance on an indoor trainer. The air temperature was in the 40-50 degree F range in the garage, and fans were running. I did not feel any limitation from heat during the testing, especially considering how short the efforts were.

5) I had a flat tire during my warm-up for the road trial and had to cut my rest time between test sets so that I wouldn't be late for work. I think I could have done a little better with more rest.

These results help explain why racing on a Cruzbike makes us better climbers and sprinters. Please add your test results and see if they compare to mine. I believe most (but not all) cyclists can get more useful power by engaging their upper body on traditional bikes and Cruzbikes. But the technique takes some practice.

-----------------------------------------------------------------------------------
Here is the protocol for the H.A. & A.C. Power Profile Test
Duration: 1:25:00

Hunter Allen & Andy Coggan

Calibrate Power meter!

Warm-up

15 minutes building gradually from 65% to 75% of FTP

1 min at 90% FTP, then recover 1 min at 65% X 3 sets total

5 min at 100% FTP, then recover 5 minutes at 65%

Test begins

1 minute all-out sprint, 10 minutes of recovery at 65%

5 minutes all-out effort, 10 minutes of recovery at 65%

1 minute all-out sprint, 5 minutes of recovery at 65%

1 minute all-out sprint, 5 minutes of recovery at 65%

15 seconds all-out sprint, 2 minutes of recovery at 65%

15 seconds all-out sprint, 2 minutes of recovery at 65%

Warm-down

15 minutes gradually decreasing from 75% to 50%

---------------------------------------------------------------------------------------

Reference:
Quod MJ, Martin DT, Laursen PB. The Power Profile Predicts Road Cycling MMP. Int J Sports Med, Published online: 2010, ISSN 0172-4622.

 

[RojoRacing]

I find you results highly suspect based on my own observations between road and trainer power readings between 5 seconds and 20 mins. I haven't done a structured test like you have but all my climbing segment efforts when riding zwift have always mirrored my efforts when outside. I don't claim to be a very fast climber on my Vendetta because I've got to compare it and myself to are a bunch of elite racers on 15lb carbon DH bike who constantly walk away from me. Now even though I claim to not be the best at climbing, you'd have a hard time finding many who can ditch me on a climb. I find going out and winning races with tons of climbing involved is easier that trying to prove the dynamics of how it can be done.

 

[Jim Parker]

I find you results highly suspect based on my own observations between road and trainer power readings between 5 seconds and 20 mins. I haven't done a structured test like you have but all my climbing segment efforts when riding zwift have always mirrored my efforts when outside. I don't claim to be a very fast climber on my Vendetta because I've got to compare it and myself to are a bunch of elite racers on 15lb carbon DH bike who constantly walk away from me. Now even though I claim to not be the best at climbing, you'd have a hard time finding many who can ditch me on a climb. I find going out and winning races with tons of climbing involved is easier that trying to prove the dynamics of how it can be done.

I admit I was a bit surprised by the magnitude of the difference in power with my testing. But the numbers are genuine, coming straight from the power meters and hear rate monitor. And the analysis and graphs/charts were automatically created by TrainingPeaks, using their own complex algorithm to calculate peak power and heart rate. I’m not asking people to take my results as the final word, but rather asking people to do their own power profile testing and share them here. You might be surprised by what you will find, too.

 

[super slim]

Jim, I have no power meter, to compare, but I think that a 30% improvement of MBB compared to fixed, seems high, and more test would be required!
I think there is a Gnome waiting to test it!

I think that the power improvement is 50% the moving Bottom Bracket and 50% to a VERY STIFF power triangle, compared to RWD recumbents, especially with some of the huge long handlebars I have seen, OR the hampster bars where no force can be applied, OR USS where the bars flex too much and the shoulder to bar, is 90 degrees to a Silvio straight line of shoulder to handlebar to BB.

Or the fact that the Vendetta is a LOT more aero than the RWB recumbents that you race against!

