John Hall (Gwin's Mechanic) on Spoke Tension

What spoke tension are you all running?

Ive been running 15 on rear 27.5 and 17 on front 29.

Same questions John Hall mentioned apply.  Carbon or AL? DS or NDS?  Is your tensiometer calibrated? Typically, I shoot for just under the hoop mfg max which on many carbon wheels is ~130kgf... so I'm shooting for 120ish on the drive side.  When I'm lacing up ex511's... which prior to the reserve hoops, was the only AL hoop I will use.  I shoot for 115ish on the drive side where the max for that hoop is 122kgf.  

dang- dialed.  

more than 0 less than stripped for me

I have the Park Tool blue TM-1 tension meter, which uses arbitrary numbers like 15 or 22 instead of kgf readings, so I really have no clue what my actual spoke tension is. This is something I'd love to learn more about, because I'm never sure if I'm running enough or too much. Just kinda go by feel on how tight is just right, and match all the spokes on that side within 1 value of each other. I will say I run significantly less tension in my front wheel, and I've needed to re-tension, true, or replace my rear wheel many times since the last time I touched my front wheel.

Glad ya'll enjoyed the podcast. John was super interesting to talk to.

Yeah tool also has numbers and a chart that you’re supposed to cross reference with spoke material and thickness.

I'm running Aluminum E13 wheels. 

So my numbers are kind of arbitrary as well.  I’ve been using the tool just to keep them on a tension so each time I ride/race they feel the same. 

I'm also using the Park TM-1, but for more accuracy, I calibrate it at home every time I build a new wheel. 

To do so, I built a frame with scraps of 2x4 (saw this on a forum somewhere), drilled a hole at the bottom and at the top to insert threaded rods with hooks, then I suspend a 300kg portable hanging scale attached to a spoke. I usually use the same spoke and nipple as what I'm building the wheels with. Then I tighten one of the threaded rods until I reach specific kgf readings and I check what the TM-1 says. I like to do 70/80/90/100/110/120kgf readings with the TM-1, write down what it equates to on the Park scale. That way I'm sure that the tool is "calibrated" (instead of shipping it to Park and get it calibrated at an unknown tension) each time.

Then I make sure that the higher-tension side spokes max at 105-110kgf without a tire mounted (using WTB rims) and I try to get everything pretty much within 0.1 on the Park meter (between two small lines). 

I usually don't touch my wheels at all afterwards and the tension seems perfectly fine for durability.

By the way, I really enjoyed the podcast episode! Thanks and keep em coming!

I've never used a spoke tension tool. I just squeeze 'em. Or pluck 'em for tone. Motorcycle wheels are way easier. They tell you exactly how tight they are with a tap of the wrench.

The tension tool is cool. You can add some compliance to your bike and check the tension from time to time so when testing other parts you know that variable has not changed.

For moto I’ve used spoke torque wrench which works well but if the nipple and spoke are seized, you won’t get an accurate reading.

Great discussion! I will never go back to working on wheels without a tension meter. I used to build wheels for myself all the time using the hand-gauge method, but they would come loose really fast during hard downhill riding. It wasn't until I worked at a shop that I was introduced to the tension meter. It looked slow and inefficient, and I figured my hand gauge worked just fine, so we had plenty of discussions/arguments with the crew at the shop about it.

So one time I'm building up a wheel for myself at the shop, same as always, and once my "hand gauge" told me everything was even and tight, one of the shop techs measured all the spokes with the tension meter. They weren't even close! The way I'd been doing it for years (by feel) left the wheel with wildly uneven spoke tension. Now it takes me about twice as long to build a new wheel with the tension meter, but they hold tension and last way longer, so it's a happy trade off.

Seriously guys I hate it to be a hater, but... From the engineering standpoint spoke tension does not have anything to with compliance of a wheel in the same way that spring preload on your shock does not affect the sping rate. And this is not mentioned in the video is just strange to me. Hear me out.

In a functional wheel all spokes are preloaded. This means that in a normal riding scenario all spokes stay preloaded/pretentioned at all times. This means that only Young's Modulus (elasticity of material so to speak) and the spoke geometry (cross-section area of the spoke, spoke length and count) impact the wheel compliance and NOT the preload (as long as the spoke does not lose tension as mentioned above). If a spoke loses tension regularly and not only on the harshest impacts, you get your typical failures when the spoke breaks either on the j-bend or the thread. Also, it results in inconsistent feel since the compliance of the whole wheel drops from one moment to another, this can't be good for winning races. So the general approach would be to go to max. tension specified by the manufacturer of the weakest components of the wheel that you're building.

