Tuesday, July 29, 2008


After lots of talk about carbon building technique, I thought it would make some sense to show some pix. So I'll be shooting some 'intermediate' stages of work to use in show and tell - stay tuned.

This season, I've been riding a prototype bike - designed to test some long reach brakes. The fork is straight-legged steel, but with fairly light weight legs - which lends it a nice ride. It's raked at about 53mm to shorten up the trail, because the bike is riding on 700Cx28mm tires.

Generally, I like the ride and handling quite well, but there is one funky handling quirk When rolling into turns at over 14-16 mph,it's sometimes necessary to turn my inside knee towards the apex in order to tighten my line. It appears that there's just a little too much tendency to hold the line, and that the bike may benefit from even less trail, but I'll have to try making another fork to test this out.

My tires have been a set of Hutchinson Top Speed. Lot's of folks look at them and don't believe that they are a real 28mm width, but true that. The key thing is that they're giving me a really cushy ride and they roll like anything. Part of the credit for rolling goes to a set of Record hubs, but note that these were purchased (as wheels) used, and haven't been serviced at all since I got them. Consequently, I think that the tires deserve their share of credit for low rolling resistance.

Anyhow, I didn't think too much about the tires when mounting them. They were in the equipment stash, and of the proper size - hence no cash outlay - good enough for my purposes. Recently, I was testing out a fixee with some nice Gran Bois 700Cx30mm tires. At about 10PSI lower pressure, the ride wasn't as good. Part of the difference is the fork. The fixee is built with an antique set of Reynolds 531, including pre-curved blades with the old English style profile. For those not familiar with this profile, it is longer and narrower at the top when it connects to the fork crown than the more common continental oval. Anyhow, these blades are definitely stiffer than what's on the prototype bike. That said, the frame is stiffer on the prototype.

The Gran Bois are great tires, and I can probably reduce their tire pressure some more. Also, they haven't had enough miles to break in. But the difference in rides between these two bikes is quite remarkable, despite my anticipation that the Gran Bois would ensure that the Fixee was more comfortable - leaving me with a riddle.

After searching online for more info on these Top Speed tires, it became clear that Hutchinson has stopped offering them in the 700Cx28mm size. What a Pity. Anyhow, digging in the equipment stash turned up another pair, same size (although different color) still in their packaging. The printed spec is for a carcass with 66TPI. This surprised me too. I don't usually think of 66TPI offering a very compliant carcass - but there you are, the ride of these is great.

Now I'm really mystified. It looks like it's time to try swapping front wheels on these bikes and riding them back to back. It's hard to believe that the tires are creating the difference in ride, and the wheel switch should help establish if this is true.

Meanwhile, its nice to have stumbled onto these tires and have a spare set, but its sad that they aren't made any more.

Changing topics, let's consider the evolution of handlebar shapes. Along with the evolution to 'anatomical' handlebars, we've seen a push away from long and deep (sometimes called Belgium drop) bars. I've never found an anatomical bar that seemed more comfortable to me than a traditional bar, especially when down in the hooks (which is where the 'anatomical' part of the design is typically located). Hence, they've never done much for me. But the loss of availability of a range in sizes of drop and reach has felt like a loss. So maybe I'm a luddite.

The very latest anatomical bar designs, however, seem to be onto something which may be indicative of modern riding styles. Several mfg's are now offering bars comprised of smooth curves, which quickly bend backwards under the brake levers. Like most anatomical bars, they have a very short ramp leading up to the brake lever.

Essentially, this leaves the rider with several riding positions: top of bars, on the brake hoods, in the hooks, and at the rear of the drops. Note that each of these positions is equally close, or closer, than their respective position on traditional bar (because the hooks pull backward more, and the drop tends to be less extreme than on a traditional bar). Meanwhile, they don't really offer a position on the ramp behind the hoods, because the ramp is so small.

So what's going on here? Are we trying to sit more upright while riding today, than did riders in the past? Absolutely not. But, many riders today seem to stress long and low stems. Often this situation is exacerbated by the use of a threadless headset without the compensation of a longer head tube. Hence, to keep the brakes in reach (not to mention the hooks), it helps to have a shorter handlebar. And to keep the drops in reach, it helps if they don't drop as far as traditional bars, plus having them extend backwards more.

