tap

I'd like to thank the guys over at Denali Skis, and the local skiers who were extremely accommodating guinea pigs, for hosting and enabling some data acquisition beta trials last weekend. We've been working on a new approach to real time data collection that produced some really cool results. In a nut shell, we instrumented the pylon of the boat enabling continuous measurement of applied load and load vector direction. I'm now sitting on more data than I know what to do with. Our primary goal was simply to see if we could make it work. I'd say for a first effort it was wildly successful. I'm still working on sorting out my post processing methodologies, but here's a teaser sample of one of my 34 mph/28 off passes. The graph is on a time scale basis, not distance. The green is angular position in degrees of the load vector vs. time (it gets a little wacky when the load drops off but is pretty good as long as there's some load on the line). The blue is applied load in pounds. The course is plotted for time reference only. There is a tremendous amount of information that can be learned from this type of data. It allows you to be completely objective in dissecting your abilities. Caldwell was nice enough to crush a 39 off pass on one of his new 2017 model skis while we were recording, so we spent a couple hours just staring at an overlay of my 28 off pass with his 39 off pass. It was both humbling and encouraging at the same time. We spent about an hour just looking at the gate approach alone. There were some pretty big differences and not at all what I was expecting. Needless to say Caldwell definitely practices what he preaches in regards to the GUT theories. I walked away with about a dozen clear opportunities for improvement, some were extremely obvious and some rather subtle. We're still sorting out ways to improve the data collection process and tidy up the post processing, so I'll keep it brief for now. But I'll say it's already proving to be extremely insightful. Here's the same pass as above zoomed into 1 ball to 2 ball with an added overlay of an extremely fluid open level skier running 34mph/32off. (not Caldwell this time). As you can tell, there's a pretty big difference in how fast the open level skier gets on the line coming out of 1 ball, he's much smoother at getting on and off the line, he has a higher peak load, gets off the pull much sooner, and is sitting pretty for 2 while I'm still pulling long. I'm the blue/green data, the open level skier is the gold/bronze. @adamhcaldwell @AdamCord

And while you're at it, pull your set screws out and see if their sharp, if so lightly sand or file the knife edge off so they don't dig in to the fin. Good advice for any fin really, for some reason a lot of stock set screws have a knife edge which causes the damage and makes it really hard to make fin adjustments later on.

@chrislandy For sure, I just don’t think it’s significant, the numbers just aren’t there. To me the smoking gun is change in stiffness. It would be very interesting to hear your results if you do fatigue one of your wakeboards. Sounds like a great use for a wakeboard :) @Horton I’m not doubting your subjective testing, I think subjective insight can be pretty meaningful. I just think it’s worth keeping change in stiffness as a primary suspect. Just need to find a NIST traceable flex tester. As far as PVC vs. PU, everything else being equal, I suspect what you're feeling is a difference in mass. As you're well aware, you would typically run a little denser core with PU to make up for the lower strength properties, at least I would if I was making skis with PU. @Jordan All I can say for certain is some of my numbers have been higher than what was written down. It wouldn’t be unreasonable (not saying this is common, just plausible) to see an increase of a couple points in flex from taking a ski straight from the factory and letting it sit out in the Florida sun for a few days (black skis get hot!). I like to think of flex as a quality control number, i.e. a comparative number from ski to ski to catch any flaws or drift in the manufacturing process. If the numbers are all taken on the same tester after the same amount of cure time then the system should work reasonable well. Plus, I doubt anyone is paying to have their flex tester calibrated, I’m certainly not, so the numbers from one tester to the next could certainly be off by a few points. My point is, if you’re looking to measure degradation where a couple points in flex may be meaningful then make sure to take your own initial readings to compare to. And if possible, do an at-home calibration of your scale with a known mass.

