The World's Most Expensive Suffix

Another company has come along claiming to be "the world's fastest wheels." It strikes me that "-est" must be the three most expensive letters in the English language.  

We've done a good fair bit of testing, and in doing a lot of testing, you learn a lot. To the good, you learn a lot about what makes your products good, and you learn how you might make them better. You learn how your products compare to others, which increases your ability to help people match a product to their intended usage. Like the torture that was Stanley Kaplan's SAT test-prep course, testing also teaches you how to test better.  

There is some expensive, but low-hanging fruit out there.  If you test your wheels with a whole range of tires, there's bound to be one that gives you an edge relative to the others.  If you are close enough in the first place, that might nudge you over the edge and make you "fastest." Doing this kind of testing is like lighting $100 bills on fire, which eventually the customer (or the bankruptcy) will pay for, and it doesn't actually make your wheels any better, but when you NEED to show the magical "-est," I guess it sounds like a smart spend.  

If you want to do the same with wheels-in-bike testing, that's also expensive but easy. Ever wonder why the copy says "these wheels were the fastest in the test on bike x" but the picture shows said wheels in bike y? Because they tested bike u, v, w, x, y, and z in order to find the one case in which the "-est" bell rang in their favor.  

You can also do some pernicious things, like this "removing the tare" thing that some wheel companies do.  When you test a wheel in a tunnel, something needs to hold that wheel in place. Some companies (oddly enough, there seems to be a correlation between companies claiming "-est" and this technique) run the support struts independently, and then simply subtract that drag from the figures their wheels test at. To use the simplest analogy that I can come up with, this is like saying that if you have an anemometer directly upwind of a brick wall, and another directly downwind of a brick wall, that the sum of their readings is what an anemometer would read in the absence of the brick wall. But set my analogy aside and use the analogy of the wind tunnel (A2) at which we've done our testing - "we don't consider removing tare to be correct protocol." Wait, that's not an analogy, that's a direct statement.  Sorry.  And it's not standard industry practice, either.   

Then you have the simple things.  When we tested 52 versus 404, we used a tube with an 80mm valve stem.  If we wanted to optimize our drag readings for public consumption and comparison, we'd use a short stem and inflate the tires using an extender, then tape over the hole.  The 6mm of extra valve stem that pokes out of a 52 versus a 404 may not amount to the difference that lets us claim "-est," but it's fairly likely that we're the only ones leaving that freebie on the table. Why do we do it the way we do it? Because that's the way people do it when they ride.  

The other challenges of "-est" are more subtle but perhaps more costly in the long run. If you play tricks, they're going to come around and bite you in the ass sometime, somehow. Mike and I put a lot of stock in what we've said in this and other channels, and there's nothing we'd ever have to backtrack and try to "unsay." We would hope that if you read the blog from post 1 to this one that you would see a ton of development, consideration, reconsideration, and incorporation, but no vast right wing conspiracy could justifiably call us flip-floppers on anything. Everything we write is as honest as it can be when it's written, there's no trickery at all whatsoever.  

The other, HUGE cost of "-est" is what do you do when it's gone?  You put yourself on the hot seat, and then someone knocks you off.  Your whole story had been "-est" - what's your story now? We like to think that being really really good at everything means a heck of a lot more than being "-est" at anything.  


Tire Size Reëxamined

Like that umlaut? The Economist does it, so I figure it's probably correct.  Anyhoo... all of this hoo-ha about tire sizes and inflation and tire size versus aerodynamic speed has had us thinking about this quite a little bit lately.  

Step 1 was to to further examine the effects of tire size on aerodynamics.  Ideally, we would have been able to naturally (i.e. without artificially manipulating bead seat width) set two of the same model of tires up to the same inflated dimensions on wheels with different bead seat widths, to test the slope that we'd earlier established that had the 404 and 52 leap frogging each other at inflated width. The 52 with a 23 is, inflated size-wise, a bit smaller than a 404 with a 25 and moreso bigger than a 404 with a 23.  It is a little closer in speed to a 404 with a 23.  What would happen if you put the same tire inflated to the exact same dimensions on a 404 and a 52?  Unfortunately, that's an unanswerable question, but we were able to test a good approximation.  

Continental makes the Attack 22.  It's a different tire than the 4000s II, but its inflated dimension on a 52 closely replicates the inflated size of the 404 with a 4000s II 23 - the Attack 22 on a Rail is .2mm narrower than the GP4000s II on a 404, and about .6mm shorter than the GP4000s II on the 404.  There are tread differences, which are known to influence aerodynamic speed, but net of everything it's the closest we could get to being able to measure exactly what we wanted to.  This was part of a totally separate trip, primarily to do some totally different testing which we can't yet talk about, and we shoe-horned this bit in.  We ran the 52 first with the GP4000s II, and then with the Attack 22.  With this testing reaching the point of diminishing returns and threatening to turn into a bottomless money pit, we kept it to just the two runs.  Since we don't directly compare different tests from different days, we used the Rail 52 with 23mm GP4000s II as the baseline, rather than the Pacenti SL23 which was the baseline for the other round of tests.  Semantics, we know, but it counts.  