Going from an under seat steering recumbent trike (Scorpion FX 3*20"-406 wheels), to the Silvio V1.0, my SLOW average speed for a 100 km flat ride went from 20 kph(12.5 mph) to 26 kph (16.2 mph) overnight, even with a vice like grip on the handlebars!!!

 

[trapdoor2]

My warning light went on as soon as I saw two different measurement devices were used. While I do believe there are gains from dynamic vs static, 30% is indeed hard to believe. I would recommend, if possible, capturing data separately from the Powertap whilst on the trainer. Best if you do so by capturing data concurrently (from both the Kicker and Powertap)...then you will have a basis for comparison out on the road. You may be surprised to find that 1-3% is an order of magnitude off (or not!).

 

[Mark B]

This discussion seems like it would be a significant cause of consternation and gnashing of teeth on that other forum. Maybe.

Mark

 

[trapdoor2]

This discussion seems like it would be a significant cause of consternation and gnashing of teeth on that other forum. Maybe.

No "maybe" about it. Something like this would escalate until Bryan had to throw folks out the door...again.

 

[super slim]

I do not think any of my comments would be appreciated by the RWD recumbent group!

 

[ratz]

Ok this is a good start but I think it's incomplete to compare for the target.
We need a few extra variables that I can outline.
I know Larry has all the pieces to do this test, and Jim has most of them.

I'm going to take a swing at this topic, the science minded can correct my errors but I think this is pretty close to the best we could do for N=1 sample size testing.

Couple of knowns from the training world

1. Some rider produce better power outdoors, other produce more indoors
2. Item #1 seems to hold constant across bike and positions for a give rider
3. The majority of people produce more power outdoors (as self reported by people when asked)
4. Consensus is that the free movement of the bike maximizes the rider power producing positions for most people

Limitation of gear

1. We need to measure power for this comparison at the pedals or crank and wheel;
2. Power at wheel is power derived, that's the one we really care about when comparing platforms.
3. Power at Pedal / crank is power input
4. There will be some loss of energy between 2 and 3 and it is not linear relative to torque
5. Wheel off trainers require a power input from second power meter to be compared.

Testing Protocol:
1. Test on 3 bikes including: DF frame, V20, and a Stick bike as close to v20 geometry as possible probably an M5.
2. Indoors simultaneously record on two Heads (probably Garmins) record the data from a Pedal or Crank meter on one head unit; record Kickr Power on head unit 2.
3. Outdoors again two Two head unit one for pedal/crank record; and one for a power tap wheel
4. In Both cases the HAAC Power Profile test would be perfectly suitable; stop after warm up phase and calibrate all power meters, outdoors it's best to test on a 1-3% uphll grade.

Analisys:
0. Take 4 data files to Wko4
1. Throw out the warm up data
2. Compute the MaxMean Power (MMP) curve for indoor session (both files). but ignore data below the 5 sec mark maybe the 10s mark the Kickr over smooths power so that data is suspect.
3. Compute the maxMean power curve for outdoor session (both files) again ignore data below the 5 sec mark because of kickr limitations
4. Compute the average offset of the two indoor files MMP curves; Compute the average offset of the two outdoor files; compute a relative offset of the kickr to powertap.
5. Compare the MMP difference indoors kickr vs powertap outdoors. Reduce any measured gain or loss by the % offset of the kickr to powertap

That's going to give a reasonable measure of the percentage gained going from indoors to outdoors. Testing on 3 bikes and comparing will give some limited idea of how much of the gains come from freeing the bike and rider from the static position and how much might come from the MBB over and above that.

Pretty exhausting test and that's a ton of uncontrolled variable; but that's always been the nature of the beast...

I would expect to see:

1) Little to Some gain on a stick bike 10-15% (bike is freed but rider stay static relative to the bike)
2) More gain on the DF 15-20% (bike is freed and rider is freed, but rider was "freed" indoors already)
3) Most gain on the V20 20-25% (both bike and rider freed from static position).

