Weight Weenies

I should have expanded on my earlier comment - I believe testing should show both - no rider and with a rider. Yes let the bikes be evaluated side by side no rider.....but there is so much impact of the rider that it then shouldn't be discounted.....manikins like Cervelo and others use seem a great step in that direction.

Not a true apples to apples comparison here, but the Velo tests include testing of frame "flex" AND also subjective rider comments related to how it absorbs bumps, accelerates, overall stiffness perception, etc.....thus they offer frame vs frame data combined with real world feedback via a rider on the bike. The wind tunnel with and without a rider isn't exactly the same suggestion (both are still analytical results) but hopefully you get what I mean - bikes do have riders.

Visit starbike.com Online Retailer for HighEnd cycling components
Great Prices ✓    Broad Selection ✓    Worldwide Delivery ✓

www.starbike.com

If possible, when conducting future tests, could you please test bicycles with the same pricepoint / specifications? I found it odd that the midlevel Madone 6.2 was tested versus the top offering of the Propel. Wouldn't it have been more of a direct comparison to use both bicycles with Dura Ace / carbon tubular wheelset specs, using a Madone 7.9?

duke249 wrote:If possible, when conducting future tests, could you please test bicycles with the same pricepoint / specifications? I found it odd that the midlevel Madone 6.2 was tested versus the top offering of the Propel. Wouldn't it have been more of a direct comparison to use both bicycles with Dura Ace / carbon tubular wheelset specs, using a Madone 7.9?

I believe all the bikes Velo News tests in the wind tunnel have Enve Smart 6.7 wheels on them. They remove the OEM spec wheels for testing.

Tinea Pedis wrote:Because without the rider you get a true appreciation for the differences between the frames. You're removing a massive variable (my body and position is different to yours is different to the next guy, etc) and getting a better reflection of how aero the frame is (or is not).

No, this is not correct. The problem with rider-less testing is that the rider significantly affects the flow distribution over the frame, as well as onset conditions to the components. The problem with changes in the flow distribution is that the sum of components contribution to the overall drag changes; the problem with different onset conditions is that the drag coefficient of various components changes.

Of course, mannequin testing is not without its own problems. Neither are accurate compared to the real-world, the question is really whether the answers are helpful in selecting a frame. Without evidence that the drag delta doesn't change between rider/rider-less configurations, I'd be inclined to class the riderless results as not-helpful in frame selection. Which is not to slate the VN test - there should be more testing, not less! - rather this highlights the real-world challenges of aero design.

If I were BMC's frame designer, I'd also have pulled my frame from the test unless, of course, I knew I had optimised my design using riderless CFD/wind-tunnel models

PS Lower body only mannequins are not worthwhile because the torso does change the overall flowfield quite considerably.

PS PS This is also the reason why you spin the wheels rather than keep then static or not include them at all, because spinning wheels quite profoundly change the flow conditions, particularly on the fork and downtube.

^ I like your username!

Agreed, you need a full mannequin. Not necessarily pedaling, but at least something that can be placed consistently

Courant wrote:No, this is not correct.

I must have missed something here, as I never said rider-less was the conclusive answer and the correct method.

Rather, it eliminated a massive variable - that you yourself admitted

Neither are accurate compared to the real-world

and is ultimately why I earlier stated

to conclusively answer that question is to take themselves and a pile of bikes and either conduct their own aero tests or get in to a wind tunnel.

The rider-less design not only has one eye to reducing that significant variable, but also appreciates that VN is a magazine. Not a fully-funded study. The time in a wind tunnel is huge money. To do tests of both rider (of which you would also immediately get people complaining "I'm not the same height or weight or width as your mannequin, the test is useless to me") and rider-less would cost a lot!

I applaud VN for doing these tests. Damned if they do, damned if they don't. As this thread just shows, there will always be people saying "no, this is wrong".

Tinea Pedis,

I think you misunderstood what I was trying to say. It's not a question of whether the results is "wrong", it's a question of whether the results are useful. All results of all experiments are always "wrong" - you can always poke holes at an experimental design on the basis of it being wrong. Instead of making results more "correct", you should seek to make them more useful, the two are not the same. This is particularly important when you have limited resource.