@Horton @chrislandy I don’t know how I manage to waste so much time on this stuff, but here we go. To start with, I’ve never subjectively experienced a ski “breaking down”, so I’m not exactly certain what people mean when they say that. However, I can theorize how it could happen. Like most materials, the stuff skis are made of (fairy dust and unicorn tears) will degrade through fatigue loading given sufficient cycles and sufficient load/cycle. The mechanisms of which are varied, you’ve described a few. All of which would show up in a static test via change in compliance (i.e. flex number). The idea that resin micro-cracking has no effect on static properties is a bit far-fetched, when a ski is flexed one side is put in compression while the other is put in tension. Maybe… maybe… the tension side could be measurably unchanged with a Harbor Freight dial indicator (doubt it), but no way would the compression side not show up. The rebound theory, as I understand it, goes something like this: The time it takes for a ski to return to a non-loaded displacement after an applied load is suddenly removed has gotten longer (i.e. the snappiness has diminished). Is that about right? To keep this simple let us consider a ski in a 3-point bend, i.e. a standard flex testing configuration. The flex of a ski can be modeled as a linear spring, F=-kx (Hooke’s Law) where x is displacement, F is applied force, and k is the spring constant or stiffness. Anyone with a flex tester can validate this, simply record multiple flex numbers on your way to 0.1” of displacement then plot applied force vs. displacement. You will get a straight line, pretty much guaranteed. The slope of that line is k, the stiffness of your ski. If you get something other than a straight line then something is wrong with your flex tester, or you really do have a ski made from fairy dust and unicorn tears, or you just crushed your bevels… that sucks. This is the reason why the amount of preload on a flex tester, whether 20 lbs. or 50 lbs., has no effect on the final flex number (within reason). So now that we can think of a ski as a linear spring the discussion of rebound should be much simpler. If I stretch a spring and let it go it will accelerate back to its starting position and then continue moving in the opposite direction beyond it’s starting position until it reaches the same displacement only in the other direction and so forth, simple harmonic motion. The time it takes for a spring to compete one full cycle (stretched – compressed – stretched) is the period (T). So a “rebound time” could be considered as T/4. The period can also be expressed as a frequency (f) or 1/T. So the question is, what affects f? Fortunately, there is a simple relationship for that as well… f=(1/(2pi))*SQRT(k/m), where k is still the stiffness and m is mass. So, the only way to change f, and therefore rebound, is to change either the stiffness of the ski, or the mass of the ski. This is all for non-damped systems. For a damped system the level of damping does affect the frequency, but it is completely negligible for anything we’re talking about here. What damping does do is decreases the amplitude of how far the spring stretches/compresses each cycle, absolutely nothing to do with rebound time. We’re looking for a change in stiffness (flex) or a change in mass. If your ski happened to have a balsa wood core I would buy the change in mass option… but it’s been awhile since anyone has been ill advised enough to do something like that. It could be interesting if someone wanted to weigh their ski over several years of use, but I doubt you’ll gain much more than a few binding screws worth of mass (assuming it’s well constructed). So we’re down to stiffness. This is where I’ll defer to “Science”, as I don’t want to give up too much. But… the idea that you can degrade your stiffness to a noticeable point without a static compliance change (flex) is hard to believe... or that the core breaks down but somehow still maintains its static modulus. If you really want to get to the bottom of this, I’d start with this… the flex numbers written on your ski may or may not be the real starting flex numbers. I have measured a few where the ski stiffened up since it left the factory. So be careful what you compare your numbers to. And just for fun to put this all into context, an actual “rebound time" is on the order of magnitude of milliseconds, or about the same amount of time for the boat to travel roughly 1-2 inches down the course. Which gets me back to where I always end up with this type of thought exercise... how much change has to occur before you can truly tell a difference. I suspect you'll notice that your ski is flexing like a banana long before you notice a change in rebound time. P.S. Horton: “I am unaware that anyone has built a better flex tester but it can't be rocket surgery.” Instron. Just need someone to fund the testing.