Note 1: The kickr power smoothing makes this whom test troublesome. A Tax Neo would be a better training to use as it would make the < 10s power usable. I suspect a lot of the V20 big gains are in the sub 10s window, I can hit some pretty sick power numbers accelerating out of a turn into a climb; +1200, I don't think I could to that on a stick bike but that's guess.

Note 2: Due to athlete's fading strength; I would do 3 days of testing. Day 1 Bike one indoor and then outdoor; repeat day 2 with bike two, repeat day 3 with bike 3. You want to compare the % changes from inside to outside regardless of daily strength the percentages should hold up even if you weaker on different days. I'm sure smarter people could refine that; but it's going to still be nebulous no matter what you do.

Note 3: There is another way to do this on a long game approach record a season worth of training and riding; flag indoor rides; and indoor rides with a tag in training peaks. Compute a season's worth of MMP for both situations and compare. The problem is to compare to another platform for a single rider you'd have to ride all 3 in equal rotation all season long, which probably isn't going to happen. Tim Turner has a vast library of his personal data for CA2 and DF and I think that's pretty unique to all the riders we know. The problem is not many people put equal mileage in on all platforms. So Assuming the platforms effect all riders the same if you get a large enough sample size to remove the curve busters, you could compare averaged data from a large group of DF riders seasonal average indoor to outdoor gains to large group of v20 riders indoor to outdoor gains and work from a much larger statistical pool.... The 4f training project is gather this MMP data on about 15, v20 riders (yeas more would be better). Currently We don't have enough data to do that comparison yet; but by next December we should. It's probably about a weeks worth of work once all the data is on hand. DF data is abundant and fairly easy to get.

 

[DavidCH]

I pdf'ed the brol discussion... it's 20 pages. I reckon chainrings have a different dynamics. The rotor QXL ring accelerates on the dead spot, therefor I would have thought the load going thru the pedal would have a bigger dynamic effect thru the boom. I know that I am faster using a normal rotor q ring than the standard de facto chainring, so I suspect the rotor QXL ring will have an even bigger effect. I will let you know when I change over after Easter. But I will have error with the results as my cranks will be 150mm instead of 165mm.

 

[trapdoor2]

Ok this is a good start but I think it's incomplete to compare for the target.
We need a few extra variables that I can outline. [big snip]

I think you have covered the overall picture quite well. But, as interesting as comparing across platforms might be, I don't think we need to test anything but the V20...at first. Yes, complex testing is always fun. However, I'm a fan of simplicity. Test the test first. If you get a result that drives you further, do it across a platform or two.
I would be very surprised if there isn't already data available for this issue from a DF standpoint. Heck, they test everything else!

 

[PeteClark]

Seems the primary objective here is to quantify the difference in power that one can deliver on an MBB vs. FBB recumbent. Jim took a swing at that by comparing his trainer and outdoor results, but the trainer vs. outdoor differences may obscure the MBB vs. FBB differences. I would focus on just outdoor testing to avoid that complication. The simplest test is often the best. If we had a few riders equally well adapted to a V and an M5, all we'd need to do is measure their power output on each bike in similar conditions.

 

[ratz]

I don't think we need to test anything but the V20...at first.

Yes indeed cross platform is only needed in the case of N=1 or a sample group under say 5; without that you have to compare 1 rider across platform. Whoever if you get gather mass data on say 15+ riders the more the better........then.

The simplest test is often the best.

And that would be why the 4F project exists it generates data; just takes time to get enough; that project cost about $100 a month to keep the software running; plus what everyone is personally paying for their TR subscriptions. But in the end the data is going to be far far far worth the expense and time. The most interesting data comes from in season when people bounce from indoors to outdoors.

 

[trapdoor2]

Yes indeed cross platform is only needed in the case of N=1 or a sample group under say 5; without that you have to compare 1 rider across platform. Whoever if you get gather mass data on say 15+ riders the more the better........then.

