Tuesday, November 17, 2009

busy month

Been pretty busy the last month or so. First went up to Edmonton for a week of work at Sulzer service center. Nice service center, nice people. It was cold (below freezing) and yes snowed as well. Very car-centric. Would have been nearly impossible to go for a walk.

Went to the huge indoor mall in edmonton. The indoor roller coaster was quite impressive! But the mall itself.. I don't know... the whole concept of it seemed somewhat depressing. I did bring ice skates and went skating for an hour at free-skate.

I flew back into PDX on Fri and flew out again that night for Chicago and Class reunion in Michigan. The reunion was great! I was able to visit with a lot of old classmates and teachers. It was hard to believe it had already been 20 years, some the teachers hadn't seemed to have changed. Saturday night we visited at Love Creek Nature center. A lot children.. :

I had forgotten how nice, tranquil and beautiful berrien county is, very little change in 20 years, which is good (although there did seem to be more vineyards). I managed to drive around on some of the roads i use to ride. Brought back good memories of the beautiful fun cycling we use to do. I also managed to visit with big Al Muldoon for a couple hours in St Joseph. He still riding very strong and look as young as he looked 20 years ago.

Baby-sat Harriet's children for a morning and carved pumpkins for halloween. They both were non-squimish and got their hands right into it!

Installing new windows in the house. doesn't take long, but what a mess !

Modified loose tight for riding to work. I really like these.. They were a very loose and are a very thin nike dri-clime ? material.. Like some warm- up pant. By sewing in a knee patch and by sewing up the side to make them less loose, they make a perfect riding tights / pants. Very light, and thin they can't hold much water. since they are loose they are much warmer than a on-the-skin-type tights, but cooler on a warmer day as well. Also they come on and off much quicker then a normal tight making them perfect for comutting.

Right now I am down in La Porte Tx and another Sulzer service center. Nice weather, but wow talk about car-centric living. Makes me happy to be in portland where at least there is more freedom !

2014 comments:  the house is sold and we are now in Hillsboro in a rented "townhouse".   Those damn tights / pants  I still have and still use ... :)

Friday, October 9, 2009

Cannondale Fork mod

Recently picked up a cannondale racing frame and fork off of ebay for $230. It is light, by far the lightest bike i have ever owned. It is fairly stiff also. It had some issues: slight misalignment issues, wore-out dropouts, a bunch of corrision spots under the paint. But, for the most part a well made frame that rides nice and is fun to ride, and will make an excellent racing bike..

It came with an integrated headset. It seemed like something was missing, like there should be some sort of "fork plate" (some called it a "base plate"?.. ). No-one could tell me, nor could i find anything that said or showed what needed to go there. Most said the Bearing just goes right on the carbon fiber.

The engineer in me said no-way, poor design.. i don't like high point loads into "plastic"... So i managed to make an aluminum "base" on my lathe. It was very delicate machining. Had to keep it thin and get the angles correct. Came out nice ! I took some short carbon fibers (made by "trimming" carbon fiber cloth), mixed with epoxy and used that under and in between.

I also added slight shim from a soda can.

2014 comments.  The baseplate is still on the fork.  the soda can diametrical shim ?  silly :)  The frame and fork I still have.  A few more dents and dings...:)  Its been hanging up for a year or two, but its about to come back into service....  

Scraping House

Here's what i was up to in September. Takes way too much time! Looks really nice now with paint on it though !

This side was really bad, and the paint was really thick and cracked, so I took it all off. I have two huge garbage bags full of paint scraping just from this side of the house and only about 2/3rds up. Some of the siding needed repair also...

The rest of the house will get a normal scraping and paint!

2014 comments.  The house was sold last summer, and where I "scraped" with the Infared heater, and repainted: the paint was still perfect !  So probably worth the extra work in the long run.  By the way that was Fir siding not Cedar.. 

Thursday, October 8, 2009

Fixing a Rear Dropout

Fixing a rear dropout does not require replacement. This method is quick, and precise and strong. Timewise, shouldn't take more then 15 minutes to 30 minutes. This one took a little longer because i also drilled out the helicoil thread repair and brazed in a proper 10x1mm "sleeve" made from a nut.

Here is a pic after i straightened it best i could. I considered welding the crack as is, but i wouldn't have been able to get it straight.

Here i broke it. The big advantage then was i was then able to fix the thread and also face it in the lathe.
Next came proper alignment achieved by using two axles, nuts and visegrips.