For those who are interested check this out: https://www.boltscience.com/pages/basics9.htm (a preloaded spoke is like a bolt). It is also explained in the book "The Bicycle Wheel" by Jobst Brandt.

Serious question: am I missing something? If that's the case, can anyone explain in engineering terms?

Surely even if they don't come loose, the spokes are always going to lose SOME tension. Is it useful to think of spoke tension as performing the same effect on the Young's modulus as sag does on a spring curve? As in it doesn't alter the curve but dictates where on that curve you start from.

Important too to remember that the loading of the wheel isn't just in the vertical. 

If the primary load path is through the hub axle, hanging from the top of the rim, what effect could spoke tension have on radial deformation of the lower portion of the rim when subjected to impacts? I always assumed spoke tension would have a bigger effect on elasticity of the rim than the spokes themselves.

Anyone ever tested spoke tension above and below the hub with a rider on board?

That's a fine assumption to make if you only ever ride in a perfectly straight line on perfectly flat ground without ever leaning your weight to either side.

 engineering.JPG

This trend is true to a greater or lesser extent of every spoke above the centreline of the hub in the left hand scenario, but true of every spoke above the centreline PLUS every left hand spoke below the centreline in the right hand scenario

I just zip tie my spokes to each other.  

I'd love to see a simple deflection test of the wheel.  Rigidly mount the hub to something- apply a load sideways- measure deflection.  Change spoke tension and repeat. 

I always thought spokes behaved like a preloaded bolt- same as milton mentioned.  

YESSS!!! Finally, the nerd battle I've been waiting for. We've got Young's Modulus, free-body diagrams, and even Dave Camp in the mix.

The simplest argument for the "spoke tension doesn't matter" camp is the "hanging hub" argument that's captured in the free-body diagram up above on the left, where only the top spokes are working to hold up the hub and the rider’s weight. I'm glad we blew up that overly simple explanation of how wheels work. As Atkisa said, "That's a fine assumption to make if you only ever ride in a perfectly straight line on perfectly flat ground without ever leaning your weight to either side."

You guys brought up another, much more compelling argument, that Young's Modulus is the only determining factor for stretch and compliance once a bolt has been preloaded. For example, preloading a coil spring doesn't affect the spring rate of the steel, just the breakaway force needed to overpower the preload.

Not to sound stupid here, but isn't the spring preload argument also the answer to this question? I mean, it’s true that spring preload doesn’t affect the rate of a spring, but that doesn’t mean it has no effect on a suspension system. Preloading a spring means more initial force is required to start moving the spring, because, in effect, the spring has already started moving. So wouldn’t preloading a spoke increase the amount of force required to start moving a static spoke in the system? The spring is experiencing force in compression and the spoke in tension, but effectively it's same same regarding preload, spring force, and YM, right?

For instance, in a suspension system you’d rather adjust spring rate than preload, but sometimes you can’t change the spring rate and preload can be be another way to adjust the system. That’s how DVO’s coil negative spring gets adjusted in their OTT system. It’s also how a lot of shim stacks get tuned in compression and rebound systems. Not the best way, but it still works. So a highly preloaded wheel with high spoke tension would take a lot more force to start moving, but a minimally preloaded wheel would start moving with less force.

In a perfect world, you would probably run the **perfect** spoke dimension to get the spring rate characteristic you want for wheel stiffness or compliance, but in the real world where no one is holistically tuning spoke size and butting profiles for different rims/wheel diameters/rider styles/terrain/rider weights, spoke tension seems like it would be a pretty good hackjob way to approximate that effect. That's my argument for why spoke tension still matters.

On a semi related subject: The clamp load of identically torqued bolts (steel shcs into tapped steel, so not exactly spokes but still) varies by about +/- 50%. It can vary much more than that if dry vs lube etc variable are introduced. I was told this a former manager of mine. And load cell data that I gathered on temporary aerospace fasteners for my old job varied about the same. What I'm saying is spoke nipple torque is not a good measure of spoke tension. Measuring spoke tension directly is far better. 

If you want to go all the way nerdy (and understand a little german), you can use the "spokomat" which includes a spoke tensiometer based on the resonance frequency of the spoke which is a function of its tension (and other parameters like its length and so on). Using your phone's microphone and a frequency analyzer app, you can build your wheel through its resonance frequency alone. It works pretty well (tested it myself) and gives comparable numbers to a regular calibrated spoke tensiometer.

https://radtechnik.2ix.de/spokomat.php

"So wouldn’t preloading a spoke increase the amount of force required to start moving a static spoke in the system? The spring is experiencing force in compression and the spoke in tension, but effectively it's same same regarding preload, spring force, and YM, right?"