@Horton I'll bite, what's your theory for how a ski can "break down" without any change in the flex numbers? This may be a setup, but I am curious.

@Krlee my wife is right in that range. I set her up on a 2015 model (v3.1) about three weeks ago. She is generally not a fan of changing skis or setups, but she's doing awesome on the new ski. Week 1 she was right there at her average, by week 3 she up'd her practice PB by a buoy. She's been skiing more and may have progressed regardless, but the ski really does look pretty damn good.

@Horton I know nothing of what you speak... confirm nor deny. @Mark_Matis that's good stuff. I apparently only got around 1 ball, so regardless of line length it was pretty weak.

Ha. Yea she's pretty decent. Probably the only reason she let me post the video.

I told her the Wakeye app was all setup and all she had to do was put it on the pylon... https://youtu.be/6Pn_wWKe84Q Too funny not to share, we both had a good laugh.

@MattP I'm a M2 guy. Full disclosure, I'm pretty good friends with these guys so I'm sure my opinion is a bit biased. But... results don't lie. I spent most of last spring/summer just trying to get consistent at -28@36mph, and fighting to get midway through -32@36mph on a really good day. I jumped in with some of the early trials and have gotten to play with a few different progressions of the ski along the way. Last tournament in the Fall I broke through -32 and got around 2 ball at -35@36. It's a different kind of ski, no question. The faster you go the happier it is. The settings make a difference. I wondered into no-mans-land for a bit with the fin and paid the price. But, with a few revisions based on some conversations with the Adam's we got it back into the realm of awesomeness.

Fun idea, but no. That's not really how that works. All you are doing in this case is moving the axis of bending towards the carbon side of the fin (instead of in the center as it would be for homogeneous materials like aluminum). It would still be just as stiff (for the most part) in either direction. And good luck getting it flat to start with (can be done but would be a big pain in the a..).

@thager The skis set for about an hour at temp. Plenty of time to make something happen. I checked surface temps with an IR thermometer.

I'm not sure which thread to post in now, but since this one has way more information it seems like the better choice. @Horton any way to merge threads? Since I seem to have inadvertently restarted this fantastic discussion it seems only appropriate to actually contribute something. The effect of temperature on the flex of the ski was brought up a couple times but I didn't see any definitive answers. So, I spend the evening out in the cold with the flex tester (cold for Florida anyway). Within the repeatability of my measurements (+/- 1 lbf.) I can say that from room temperature at 70 deg F to 45 deg F there was no measurable difference in ski flex. I did a sampling of three different skis (blank) at two different flex points per ski at 25" and 41". The skis: 1.) lower end ski with polyurethane core and fiberglass/epoxy skins 2.) an older top end ski with a polyurethane core and carbon/fiberglass/epoxy skins 3.) a newer very top end ski with a PVC core and all carbon/epoxy skins All three skis were "big brand" skis. I allowed the skis to acclimate to temperature for 1 hour prior to taking measurements and took all measurements at temperature. Testing flex numbers ranged from 110 - 210 lbf. It'd be great if someone else wanted to validate (or contradict) my results, but as far as I can tell temperature has no measurable effect on ski flex (within normal use temperatures). And for what it's worth, I'm in the cold water has more drag and viscosity is the cause camp. And I did take a fluid mechanics class at one point in time, but I have no functional memory of it.

@AdamCord thanks. somehow I completely missed that thread.

I was skiing this last weekend and got into a conversation regarding why it felt like such a struggle (probably just an off day). Somehow we ended up on water temperature and viscosity. So... I looked it up. At 90F the dynamic viscosity of water is roughly 0.8*10^-3 Pa*s. At 60F it is roughly 1.15*10^-3 Pa*s. That's a 44% increase in dynamic viscosity! I am in no way an aerospace engineer, so I can make really bold statements and clam full ignorance. What does this increase in viscosity do for the force of drag??? The easiest theory I can find (and here is my ignorance) is the Stokes' equation for spheres traveling in a fluid where Drag Force = 6 * pi * mu * R * V. Where, mu is the fluids dynamic viscosity, R is the radius of the sphere, and V is the velocity. Using the Stokes' equation, a 44% increase in dynamic viscosity equates to a 44% increase in the Drag Force!! That all assumes pure laminar flow, small Reynolds number, and so forth (which I'm sure is a gross oversimplification). Any aero/hydro guys/girls out there that can provide some better insight? For now I'm thinking of changing my wing from an 8 deg to a 5 deg until Spring gets here.