"So what?" you may say, the stem goes one way, the bars in another, and we end up in the same place. But that's just it, we don't really end up in the same place. We lose the ramp as a hand position, and in my experience its a great position - with a low likelihood of aggravating the ulnar nerve. In my book, that's worth thousands of dollars by itself. Also, the difference in body position is reduced when moving the hands from the tops to the drops. Old bars had the tops closer to the seat, and the drops lower and farther away from the seat, as compared to the latest examples of anatomic bars. So, one's body position is less likely to change as much on an anatomical bar when moving between these grips - and that's the first reason to change grips.

Does anyone besides me care about these changes? I don't know, but it's food for thought when trying to tailor your position on the bike.

Finally, in another ludditish (word?) rage, let me take on cassettes (cogs not music). It took me a long time to understand the fascination with cassettes using 11 or 12 tooth small cogs. Let's face it those are for speeds in excess of 40mph, which very few people will achieve except going down hill. You may point out that we all spend our share of time going downhill, some would say I go downhill at an ever accelerating rate. But, pedaling down hill is largely a waste of energy. The exponential growth of wind resistance means that pedaling will add very little to your speed and is unlikely to decrease your elapsed time.

Yes the pros pedal down hill in the Giro, TdeF, and Vuelta - and you should plan to do so also, when you're riding for a pro team in the Grand Tour. But, that's probably not likely to happen (if only because you're reading this instead of training), therefore you shouldn't worry about pedaling down hill or about having an 11, 12 or even 13 tooth cog.

This leaves the question of why the big 3 push these over-geared cassettes on us? I think its because we're all weight-weenies at heart. Lower the size of each cog by 2-4 teeth and you'll lower the weight of your cassette. Have you seen what folks will pay for a Ti cassette just to save a little weight? And the mfg can save weight just by using smaller cogs, which probably also reduce the cost to mfg (less material). Hence most cassettes start with an 11 or 12 tooth cog. Ugh. I wish SRAM would come up with a 13 or 14 by 27 tooth Red cassette. That would be cool. But that wish isn't likely to be granted soon.

Ok, carbon pix soon, and maybe some more surprises.


Thursday, July 10, 2008


Here's just a little update. I'm getting ready to sag a bunch of my riders on RAIN (Ride Across INdiana) on Saturday. We'll leave town about noon tomorrow with two sag vehicles and 8 (I think) riders. Tonight I've been pulling together my pit gear. I won't be able to replace shifters, bottom brackets, headsets, or anything but 10 speed Shimano compatible cassettes, but should be set to tackle anything else. Either the variety of standards, or size of tools required, were my two criteria for what not to bring. It's not that I expect problems, but being prepared is the best defense against having to fix anything.

Moving on, my new Ti parts arrived this week. So sexy. Between the BB and head tube I expect to remove about 200 grams from my typical carbon frame with these parts. They're also beautifully made.

My current build is coming along nicely. If you're a long time reader, you know that I'm a keel builder. That is, focus on the head tube, down tube, & chainstays, to ensure a straight keel between the wheels. Then fill in the rest. The reality is, however, that I usually connect the seat tube to the BB first - and then let it just wave in the wind until the keel is done.

With a bagged carbon frame, there are a number of ways to do things, and I chose a sequence slightly different from how I build steel. I begin by mitering the chain stays (which have a mono-yoke) to the BB, then glue them together in a jig with aerospace epoxy adhesive. Once this sets up, 8 layers (more will be added latter in the process) of uni-carbon are wrapped around this joint - five run straight and form a 'U' when viewed from the side, the remaining layers are angled about plus/minus 25 degrees. Each angled layer includes both plus and minus angles - as I'm using narrower strips of carbon - so the layer has a crossing of the two angles, but nets about the same amount of fabric as one of the straight layers. As a final step, three layers are wrapped around the yoke of the chain stays (90 degrees to the main reinforcement). After fiddling to make sure that all is flat and smooth, the yoke portion gets wrapped with heat shrink tape. This serves to flatten this area nicely, indicate if there are problems in the wrap around the BB shell, and helps hold the rest of the layers in place until the vacuum is applied. Note, there is a concave space where the top and bottom of the chain stay butt up to the BB shell. If the main wraps are too tight, they will lift out of this area - making a bubble and potential stress riser. So with gloved hand, I check to make sure that the tension on the wraps is correct.

This then gets put into a vacuum bag, and as the air is evacuated, I work the bag to lay as flat as possible all the way around the BB/chain stay joint. The uni-carbon comes with something (it varies) on the back to hold the threads together and in parallel. With a good vacuum job, its possible to see this backing through the carbon when the joint is later unwrapped.