Concretely, it's very hard to say off the basis of these test results whether one frame is in reality quicker than another. I.e. the certainty with which you can say a given drag delta has the correct magnitude or even sign is low (e.g. among the top tier frames in the test). In this respect, testing with a mannequin, while introducing it's own complications, is potentially more useful (even though it's still wrong), assuming you put the effort into correctly accounting for the mannequin's drag, because it increases the certainty with which you can call the relative (real-world) drag deltas (if not the absolute values).

In other words, and why I said your initial statement was incorrect, was that if you define:

Correctness as the probability that the results of the test are correct.
Usefulness as the probability that a decision based on the results of the test are correct.

Riderless and mannequin tests are about as correct as each other.
A mannequin test is likely more useful than a riderless test.

(Ideally you'd spin the legs too, as actually Giant appear to have done in their dev program; and ideally, you'd actually get out on the road with a real rider too).

I applaud the VN test absolutely, this is not a case of damned if they do. However, if you try and up the scientific content of magazine testing there also needs to be, hand-in-hand, a free and open discourse about the meaning, value, and potential errors in the result. Aero design is extremely complex area and it's hard to reduce aero performance to a single value -professionals have issues interpreting aero testing results, let alone lay members of the public (as the misunderstanding above shows ). The danger is that well-intentioned but poorly understood testing has the potential simply to add noise, and give people (and marketing departments) the sensation of certainty where reality is far more subtle. The bit I actually object to more than anything actually, is that if marketing say, design me a frame that wins the (less useful but easy to run) test because this is worth $X in sales, then there's a commercial imperative to not attempt to design a frame that is actually faster in the real-world (i.e. your engineer gets screwed!).

Hope this helps to clarify things!

There is absolutely no evidence whatsoever that so called aero bikes are in anyway quicker in long distance mass start events. There is. Plenty of evidence to suggest that external influences more than negate any alleged findings in wind tunnel tests. Given this, wind tunn tests are irrelevant pseudo science if the purpose of them is to show whether one bike is quicker than another in the real world. It's a win win for marketing departments because it's seemingly impossible to disprove their claims. Meanwhile Garmin have guys in the break riding R series and Bellkin riders on TCR's. A clue?

With the new S3 out, that might change... The peloton has always been very traditionalist. Above all else, they value stiffness, comfort, and feel. Aero is one of those things that you can't really feel (maybe if you're really sensitive you can), but it is still there. Also, lots of riders on the Propel

Courant wrote:I think you misunderstood what I was trying to say. It's not a question of whether the results is "wrong", it's a question of whether the results are useful. All results of all experiments are always "wrong" - you can always poke holes at an experimental design on the basis of it being wrong. Instead of making results more "correct", you should seek to make them more useful, the two are not the same. This is particularly important when you have limited resource.

This is debating the protocols of the test, which people will differ in their opinions of.

'Misunderstanding' comes from

Courant wrote:No, this is not correct.

so maybe just more of an eye to your choice of words, and there will be less misinterpretation of what you're looking to say. Especially when I'll disagree on the lack of 'usefulness' you're implying rider-less testing does not provide.

Again, with an eye to keeping it reasonable on the budget a mag has for this testing in the first place.

Courant wrote:let alone lay members of the public

I write for a bike mag.


airwise, Russell's Teapot again? Belkin also had guys choosing Propel with clip-on bars over a Trinity. But I know where you stand on this, not going to try and convince you otherwise

Courant wrote:

Tinea Pedis wrote:Because without the rider you get a true appreciation for the differences between the frames. You're removing a massive variable (my body and position is different to yours is different to the next guy, etc) and getting a better reflection of how aero the frame is (or is not).

No, this is not correct. The problem with rider-less testing is that the rider significantly affects the flow distribution over the frame, as well as onset conditions to the components. The problem with changes in the flow distribution is that the sum of components contribution to the overall drag changes; the problem with different onset conditions is that the drag coefficient of various components changes.

Of course, mannequin testing is not without its own problems. Neither are accurate compared to the real-world, the question is really whether the answers are helpful in selecting a frame. Without evidence that the drag delta doesn't change between rider/rider-less configurations, I'd be inclined to class the riderless results as not-helpful in frame selection. Which is not to slate the VN test - there should be more testing, not less! - rather this highlights the real-world challenges of aero design.