Next i simply stick welded it with. Weld very hot and fast, chipped slag, and hit it again, and repeated 3 or 4 times until it looked good. Filed it flat (not shown) and was done. Will now work perfectly. This will be my sweethearts bike someday.., an old but cute and sturdy Bridgestone RB-1 road frame that i picked up for $40..

2014 comments.  As I look back at some of the things I did, this one was actually really good work, a great quality repair... .  I ended up not doing anything with this frame, and finally sold it, and hopefully it is being ridden around by a Portland hipster :)  

Monday, August 17, 2009

Tandem !

Bought an old burley tandem for $400. ~15 yrs old ? but lightly used and sensibly made (i like the internal shifter cables..). Fairly heavy, but sturdy.
I added my new tandem wheels (the original tandem wheel had a bent axle even from light use). I re-machined the front bottom bracket eccentric to take a 110 mm spindle rather than the bowlegged length of 126 mm.

Added V-brakes which of course are huge improvement over cantilevers, especially for the rear brake (with its long cable). [Note must use "v-brake" road levers and not say a travel agent with normal road levers, to get the advantage of less force in cable and subsequent less "sponge"]. We did a couple max-stops and yes can almost slide the tires ! Its stops like single racing bike. Kool stop pads of course. The rims are a fairly inexpensive Alex Rim, wide but 30mm deep , so plenty of aluminum for a "heat sink" or at least i hope. Panracer 35c tires; way wider than i expected. Should have went with 32 or 28.

We did the Portland sunday Parkway and rode up mt tabor. Felt great both going up and coming down. The steering is a little strange and will take some getting use to.

2014 comments.   This was a great tandem.  It is sold now and we have new Tsunami frame, which I also built new wheels for (which I will do a new post on...)

Tuesday, August 4, 2009

Shelves and Shaper

Been making some nice garage shelves. The secret to making them fast and accurate is to Dado cut (shelve slots) the fours corner posts all together. In this particular case i had a 2x8 that I slotted, then i ripped into 4 pieces. With subsequent bracing makes for a lightweight and strong shelves.

Finally got the metal shaper body hauled away to Metro Metals. 740 lbs. Using a pry bar, 4x4's and jacks and a cable winch it took an hour plus to load without injury... The tailgate barely handled it, and is now slightly bent.

It was unloaded in all of 5 seconds by a big magnet (which for a second or two, when only holding on a corner and about 5 feet above, thought was going to let go and really bend up the the pickup..)

2014 comments:  still have the shelving, and since have made _a lot_ more shelves (especially for the new rented townhouse that has no built-ins...).  The Pickup has long been sold, but was of course a great little vehicle. 

Saturday, July 11, 2009

Crank Brothers Eggbeater Pedals

If the little bearing on end of the pedal fails the pedal body can come off the spindle. The rest of the ride will be one-legged (if your lucky, if un-lucky a bad crash and ambulance ride) . Now i have had this happen to me, but not with an eggbeater, and I was lucky and didn't crash.

To help prevent this from happening an extra bearing can be added. In addition to another bearing (686 sealed, vxb sells cheap), Will need a longer 4 mm bolt (local hardware) and a washer overlapping the inner seal lip, helps shield from the elements (since rubber end cap now can't be used..). Place 2nd bearing in after clip. Note will not work with all eggbeater spindles. (some use nuts rather than bolts to hold bearing in / hold pedal body unto spindle. )

Also will need to use thread lock or have the bolt bottom out perfectly without putting too much preload on outer bearing. Also should probably have a washer in between the two bearings the same or slightly greate thickness than the snap ring clip. Ok possibly if you get the spacing correct on the in-between-washer (not shown below), you might not need thread lock, etc.

2014 comments :  Hmmmm.  I havent' done this in years..  I don't even know if its possible . I am still riding Candy's though. 

Hauling Steel

Been taking steel to Metro Metals. They have a pretty nice and fast process and helpful and friendly people. Weigh on the way in, dump steel, weigh on the way out, they print a check a give to you.

This was $50 worth of steel,(I added a couple more items, and had 700 lbs of steel !)

Disassembling the metal shaper, still need to get the body hauled. Maybe next weekend if my little truck can handle the weight.

Building Tandem Wheel

Some pictures of working on tandem wheel.Here are the spokes, the nipples, the washers, nut driver. The nut driver bit had to make myself, with a file, brazing and then turning on the lathe. It is square to fit over the end of the nipples. The nipples took a long time to prep. I had to lathe-turn each one to get rid of the slot and to put a slight curve to them.