No argument that mo data = better. The problem is that it is often impractical to do so. Right now, this data set lacks correlative data between the two test sets (Kicker vs Powertap). I have sit in some seriously boring meetings while test engineers hash out correlative issues between "test set A" and "test set B", even though they were virtually identical. Is 0.00004 bal/m2* important?

Once the test is determined to be valid...then, yea, go forth and gather thy multitude.

*bal/m2 = balrogs per square meter

 

[ratz]

Is 0.00004 bal/m2* important?

If you want to come back from Mars; yep; if you are staying for a lifetime probably not.

My 6 year old daughter declared last month she want to be the first woman of her generation to walk on Mars.
Two days later she modified it to say she wanted to be the first woman to walk on Mars and come back.
Best to know you assumptions and goals before you start planning

Too all of these data; ANYONE that has a Power Meter and a recording device; even if you dont' train; consider getting a free Training Peaks account and using it to submit your data to the database. cost is free and the data is going to get used; eventually.

 

[McWheels]

I've long wondered about the additional help one gets from the dynamic boom, or MBB. I've also seen the piece on increased torque when set against feet/barbells/bungees. Caveat: I'm not a racer and I don't have test equipment, but it's going to fun writing this.

Starting way back, the power of vacuum was first demonstrated with 2 teams of horses unable to pull apart 2 hemi-spheres. Thing is, it would have been the same effort on the hemi-spheres had it been 1 of the horse teams pulling against a fixed pole. Newton's 3rd applies, but not to marketing. Therefore in the purest mechanical sense, whether you pull against the handlebars or not, and whether you have a stick bike or a dynamic boom, bracing the human against something for the benefit of the pedal will make more torque, it's merely the connectivity and efficiency of one to the other we are disputing.

On the CB, I brace the effort in the pedal through:

• pedal material
• pedal bearings
• crank arm
• bottom bracket
• boom
• chain-ring
• chain
• sprocket
• driven axle
• spokes
• rim
• tyre
• asphalt

So we discount what is common and it comes to me that the inefficiency of the booms in bending and flexing which isn't recovered is where we are winning, and not necessarily in the first-order design differences. If something gives, then I lose that power at that time. Mechanical spring efficiency is never 100% so it's unlikely I get all of it back, especially if it's something I wasn't planning on bending.

And this sits more easily with me as an explanation about a CB's climbing prowess:

• I lose less energy in the front triangle as it's inherently stiffer than a 2D stick.
• I have better connectivity from other muscle groups to the pedals because of the dynamic boom.
  • Is it better when the shoulders through the hands to the feet are in a straight line? Pulling off-axis strikes me as also inefficient as only a component of the vector goes where you want.

 

[Mark B]

No "maybe" about it. Something like this would escalate until Bryan had to throw folks out the door...again.

Hmm.. Sorta like throwing a skunk into a packed room and standing back to watch the mayhem. Sounds fun!

 

[Rampa]

A rowing machine really should make it clear why it works better to anyone. Or a game of tug-o-war. When you pull at 180 degrees from your push, you exert the most force.

 

[Jim Parker]

More on the physics: Yes, I would agree that even if we are merely "bracing the body" by pulling on the handlebar, then that would allow more power generation. And this may be all that many successful Cruzbike riders ever do, and that would be fine.
But if we go a small step further, and actually move the BB closer to the hip during the downstroke, then there is a direct mechanical increase in torque at the crank. This diagram illustrates force from the lower body is acting to move the pedal forward while force from the upper body is moving the BB backwards, thus increasing torque at the crank/BB.
Conceptually, it is essential to see the crank as a lever arm between the pedal and the BB. The upper and lower body are not acting directly on the same point, but rather their actions culminate at opposite ends of the crank arm.

 

[Rampa]

I suppose the "waggle" lets you keep a greater amount of leverage for a longer time. Kind of like if more of your pedal-stroke per side were at twelve 0-clock. Or something like that.