Well, no, it is not the same. With a hub suspended in a rim between two spokes, one spoke pulling up and the other spoke pulling down, the preloads of both spokes are forces of the same absolute value but in opposing directions. In an equilibrium state, these two preload forces will cancel each other out. Therefore the stiffness of the wheel will not be a function of the spoke tension(preload) but only of the young's modulus of the spokes and their crosssection. As long as no spoke is losing its tension fully.

In a shock, the preload force on the spring is countered by a tensile force in the shock's shaft. To compress a preloaded shock, you will first have to overcome the combined stiffness of the loaded shaft (basically a very stiff spring) and the spring of the shock till the shaft does not experience any tensile force. After the shaft is in an unloaded state, the stiffness is only dictated by the spring rate because the shaft will not take any compressive force due to the shaft moving into the damper body. This case is comparable to a loaded wheel when a spoke loses its tension fully and suddenly the stiffness drops because the spoke can not take any compressive force.

At least that's what I came up with after a little time. Correct me if I'm wrong.

All great points. I think you're right about the wheel when all spokes are in tension, like in a simple vertical orientation rolling along in a straight line hitting bumps, but when we're talking about wheel flex and compliance, I think we're primarily talking about side to side (lateral) wheel flex and spoke compliance. In the case of lateral wheel flex, one whole side of the wheel is losing tension completely on the bottom half as the wheel folds over, like in the free-body diagram up above on the right.

In that case where one whole side of "the spoke triangle" loses tension, I think the amount of force required to move the rim laterally is going to be dependent on the "initial" spring rate of the spoke, which is a function of preload. And because the bracing angles on the side of "the spoke triangle" are so steep, one side of spokes is going to lose tension very quickly as the wheel deflects laterally. That's why downhill hubs and boost hubs have strived for wider and wider dimensions between the flanges since forever, or taller flanges in the case of the Hope Big-un hub or the old Hive hubs.

Adding preload to a spring does not change the spring rate, however, it does change the functional range of the spring. A 500 pound per inch spring on a 3 inch travel device with 0 preload requires a force above zero to start compression and a force of 1500 pounds to reach 3 inches of travel. the same spring with 1 inch of preload would require above 500 pounds of force to start compression and 2000 pounds of force to reach 3 inches of travel. The same would apply to pre tensioning spokes.

A light in the darkness! I feel there is a chance that we actually all kind of agree in a weird way.

 This overview from the following video https://www.youtube.com/watch?v=QXJd3ovqPhs&t=1701s was really helpful for me to visualize and understand spoke tension and its effect. In mtb and mostly gravity lateral load is mostly what matters and spoke tension doesn't affect the maximum displacement (rim stiffness, bracing angle, hub geometry, lacing pattern could effect maximum displacement) but make it so that a higher force is required to get that same maximum displacement.

thanks for this find.  

Soooo... spoke tension matters? [in the sense that it can measurably affect the force required to laterally displace a rim, eg in a bumpy off camber corner]

No- i think the graph and video is showing that varying spoke tension does NOT (measurably) change the amount of wheel deflection given the same lateral or torsional load.  The boxed in data points are not moving up/down on that chart (displacement on y axis)

This would agree with treating a spoke as a preloaded bolt.

Edit: I'd believe proper spoke tension can help durability but I'm guessing for us normal people no one could notice differences in spoke tension until it goes to 0.  

Spoke tension does not change the maximum displacement but it affects the force required to get to that maximum displacement. That is where the table at the top is useful.

I agree with the previous poster, I think the graphs only refer to maximum displacement of the spokes, not force required to get to maximum displacement. The table says it takes ~30% more force (1420 vs. 1810 Nm) to get to maximum displacement laterally and more than double the force (302 Nm vs. 697 Nm) to get to maximum displacement vertically when you compare a low tension steel spoke vs. a high tension steel spoke.

I haven't watched the video yet because, despite appearances, I still have to get some work done sometimes.

[EDIT] I just watched the video, and apparently the software is using a constant 100kg lateral force on all of the tests at all spoke tensions. So I have no clue what the "Maximum force" column in the graph represents, because the force applied should be constant across tests, and I don't think the simulation is calculating failure point?