Here is a picture of the Wheel half-laced.

And here is a picture of the nipple (flipped) and washer inside the rim, to help spread out the load. Higher spokes forces are possible with lower peak stress onto the rim and nipple. Fatigue life should be improved with a stronger wheel. Also improved aerodynamics with hidden nipple. Fortunately with these rims, the rim bed was nearly flat for 9 mm od of the washers, So the washers are only slightly taco shape. The ideal is to have enough surface area in between the under force components that everything stays elastic.

2014 comments. since been sold.  New tandem wheels are only 24 spoked, but built using the same principles... 

Spoked Wheel - Part 4 - Elbow Strength

From previous post , the tensile strength and cross section areas are
233 ksi (2.34 mm) 0.0066658 in*in
240 ksi (2.0 mm) 0.0048695 in*in
245 ksi (1.8 mm) 0.0039443 in*in

Then it follows that:
The ultimate tensile load of 2.34mm spoke is 1553 lbf
The ultimate tensile load of 2.0 mm spoke is 1169 lbf
The ultimate tensile load of 1.8 mm spoke is 966 lbf

The spoke is predicted to break in tension at those forces. But for infinite spoke life, what we are interested in is is fatigue strength. For ductile steel, fatigue strength (or endurance limit) is approximately 0.5 of tensile.
Taking the elbow as the limiting factor, which we assume to be in shear we want the shear fatigue strength. Using distortion energy (von mises) failure theory, this is 0.58*tensile fatigue strength.
To summarize the max force that should be put into a spoke:
2.34mm: 233*0.5*0.58*0.0066658 = 450 lbf (2004 N, 204 kgf)
2.0 mm : 240*0.5*0.58*0.0048695 = 339 lbf (1508 N, 154 kgf)
1.8 mm : 245*0.5*0.58*0.0039443 = 280 lbf (1247 N, 127 kgf)
What this means is that for a spoke that has 1.8mm elbows, if the force in the spoke is kept under 280 lbf, the spoke should have infinite life.
In the elbow, the force transitions from an axial force into a shear force . Someday I may model this with FEA to see the stress contours. If there is bending stress in the elbow, which is definitely possible from not adequately stress relieving the spokes, then the stress in the elbow could be above the fatigue strength, and thus infinite spoke life will not be achieved.

Also note that just because this much force can be put into the spoke, doesn't mean it should be. The rim and nipple strength might not match. In the days of box section rims, the amount of spoke tension was determined by tightening the wheel's spokes to within a half turn of the onset of rim buckle. With today's deeper profile and/or sturdier rims, buckling is not the limiting factor. I have only experienced it with a very light box section rim, long ago...

Also of course, in the future, will look into the nipples and rim. The goal is to have as highly tensioned spokes as possible with infinite fatigue life of the spokes, nipples, and rim.
2014 comments:  Ahh this is good stuff. I still use this, my own blog post,  as reference :) 

Spoked Wheel - Part 3 - Fatigue Cycles

How many revolutions should a wheel last (need for fatique calcs)? Infinite is great of course, but what is reasonable? 10-20 years ? Rim-braked rims wear thin of course, so... For now I will go with15 years. Say 25 miles per day. 9125 mile per year. 210 cm (0.00130488 mile) circumference wheel (700x23c). 9125 miles/0.001305 miles per revolution = 6,992,882 revs per year or ~ 7 million wheel revolutions per year.
In 15 years, 105 million revolutions.

For a racing wheelset: 2 or 3 races a month, maybe 200 miles for 9 months a year. 1800 miles a year. Keep only 5 years. 6-7 million wheel revolutions.

However, typically the knee for fatigue life or endurance limit for steel is right around 1 million cycles. In other words, if the steel component being dynamically stressed doesn't break after 1 million stress cycles, it probably will have infinite life.

One issue is that Aluminum does not have an endurance limit.
Summary: Wheels see enough stress cycles that they should be built for infinite life.

Saturday, June 27, 2009

Inexpensive Tandem Hub

Since I am too cheap to buy an expensive tandem hub (at least for the cheap tandem I am working on), i modified a track hub. Dual threaded for freewheel and drum brake. The Miche hub worked nicely for several reasons.
1. More threads, more nicely done, than other track hubs. As many threads as on a normal hub. I don't want to "strip" the threads with the extra power of a tandem.
2. 12x28 bearings. The bearing are large enough to handle the extra weight of a tandem, while still leaving enough aluminum thickness on the hub, and allow a 12 mm axle. Any other size and you in danger of burning out bearings, breaking the hub shell, or bending axles.
3. Allows the use of an inexpensive but strong freewheel. I know freehubs are great, but since the ratchet pawls are at a smaller diameter, they can't take the same amount of abuse a freewheel can. Obviously a phil wood hub would work great, but they are over $400.