A go to source for all things G10: McMaster-Carr It is sorted under Garolite in their Raw Materials section. Or just use the link: McMaster They are a distributor and do an excellent job if you are looking for less than full sheet sizing. FYI, G10/FR4 for the purpose of binding plates is equivalent to standard G10, the FR4 is a designation for flame retardant. Also, G10/FR4 is often less expensive than standard G10 (supply and demand). There are lots of other good sources our there as well. Look for certifications to MIL-I-24768 to ensure good quality material with minimum strength requirements.

I love this sort of thing, but you lost me with the title of the thread. Can't tell you how many times I passed it over before I finally clicked on it. Add a load cell to the setup and then you have everything I was looking for. Can't say I'd ever pay money for one, but would be cool to get measured along with a coaching set, assuming the coach had a gold standard to compare against. I put a load cell in line with the rope a couple weeks back, I know it has been done before but it was still fun. I'll post the data if I get a chance this weekend.

@Than_Bogan , what's your shoe size? I'm working on a binding placement theory and need some more data points.

@Than_Bogan , I'm keeping the setup and analysis as simple as possible. I strip the ski of all hardware (bindings, plates, fins, fin blocks) then fixture it in a simply-supported condition at both ends, stick an accelerometer to the mid span, then give it a whack! I convert the acceleration data to the frequency domain using a fast fourier transform to find the natural frequency of the ski. Right now all I'm doing is tracking the natural frequency. All else remaining constant, a 1% change in frequency (Hz) equates to a 2% change in stiffness (EI), modeled as a simple uniform beam. What is neat about the frequency assessment, in theory, is that it measures the entire ski with one single number, which is great for monitoring the health of a ski. The accelerometer I use has a pretty high sample rate, but if you get really bored just about all smart phones have a built in accelerometer. There are free apps to capture .csv data. Just remember, if you strap a cell phone onto your ski you are changing the mass, and therefore the natural frequency. The key to the analysis is consistency in the setup.

The setup and data collection for dynamic stiffness can actually be very simple (more so than a static flex tester), the data analysis on the other hand is much more interesting and requires about a sophomore level understanding of undergraduate physics. I think the question is not so much the how, but whether or not it brings any value (yet to be determined). @Horton is very much correct, in general static stiffness is not dynamic stiffness. Although, I suspect for this application that a knockdown in the static numbers would be a pretty good indicator that the dynamic numbers are also changing. I measured a few dynamic baselines at the start of this season on a couple of my skis. I am planning to measure again at the end of the season to see if there is any variation. I may have some data to present then, but for now it's all conjecture. I love the analytical side of the sport as much as the physical side. I do it just for fun because I like the science of it, but then again I'm quite certain I'm the odd one out. Most of the time I just get the deer in headlights expression when I wonder too far into the science. The problem with analysis seems to always come down to measurement resolution and establishing a threshold. If I measure at the beginning and the end of the season and find no measurable difference then are you convinced that nothing has changed, or simply that the instrumentation was unable to measure the change? And... even if I am able to measure a change, how big of a change is needed before it truly affects the performance of the ski? Maybe someday I'll have a large enough data pool to draw a general conclusion. Until then I say try a new ski, if you love it and can afford it then buy it (but don't throw away the old one, donate it to science!).

Flex test the two skis. If you've checked everything else flex is next on the list. Something could be wonky.