As predicted, the narrower tire gains some speed.  Whether it's the same speed gain we'd get from the theoretical GP4000s II that set up to the same dimension that the Attack 22 set up to, we can't say.  We're getting well into the realm of splitting hairs here.  

So now, the million dollar question: which tires should you be using?  We've now examined the relationship between inflated dimensions and effective pressure, and the relationship between aerodynamic speed and tire dimension, and the fact that people don't want to ride around on the road at 60psi, and a few things show up.  The first is that you should ride the tires that you like, at the pressure that you like.  There's no "right answer."  That said, it seems to make little sense that a 110 pound triathlete should choose a 25, on which she'd have to knock the pressure down pretty low to get a smooth ride (also taking into account the rims on which she's mounting them), when she could instead get more aerodynamic speed, better protection against pinch flats, and a comfortable and secure ride on a 22.  Conversely, a 200 pound guy doing performance recreational rides - group rides, gran fondos, etc - on beat up roads can choose a bigger tire to advance his priorities.  

Varying tire size, type, and inflation pressure gives you an array of tools to tailor what's between you and the road more specifically to your priorities.  Rather than try and drill down into making you feel as though there's a specific answer of what, how much, and how wide, what we've tried to do is to give you the tools and information to help you discover your own perfect ride.  


Steady as she goes redux

We knew that yesterday's blog would have some birthing pains, simply thanks to the new-ness of the data we presented, the lack of a vocabulary and nomenclature around it, and the simple lack of history around it.It hadn't gone through the crucible.  Everyone's familiar with seconds saved in 40k TT, but the quantitative measurement of the off-axis forces acting on a wheel is new stuff.  Heck, we've barely just settled on using "bead seat width" for the measurement of the distance between brake tracks inside a rim.  

Anyway, we had one bobble in how we ranked stuff, which knocked the alloys down off their rightful spot in the hierarchy.  If you take the wheel in isolation, having the Center of Pressure (CoP, think of it as the location of the push from a cross wind) ON the hub axis is ideal.  When you put the wheel on a bike, it's not.  The steering axis, which is the actual line on which your front wheel pivots, is behind the hub.  This is where we get into headtube angle and fork offset and trail dimension, but at the end of a long road somewhere around 4.3cm is a good number.  That simple change made the sniff test go from "something's funny" to "yeah, that seems about right."

The alloy-rimmed wheels are the only wheels which put the Center of Pressure (which I will inevitably call Center of Effort somewhere around 1000 times, because that's what it's called in naval architecture) behind the hub, in it's ideal place.  Why exactly that's the case is a subject we will leave alone for now.  So instead of using the hub axis as our zero point, we are using -4.3 as the zero point.  That is to say that a pressure that is centered 4.3cm behind the hub will exert no steering torque on your bike.  Any center of pressure behind that will turn the front of your wheel into the wind, any center of pressure ahead of that will turn your wheel away from the wind.  


We also had represented the Center of Pressure and CdA (coefficient of drag) both as separate and related values.  For the time being at least, it probably gets this conversation further down the track to emphasize their separate values, as we have on our shiny new graph.  

The emphasis falls more squarely on CoP for now.  Combining the two, with the steering axis correction, does nothing but flipflop the relative position of Rail 52 and 34.  

All told, this is a more correct, cleaner, more easily understood and digested presentation of the data, and a better starting point to the conversation than yesterday's.  

A couple of quick notes:

1. Zero yaw points are excluded.  The measurement formula doesn't work at zero because the formula math doesn't work with a zero in it, and at zero yaw there is no crosswind anyway.  The values for zero are all over the place, and we were told straight away to remove them (which we had done yesterday)

2. As with yesterday's graph, this is weighted per Tour Magazine's yaw-oocurence weighting for 25mph bike speed.  That may or may not be ideal.  The differences narrow down a little bit at wider yaw angles which are more represented at lower bike speeds.

3. These are all measured with 23mm tires.  Putting 25mm tires on does change things a little, but that's a jar we aren't opening for now.  

4. At the end of it all, there is precious little context around this.  How much of an actual "on the road" difference in handling is represented by the gap from worst to first?  There is just not the landscape to put that gap into context.  Where this group of wheels fits into the overall picture is unknown.  

5. The CDA figures (which are measured in m^2) have all been multiplied by 100.  This changes nothing in their relative ordering or the magnitude of the differences between them, it is a facility to make the chart easier to present.  

Soon enough, we will all have a common language and ease with this stuff.   Until then...





November in the wind tunnel: Steady as she goes

One of the chief aims of the Rail series was to provide a stable and manageable ride.  This was a major consideration in the 52 being the 52, and not the 58 or the 60 or the 62.  The 34 was in large part born from the market demand for a wheel that would be nigh on invisible to crosswinds, even though early 52 reviews were near unanimously positive in regard to manageability.  

The cross sectional design of Rails attempts to create a near symmetry between rim side and tire side.   Obviously, tire choice affects this.  We had a 23 in mind, aware of the size to which most 23s would inflate on our chosen 18mm bead seat width.  Different tires have different shapes, and many people use different sizes. Nonetheless, the general gist stands.