Made an axle for the hub. Not hard to make, other than having to be ultra careful to get the axle to bearing a slight slip fit. The non drive end I made of aluminum. The axle ends, still needing a trim, I silvered into thick tubed 1/2" 4130 tubing, that i then machined down to 12mm.

2014 comments..  Hmmmf.  looks great, but didn't work out in the end.  My favorite tandem rear hub now is siply the $40 or so Shimano 29er (629 ?)  cassette hub.  With 10mm steel axle.  Has held up great for us.  Drum brake ?  sold long ago.  Just not necessary for us..

Friday, June 26, 2009

Pa and Oh visits

Been in Allentown PA for interaction / training with the Pro Pump Services guys. Been here nearly two weeks. Visited two Powerplants, a foundry, and our Sulzer service center. Here we are at Cogentrix plant in Northampton that burns waste coal (basically coal that was "dumped" and is sitting in piles around Pennsylvania).

Here we are measuring vibration on a booster pump (a small pump, yes relatively, this is a small pump). I have the blue hardhat on.

I managed to make a quick trip to Ohio for the weekend to see Joel Sharon Caleb and Cameron. Here are the boys playing golf. A lot of fun. Here Caleb is getting the ball back to the green after launching if off the tee well beyond the the green and into a hay field. This was Cameron first time swinging a real golf club and he did quite well! Even managed to skip the ball off the water.

2014 comments :   These guys are big boys now ! 

Friday, June 5, 2009


With the new digital TV, I was able to watch the Giro. I noticed slideouts occurring on corners with dry pavement. Watching the riders go down, it was apparent what went wrong: Not enough weight on the front wheel.

If you watch carefully, just prior to the front wheel slipping, the rider's weight is back on the saddle. One can find all kinds of information on how to corner, the majority of course good advice, but I have yet to see emphasis on, or to keep awareness of the weight balance between your wheels. I do have personal experiences as well (e.g. a crit in michigan in 1993? second to last corner my front wheel start sliding. In hindsight I am pretty sure I was pedaling/pushing so hard that I also pushed back into the saddle. I did stay upright, but scared the guy behind me .)

With a tweaked front geometry and by coming out of the saddle a little on corners, I have found I can corner better, and have hit nearly 60 mph on roads in west hills here in portland ( only possible if you stay in a full tuck, don't touch brakes in corners, and of course weigh as much as me...)
From BLOG Pictures

If you take a motorcycle training class, quite a bit of emphasis is put on cornering. One key point is that in most situations if you lean the bike over (done by an initial countersteer) and look into the corner, the bike will turn. And that speed (within reason) is not the issue, if you lean it, it will stick. (Assume dry pavement...I am always wary of wet roads ! )

The same with a bicycle. If you lean it, look into the corner, you will stick. But that assumes you have the front wheel properly loaded. It is easy to transfer weight to the front wheel, come out of saddle and put weight on the pedals. Inside pedal, outside pedal doesn't really matter, likewise, what is happening to your weight side-to-side, or up-down doesn't really matter compared to your fore-aft weight. Weight on the pedals means most of your weight vector is now at the bottom bracket which is nearer the center of the bike, and viola, with the rest of your weight on the handlebars, your front wheel will be sufficiently loaded with weight and slideouts should not occur. (on dry pavement and assume no pedal clip.)

Saturday, May 23, 2009

Large Pulley

Ever wonder what is up with the large pulleys recently showing up on the inexpensive shimano derailluers? Decent engineering. Fairly common knowledge since the 18th century days of mechanical clock making*, to minimize friction loss, the ratio of axle diameter to wheel diameter should be minimized. ( e.g. the axle/bearing diameter as small as possible, the wheel/pulley as big as possilble).

For rotational systems, power (or power loss, in this case) = torque * rotational speed. Torque = force times a lever arm. If the lever arm, or radius of axle approaches zero, well of course so does the torque and so does the power loss. Or for identical axle diameters, if the Pulley diameter increases, the rotational speed decreases, and thus also power loss.