For those wondering how much the riser plate can change your flex numbers... I use a reflex front with a standard toe hold for the rear. I added a constant thickness riser plate under the rear toe plate only. Everyone has their own theories and reasons, mine was simply to make getting forward on the ski a natural position instead of a forced position. Regardless of why you would decide to try this, I thought I'd provide some insight into the change in flex of the ski. (I pretty much just wanted an excuse to play with my flex tester). The riser plate I am using is 0.345" thick (constant thickness) with no attempt to make it more flexible, i.e. a solid plate. The plate is a Walmart plastic cutting board. I removed the front binding to make the testing dependent only on the rear binding setup. I flexed the ski at 22" from the tail, I chose this simply because it was as close as I could get to the center of the rear binding. All measurements were taken at the exact same location. With a blank ski the 22" flex number is 100 lbs. even. With a standard toe plate the flex number is 104 lbs. With a 0.345" riser under the rear toe plate the flex number is 115 lbs. That's a 10.6% increase in the flex number from toe plate to toe plate with riser. As I said, this is just a base line to provide some insight. The thicker the plate you use the more it will change the flex number. Also, I don't know if making the riser plate more flexible will make the assembly any more flexible because the aluminum plate from the binding is most likley doing most of the work, the riser plate is acting as a core material. But the thicker the riser, the further the aluminum plate is from the neutral axis of the beam. If I find some more free time I may try putting the riser on top of the plate similar to @davemac 's setup to validate the theory. Or cutting some slits in the riser and seeing if that does anything. I have no idea how much flex needs to change before the skier can notice anything. However, I will say that I did have to make a fin adjustment with the riser plate in place. I have no idea if that's attributed to the change in flex or the change in my natural position on the ski. And for what it's worth, the new setup is working well for me.

So I've been putting some effort into finding a means to track skier path. I first tried latching an accelerometer to myself while skiing, which worked great for acceleration data but not so well for calculating velocity and position data without ridiculous error in the numbers. Next I played around with acoustic location (which I still think has promise), but it required an annoying emitter to be worn by the skier. I read quite a bit about GPS location but, at least for now, I don't think it's quite able to provide clean enough data for what I'm looking to do. Then, primarily from reading about the work Bud Davis was doing, it occurred to me that motion capture was the answer. I took what Bud was doing and moved it a bit further into the digital world. I mapped out some geometrical reference points on a boat, and grabbed a video that was shot from the pylon. Using pixel location I was able to track the skier's center of mass, handle, and front foot. The graphs below are from an open rated mens skier skiing 32 off at 36 mph. The process is fairly manual for now, but seems to provide reasonable results. Pay attention to the labels and units. The first graph is the angular position (in degrees) vs. time. The second graph is angular velocity (in degrees per second) vs. time. The third graph is a bit of a stretch, but I think it is rather informative. It is a graph of the net force on the skier (assuming a 190 lb. skier) in the direction of the center of rotation (the pylon). It is reasonably equivalent to the line load, at least for certain portions of the path. It should be reasonably accurate from center line to apex. From apex to center line I would expect the actual line load to be slightly higher than what's calculated. I could translate the polar coordinates (reference frame of the boat) to Cartesian coordinates (reference frame of the course), but frankly I don't see the point. As far as I'm concerned the angular data is way more interesting (although I reserve the right to change my mind).

The physics are sound. If it all goes south I'm claiming operator error.

@gsm_peter‌ , if I read your post correctly I think what you are describing is a "wet bag" technique. A similar process to what @AdamCord‌ is getting after, but different. Wet bagging generally implies wetting out the fiber during the layup process then using ambient pressure, via use of a vacuum bag and consumable stack, to consolidate the layup and press out air and excess resin. @AdamCord‌ is hinting at using more of a vacuum assisted resin transfer process whereby the layup would be constructed dry, then the resin would be forced through the laminate using ambient pressure, via use of a vacuum bag or vacuum cavity.