To date, crosswind stability has been measured subjectively and anecdotally, never directly measured and quantified.  Thankfully, a wheel company from Indiana had been pestering A2 for an actual measured match to their CFD predictions.  Recently, A2 completed the measurement apparatus and algorithms to provide these measures. When you test a wheel now, your data sets include two new columns -  coefficient of drag, and center of pressure.  

Ceofficient of drag, simply stated, is how much pressure is pushing against your wheel - how strong is the push.  Units of measure are non-specific, but linear - meaning that .20 is twice as hard a push as .10, and 2/3 as big a push as .30.  Center of pressure describes the position of the push, relative to the hub, measured in centimeters.  Center of pressure of 2.35 would describe a push centered 2.35cm in front of the hub.  -3 would describe a push 3cm behind the hub.  

To date, we've never seen any graphical presentation of this data.  It's too new a concept, so we've taken a stab at it, which we think provides a clear picture of the relative power and placement of the crosswind's push on each wheel.  We have once again used the Tour Magazine angle of attack weighting in creating this chart, but we have used the 25mph weighting.  Our reasoning for this is that as wind speeds increase relative to bike speed, the likelihood of wider angles of attack increases.  We expect and welcome questions about this information and presentation, simply because we want it to be easily understood.  

Here is a link to a page that allows you to calculate apparent wind speeds and angles for any given combination (be sure and use the second box, the first one calculates to true wind speed).  Be aware as you do this that a windy city will have an average windspeed of somewhere around 10mph (the calculator uses knots - 10mph equals 8.7 knots), as measured at that city's airport.  Airport windspeed is measured high off the ground in an unobstructed place, and will overstate what your wheels are riding in by quite a bit - like 50% or more.


To say the results pleased us would be an understatement.  As the initial test of the 34 was underway, I was busily prepping the next wheel to test in the work room and poked my head into the control room to ask if I was in a good mood.  Dave, A2's engineer and a man not given to subjective statements or value judgments, said the aero drag measurements were going right along, but that the pressure measurements should put me in a very good mood indeed.  As the 52 ran and the data came up, my mood improved even more.  


As you can imagine, we're excited to see that our consideration of crosswind stability in the design of the Rail has been confirmed with such excellent results.  


Full of air: tire inflation

Last week's discussion of how tire size affects aerodynamics set off quite a little bit of discussion.  We've provoked some big responses before, but nothing quite like that.  The one thing that we hope people started to think about as a result of it, other than the direct component of narrower tires doing better than wider ones in the wind tunnel, is the importance of measured width.  It's a big factor.  

Now, measured width is a bit of shorthand, what we are really referring to is the actual volume of a tire, which includes the height as a variable.  Height and width aren't in lockstep, as some rims actually hold the tire lower down within the rim, while some let the tire sit a bit higher.  To investigate this more fully, we measured inflated width and height of 23 and 25mm Continental 4000s II tires on every rim we took to the tunnel, as well as estimated what they would be on a representative rim of the old standby 14mm between the brake tracks.  

There is debate over what "counts" as tire volume - does only the inflated portion outside the rim's circumference count, or does the volume in the cavity count as well?  Fortunately, the variances there weren't so extreme that they threw things out of whack.  Our calculation was fairly rough and simple - average the width and the height, take the surface area of that circle, and call that overall tire volume.  To eliminate the debated "dead zone," we took 5/8 of the overall tire volume and called that "effective tire volume."  5/8 simply because we are measuring the "outside half" of the tube, as it were, and that's bigger.  As I said, a bit quick and dirty, but when you reference it against a bunch of other calcs, the way you peel that carrot doesn't amount to much in the wash.  

Using a law of chemistry called Boyle's Law, which simply states that for a given mass of a gas, if you decrease the volume then the pressure must rise, we normalized how much pressure a given volume of air would yield in each tire/rim setup.  The results are shown in the graphic below.


So what does this have to do with anything?  It shows that as you increase tire volume, in order to keep the same "buoyancy," you need to decrease pressure.  There are a lot of different ways to express buoyancy, probably the best of which is illustrated here - the wheel drop methodAsk 10 people what the ideal pressure is for any given tire and you are likely to get 20 responses.  The point we're making here is that tire volume is probably the biggest determinant of how much pressure you should use in your tires, and it varies by a ton.  Put 30 psi in a road tire and you are going to be riding around on the rim, put 30 psi in a 2.2" mountain bike tire and you are going to be bounced all over creation, put 30 psi in a cx tire and you are going to be pretty close to ideal (I know, I know, tubulars, lower psi, etc - I'm making a point).  How big your tires inflate on any given rim will have a big effect on how that tire feels. The 100 PSI default for road tires was established when rims were much narrower than today's. The chart shows that to achieve the same tire volume as 100 PSI on a traditional skinny rim, you should only run only 79 PSI on a set of Rails with a 23mm tire, and only 66 PSI if you've mounted 25mm tires on your Rails.