10 tooth ~ 32 mm dia
11 tooth ~ 36 mm dia
13 tooth ~ 44 mm dia
15 tooth ~ 52 mm dia

So by going from 10 tooth pulley to 15 tooth pulley the friction loss is nearly halved (to 60% 32/52)
Another advantage is the chain isn't flexed as much, albeit not saving probably a huge amount of energy, but some nonetheless.

Obviously if you have frictionless pulley bearings (i.e. standard deep row ball roller bearings), then this is not as big of deal. Yet it still applies since many bearings have tight rubber seals. I have seen where this seals have way more friction than standard bushing pulley, and regardless of what is claimed, they don't "loosen" up all that quickly (these are not electric motors spinning at 6000 rpm).

So if you are going to use sealed pulleys, pick those with smaller bearings, where the seal friction acts at smaller diameter. I personally use shielded bearings, but for a lot of wet weather or washing (..ccx..), it might pay to use simple bushing'ed large diameter pulleys.

On subject, how about Ceramic bearings for Pulleys? Not logical, since the way ceramic bearings save energy, is that they don't deform as much under load (hertzian contact...). And if the load is next to nothing to begin with, such as in a Derailluer pulley, then the relative energy saving will also be nothing. If they seem like less friction, it is probably only because the seals are really light. Don't waste your money!

*Book about John Harrison: Longitude: The True Story of a Lone Genius Who Solved the Greatest Scientific Problem of His Time).

2014 comments.  Friction Facts now has tested pulleys (among other things) and actually measured this.  Go to their site and support them !

Monday, May 11, 2009

Chamfer Pedal and Crank

The pedal to crank joint is a pretty poor design.  Obviously thought up before bolt joints/preload was understood, nonetheless it is functional.  For example, a 9/16" bolt can easily be preloaded to around 5 tons of force! Which one should not do or you will deform your crank threads (or maybe even make a miniature slinky).  So the 9/16"  axle size is obviously chosen for stiffness.   Yet if not fairly snug, movement and fretting can occur and your cranks may form a crack and fail.  It doesn't help that a sharp shoulder is digging into the aluminum.

To digress, a better design would simply use the pedal axle as a "nut" and run a bolt from the back side (with a proper washer) sandwiching the crank.   a high grade 5/16 bolt would probably be adequate with around 3000 lbf of preload...

Now, I have never had a crank fail at the pedal threads.  However evidently older campagnolo cranks sometimes would fail.  I remember reading Jobst Brandt's postings (5-10 yrs ago..),  his fix for this (he had cranks that failed)  was to simply  to chamfer the interface so that the pedal force would no longer be forcing threads or a sharp shoulder into the aluminum, but rather a chamfered smooth surface.   Good idea !

My main reason for chamfer is to move the pedals in closer for reduced q-factor (bowleggedness), and thus increased bio-efficiency (ok not proven), increased aerodynamics (less "width"), increase cornering clearance, and just general comfort.  Nothing worse then pushing down and to the side when you pedal, I like to push straight down !   

This also why I am still using square taper spindles..  can mix and match for less "q-factor".  Fortunately my feet line up straight and my heels don't hit the chainstays.. 

I bought a little 45 deg dovetail cutter.  Using old pedal axles, made a bushing that is first installed into the crank, and then the dovetail cutter is put in and turned to cut a concentric chamfer.  I set it up in the lathe with a spring for smoother and more controlled cutting. No I did not turn the lathe on!, only hand turned the chuck.  
The axles were removed from the pedals and cut with compound set at 45 deg. 

2014 comment:s I still have these tools and will do this.  Not necessary, but if I can get away with a lower q-factor than why not... 

Tuesday, March 24, 2009

Spoked Wheel - Part 2 - Spoke Material Properties

Spoke Material and Properties
From spoke manufacturer's websites:
Dt spokes: 18/10 stainless
Wheelsmith: 304 stainless *
Sapim spokes: 18/8 stainless. Strength = 1080(157 ksi)-1180 mpa (basic spoke, middle section) , Strength=1350 mpa (196 ksi) (normal double butted spoke, 2/1.8)
"Strength" unfortunately don't know if this is Tensile or Yield Strength.

*(304 ss is 18-20 Cr and 8-10 Ni, so basically we have identical spoke material..)

Annealed, non cold worked, 304 is not that strong. From one material table TS=75 ksi, YS=30 ksi.

But looking in machinery's handbook (25th p384), they do list the Tensile of 18/8 wire:
2.34 mm TS = 233 ksi
2mm TS = 240 ksi
1.8mm TS =245 ksi
1.6 mm TS=251ksi

This is really quite high! I have to admit, a lot higher than I expected.