If I were BMC's frame designer, I'd also have pulled my frame from the test unless, of course, I knew I had optimised my design using riderless CFD/wind-tunnel models

This is what I was trying to say, and agree wholeheartedly.

Tinea Pedis wrote:

Courant wrote:No, this is not correct.

so maybe just more of an eye to your choice of words, and there will be less misinterpretation of what you're looking to say. Especially when I'll disagree on the lack of 'usefulness' you're implying rider-less testing does not provide.

Sorry, usefulness of the testing aside, the statement "without the rider you get a true appreciation for the differences between the frames" is profoundly incorrect and represents a misunderstanding about what is being measured (aerodynamics is a function of both the geometry and the flowfield, not just the geometry).

An easy-to-understand analogous situation is if I claimed to have come up with new saddle covering material that I claim reduces aerodynamic drag by 20% and test this in a wind tunnel without a rider to "get a true appreciation" for the aerodynamic performance of my new material. The results of the test in and of themselves may be sound, but it's a completely worthless result. The rider-vs-riderless frame test is the same, it only differs in degree - you can't, without further testing, differentiate e.g. which components contribute high drag counts in the riderless test and may show nothing like the same with a rider on, and this may well differ between frames. The picture on the test title page is the dead give-away - the smoke trail passes round the front wheel nicely and then, oops, goes exactly through the kneecap of the right leg and then, oops, continues unimpeded through the frame to impact the kneecap of the left frame. Are those missing legs going to impact the flow on the downtube and seatube, and even the front fork? Sure they are! Given the radically different profiles being used by the different frames, would you expect the drag deltas to be different between frames and between flow conditions? Without a doubt (e.g. the downtube profiles given, in yaw, and with legs so the onset flow angle on the downtube being more pronounced, the Trek Kamm profile seems actually quite sensible while the ridley will likely get hammered - i.e. you'd expect the relative performance of the trek to improve with respect to the ridley under a mannequin test). Same goes for all the components on the frame. This is the source of error and doesn't allow you to draw the conclusions that are being made with any certainty. This is why a mannequin test is more useful because the delta between one human and another is much smaller than no-rider to rider (i.e. this is the truer test of aero performance), though you have to account for the error in the estimation of the mannequin drag, which adds complexity - the error term here comes from the drag measurement, not deltas due to different flow conditions.

(Note, which is not to say that the Giant is not a good aero performer, you just can't derive that from this test; and back to the original question, BMC or Giant, again, the non-inclusion of the BMC should not be a factor).

It would be great to see a public-domain rider vs riderless comparison, across different frames. Because, consider this, if riderless testing gave you accurate drag deltas off a simple testing protocol, why do manufacturers use mannequins and moving legs, and why the interest in track based measurement systems bypassing all the issues with tunnel testing completely?

Tinea Pedis wrote:Again, with an eye to keeping it reasonable on the budget a mag has for this testing in the first place.

I absolutely respect this, but I think there needs to be a clear understanding of the value and the meaning of testing on both the consumer and the commissioning side of the equation. Presenting results without debate/discussion is not useful to anyone.

Tinea Pedis wrote:

Courant wrote:let alone lay members of the public

I write for a bike mag.

With respect, the subject matter is aerodynamics, not journalism!

I would buy the Giant simply because it looks cool.

Not sure what to do about testing. Objective tests are all very well if they make useful conclusions for the intended market. I sympathise with publishers. Readers demand something more than subjective opinions and tests like those mentioned satiate them to a degree. But I cannot for the life of me see how they identify any ride quality that may influence a sensible buyer's decision making.

And therein lies my problem. They seem to suggest that because A is demonstrably more aerodynamic in a wind tunnel then the bike will be quicker should the rider buy one. The leap from A to B here is pure fantasy and only of benefit to manufactures and their marketing departments.

So buy the Giant - the blue on black is great in the flesh!

Visit starbike.com Online Retailer for HighEnd cycling components
Great Prices ✓    Broad Selection ✓    Worldwide Delivery ✓

www.starbike.com

Courant wrote:With respect, the subject matter is aerodynamics, not journalism!

With respect, that's not the point I'm making.

Happy riding


airwise - that's the sweet and the sour of the whole 'review' process summed up right there really. What I value and will publish in a review will be useless info for someone else.