2014 comments:  I now use this, my own  blog post,  for reference :)

Saturday, March 21, 2009

Bike Projects, this and that.

A few projects I have worked on in the last couple months:
After quite a few other methods tried, by far the easiest and fastest way to take a old steerer tube out of a fork crown. Use bimetal hole saws.  Started with a 7/8 " ended with a 1".  Took only a few minutes. Now it is ready for a new steerer tube. 

There is a story of welders needed up high on a water tower.  The job was advertised as no certification required, and a long line of people formed to get the job.  However, it thinned considerably when it was learned you first had to weld the seat on which you would be sitting on, 50 ft or so in the air.   This is similar,  don't do it unless you fully trust your welds.   

I used an arc weld, 1/16 7014 rod.  Stick weld?  Yes.  It's what I am good at and its very fast as well.  And very reliable. (see http://www.weldingtipsandtricks.com/stick-welding.html )    The setup was perfect.  No gaps in between the tubes, and a slight chamfer.  I also did a test weld, and cut it apart, There was actually an inside bead as well as of course an outside bead.   I used a Lincoln invertec welder (which is also a tig welder).  Fairly new welder for me, and it is sweet.  It stick welds better than any welder I have used.  It also took some care to keep the the tubes concentric during the welding.  

Changed up my junker bike a little.  It is an old 27" (wheel) schwinn that i use 700c wheels on.  I made a new fork for it last year to accept big tires, like cross tires.  I wanted a little more clearance on the rear brake, so i cut the brake bridge and brazed it on a little higher.  Took maybe a half a day, a tad longer than i wanted, but it came out well.  The brass and flux that Henry James sells sure is great stuff.  Amazed again at how nice it flows. 

Last week I also built up a set of wheels.   Just a normal set 32 spoked.   When running with 32, 28 and even maybe a front 24 and a mid or deep v, you really don't need to do anything special to the rim to nipple interface.    The rear rim is an offset design, which, if I am going to do a 1 to 1 ratio of Drive Side to Non-Drive side spokes and and 8/9/10 speed cassette, is all I will use anymore.  It allows a little better tension ratio and thus a better wheel.  The front (not shown) was a 28 hole deep v design.  I paid maybe $25 a rim, $5-10 each for the shimano hubs, and maybe 10 cents a spoke.   So maybe $75 dollars and a half hour a wheel to build.  I suspect these will still be running fine 10 years from now, unless the sidewall wear from brakes.   In the background is a wood box with drawers I made to hold spokes.  I have quite a few spokes that i have purchased dirt cheap off ebay over the years.

 Here is a Rolf Rim modified from an 18 spoked rim to a 27 spoked wheel.  I used 221"two-to-one" spoking pattern.  I actually built it last year, but had to rebuild cause i screwed up the fit of the nondrive side flange unto the hub and start squeeking!  I was able to move it further outboard this time around, and now all 27 spokes have the same tension (a function of flange distance from wheel center) .  

The nipples are actually inverted and inside the rims.  I used some small washers as well (2.2 mm ID) in between the nipple and rim.  Not quite ideal, because the nipples had to be chamfered, rounded somewhat and the washers curved taco style.  So two lines of contact in between the nipple and washer. And it took some time. The rim will probably definitely not ever fail, given the washer spreading out the load.  The nipples, are quite a bit stronger as well, being up side down and being on top a washer that has a 2 mm hole (instead of typical 4 mm).   And finally, the rim should be more aerodynamic, with hidden nipples.  Truing and tightening was almost easy as with a normal spoke wrench using a square driver.   

2014 comments.  Whoa too much monkey business...  Now carbon forks for me :)  (Well as long as you can get the rake/trail correct).     That poor bike was stolen.  I think I still have that wheel though ? 

Saturday, March 7, 2009

Spoked Wheels-Part 1 Intro

I've built spoked wheels since 1989. A lot of engineering mechanics and design concepts are nicely illustrated with a bicycle wheel. The concepts are not necessarily complicated but can still be hard to grasp. Even with engineering degrees (although I understood the physics behind the wheel) I did not really start to "know" spoked wheel until i sat down and starting doing calculations and thinking. For example things like 2-cross versus 3 cross, hi- versus low-flange, etc. Jobst Brandt's book and internet postings of course are great.

And here I am 20 years later, and i decided i want to start actually putting numbers, and showing the engineering behind the concepts. Although I am able to make generalized statements, it would be nice to be able to be specific, and dispel mis-information. Thus the reason for doing some calculations and sharing it in a blog. This will be a learning experience for me as well.

There are a lot of very expensive boutique wheels being sold, some are very good, some not. With some care and by incorporating a few techniques, a home or custom builder should be able to build a wheel as good as any (meaning as light, possibly more durable, and way less expensive).

This Part 1 blog entry will be the "generalized statements" and notes and current opinions (subject to change!). In subsequent blog will start to show calculations.

Wheel building an Art?
The wheel, when done, may look like, or be art, but the actual wheel-building process? No. It is aluminum and steel, tensioned. It's a bolted joint(s). If things are kept elastic (as they should be) we can do calculations and fully analyze a wheel.

Equal Spoke tension in wheels.
This of course is the holy grail (as it should be). Most wheel makers will even tout how even their tension is. (be wary if they state that the wheels may be a little out of true due to equal tension being their first priority. Wheel trueness should always be the first priority!)

But then a highly dished rear wheel gets built where the spoke tension on the NDS (non drive side) is about half of the DS(drive side). This is certainly not equal tension and the result can be a very poor wheel. Obviously not an issue with single speed wheel or really even 5,6, 7 speed wheels because the "dish" is not as great. [also Mtb (135mm spacing), and tandem(145mm+) wheels]. But with 130mm spacing and 8,9,10 speed road wheel, the issue is pushed just a little too far and problems arise.

Yes many wheels with this drastic un-equal tension have been built, and have functioned fine. But it is not ideal. Either the NDS spokes have too little tension and thread lock is needed (or constant "touchups"), or the DS spokes are so tight that either the nipple or rim fails (usually a "fatigue" failure a couple years down the road).

One solution is in the form of 2 spokes on drive side for every one spoke on non drive side (aka, G3, Triplet, etc, or 2-to-1) . This works well because the tension ratio (which is simply a ratio of the flange width distances) is 0.57 (for DA 7700 9s). There are other spoking patterns as well that I will present.

There are other techniques such as Offset rims, which is a good idea. Or a very-very poor idea such as moving the flanges closer together.

Paired Spoking.
It works. With such low spoke count, the rims have to be very strong (ie heavy and also means welded) and very straight. I have taken a number of Rolf Rims, which are super quality, drilled a few extra holes and built a nearly normal wheel with them. One of the reasons for the design of paired spoking was with nomally spaced low spoke count wheels , a "sine wave" could be put into the rim, from one spoke pulling one way and the next spoke pulling the other way.
So by pairing the spokes the problem is "solved". The paradox though is that paired spoke design requires a stronger stiffer rim. Which also solves the"sine wave" issue if used with normal spoke pattern/spacing. So you get there either way. However for an extremely low-spoke-count wheel, Paired is probably better. (future Blog Entry with some calcs..)

Low Spoke Count / Rim and Nipple Strength
I personally really like deep profile low spoke count wheels. (I consider low spoke count to be ~24 spokes or less). A lot Easier to build and more aerodynamic. The main issue is that a lot more spoke tension is required.

This then means extra care is needed at the nipple to rim interface to spread out that force. Rolf, and well most wheel builders, unfortunately do not do this (at least as far as I know) and nipples will break or the rims will fatigue and crack out sooner or later (might be 1 year, my brother's neuvation wheels.. or maybe 5 years depending on use ?)
What is the tradeoff between low spoke count wheels and rim profile?

Lacing PatternRadial lace is not a problem if your hub flange can handle it. If in doubt, or if a low spoke count wheel that requires high tension, don't do it.   Also the radial force can cause the but bearing fit to be lost.  

Obviously to transfer torque from the hub to the rim, the DS (drive side) flange needs some sort of 2, 3, or 4 cross lacing. with low spoke count wheel, I aim to have the spokes be nearly tangential to the flange. With higher spoke count ie 32 or 36 your don't need to be as concerned. Almost always have done a 3 cross. A 4-cross only to accomdate a given spoke length.

Note if have large diameter rear hub body, torque can be transferred to the NDS (non drive side) and you can do the 3-cross on that side and radial, or etc on DS. Some wheel wheels are seen with this. Any advantage? No. Keep in mind that most of the force in a spoke is preload from building the wheel. When torque is applied (to a wheel with lacing pattern, as just discussed) , the amount of additional force adding and subracting from the spokes is nearly insignificant, no matter how super strong the rider might be. (later blog entry to put numbers to this...)

An exception, obviously is a full radial lace to rear wheel, which is basically a "toggle-over-center" which can develop huge mechanical advantage (actually at point of toggle, it's infinite!) and possibly huge forces dependent on elasticity of members involved. (the concept is the same as Vise-Grips, Quick Release, or even why electric tension wires are purposely "sagged".)

2014 comments: I still think my comments are good.  I really like 2 to 1 spoking for a 8/9/10/11 rear wheel: equal spoke tension is just plain smart.  Unfortunately I have noticed a lot of hub manufacturers are decreasing the flange distance on their rear hub.  Bad Idea ! 

Tuesday, February 24, 2009

Quick Release Force

Have you ever wondered about  quick-release clamp force? (that which holds your wheel in the frame) ?   Way back, this came up in a forum, and it was mentioned that Bicycling Magazine actually tested/measured it! (imagine the time wasted!) At the time,  I commented that, that was pretty silly because it is easily calculated. Which I didn't bother to do until now and yes it is easy calculation!   

From BLOG Pictures

It is a statically indeterminate calculation, meaning it is not solvable with force equations alone, deflection/stiffness of components  is also considered.  Since Metal (steel ti or alum in most cases for bicycles), under elastic conditions, behaves linearly (simply meaning that if the force doubles, so does the deflection, and so on), this is not an issue.  The components can be thought of as Springs, where Force=SpringConstant*Deflection  or  (F=k*x).

In the case of a bicycle hub, the main components are the quick release rod elongating, and the hub's axles, compressing.  Also involved are the nuts, and frame dropouts. 

For items under axial loading the equation to calculate the spring rate is ElasticModulus*CrossSectArea / Length    or   (k=EA/L).

The amount of deflection introduced to the system is the "throw" of the quick release lever, which is typically reccomended to be 90 degrees (so we will use that).  With most quick releases this equates to 1 mm of deflection.   

In the table below (done with MS Excel) the measurements came from a typical shimano rear hub setup.  Everything is assumed to be steel (thus E=30e6psi).   Most of the measurement are not exact and for these purposes really don't need to be, with maybe exception of the quick release rod.  Assumptions also have to be made as to where in the threads does the force path change from nut to bolt. (good assumption is 3 threads in...)
ItemModulusLengthOuter DiaInner DiaOD of clamped volX-Sect AreaSpring Rateamount of deflection
ELdodid3AAk (EA/L)
Axle Nut30E638179.5156.10.2421910150.00091
Axle Nut30E638179.5156.10.2421910150.00091
QR Nut30E66154164.10.25412721510.00014
QR Nut30E65160201.10.31218698800.00009
QR Rod30E61565019.60.03058530.02958
Overall Spring Rate [lbf/in]4398
amount of throw [in]0.039
Amount of Clamping Force: [lbf]173
Amount of Force if consider only QR rod:  [lbf]230

Only 170 lb !  Not much.  Mostly dependent on the Quick Release rod, as expected since it is considerably more flexible.  Other notes:
-Compare to a clamp force of bolting: ~8000 lbs vs 170 lbs.  Thus the need for locknuts that "bite" into and deform dropouts.  Thus  vertical dropouts, Lawyer Lips, disc brakes causing pullout, etc. 
-If you going to use a Ti-rod, be wary.  With a modulus of about half of steel, you will only get 1/2 the clamp force (<100lbf!)>
-The calculation is a lot easier and can be almost as accurate if you consider only the qr rod.  It =   E*A/L*AmtThrow.  If the axle is titanium or aluminum with small x-sect area then you will want to include that as well.  
-Obviously if you are able to apply more Throw/deflection into the quick release, the amount of force goes up linearly.  Ie in this particular case, 2mm would give 340 lbf.  
-If "things" plastically deform during the clamping, then all bets are off on clamp force. 
-How about quick releases with plastic (actual plastic) in the force path ??   Sucks to be you, Good luck !!    No actually the plastic is so thin, ie the "L" is small thus the spring rate is still high. But yeah, personally I do not use them.  
-Note the approx 7 mil of axle compression (for a typical 10mmx1 steel axle..).  Thus the reason to have a little play in the bearings under no-load. 
-Note the nice mix of American and Metric Units.  Ametrican!  Like a lot of other American Engineers, we tend to use what we are use to.  Another quirk,  7 mil = 7 thou (thousandths) = 0.007 inch (mil is not millimeters, but rather is short for a milli-inch, crazy huh?)