The finished assembly is inspected, and then sanded with 180 grit to prepare to bond tubes and more CF. Note, I could use a peel-ply that leaves a thicker layer on epoxy on the surface, with the pattern of the fabric embossed in the epoxy. This makes a good surface for bonding other bits too. However, I find this fabric a bit stiff and unwieldy for working around this sort of joint. Hence the sanding of the finished surface.

Next (reverting to my old habits), I miter and bond the seat tube into place. With steel, I'd just use a pattern from Bike Cad to mark the miter. This works here, as well, but only up to a point. At the rear, the yoke of the chain stay interferes with the fit of the seat tube, and the seat tube therefore needs to be trimmed down carefully. Once the proper fit is established, all surfaces are cleaned up with rubbing alcohol and allowed to dry. The fit over everything is checked one more time in the jig. Then a layer of aerospace epoxy adhesive is applied around the base of the seat tube, and the tube is put into place, and the jig is closed down on the seat tube to hold it in place. Supposedly the parts can be worked within two hours of bonding, but I generally give them overnight.

Meanwhile, I've been preparing the head tube. This one gets my traditional aluminum head tube with a CF wrap. Actually, a fine layer of fiberglass goes down first, then the CF. In this case, I used some 5.7 oz plain weave CF. Over the CF goes a plastic peel layer in which I've punched a lot of small holes (pin pricks actually). Over this goes a layer of synthetic cotton batting - which serves to suck up any epoxy squeezed out of the CF. Finally, I give it a tight wrap of heat-shrink tape. Once its all stabilized (tape on the ends or whatever), I use an electric heat gun (like some folks use to remove paint) to quickly shrink the tape. With a head tube set up for cure, I usually put it in the oven at about 175 degrees for 30 minutes. This speeds up the cure, but also helps to bleed off excess epoxy.

This process seems to work because the whole piece is evenly coated in epoxy (which shows as a glossy sheen), but very little of the texture of the CF cloth is lost. On the final layer (once joints are done), we'll want a thicker top coat of epoxy to provide a smooth base layer for the painter - but until then, we want to use the least epoxy possible to do the job - and this method seems to work very well at meeting the goal.

Now its time to do some measurements to mount the head tube in the jig, and position it correctly relative to the BB. Key issues are, of course, the head tube angle, effective top tube length, height of the bottom of the head tube. The later are designed around the fork and headset which will be employed, to ensure that the prescribed head tube angle is realized in practice. After a bit of fiddling, a satisfactory positioning is achieved.

Then, its time to begin mitering and fitting the down tube. On a lugged steel bike, its possible to make the down tube a half inch long, fit things up in the jig, then mark the excess from within the BB. That doesn't work here with a solid BB. So, I start by fitting the down tube to the head tube first - using a protractor to check that my angle is correct. Then the BB end is mitered, but it's left intentionally long. Now I fiddle to see that my down tube/seat tube angle appears correct. If not, there's something wrong with the head tube positioning. This is just a double check, but nonetheless an important step.

If positioning looks good, I carefully start to carve the BB miter back until I can fit the tube into place. If all has gone well, both ends have nice tight fits and I don't have to recycle an expensive piece of CF.

At this point, I once again check all fits on the jig, then clean the down tube, head tube, BB, and seat tube with alcohol. Again, adhesive gets applied, this time to both ends of the down tube. At the bottom it is fitted to the BB and to the seat tube. Once more, the jig and fits are double checked - before anything can set up.

OK, that's how far the current frame is. We'll next have to modify the BB joint to ease the process of draping layers of CF thereon. So stay tuned to learn more about finishing this important joint.


Post ride update... The only mechanical I had to deal with was.... a bottom bracket. At the first stop, one rider had a crunchy dragging BB. Fortunately it was a cup and cone style, as I hadn't brought any spares. The bike was borrowed, and both the axle and one cup had some damage in the races. Also the bearings were caged, which are easier to keep track of, but which I find to offer less good results than loose bearings.

Anyway, I was able to repack the BB and adjust it so that it would spin smoothly. Then I drove off to purchase a spare (just a basic 113mm Shimano cartridge square taper BB). Anyhow, the rider made it through the ride without further issues and I never had to install the spare.

Congratulations to all the riders who completed RAIN!

Wednesday, July 02, 2008

What's new

Let's see, he said as he stroked his chin. "Hmmmm.....

Now that the weather is warm, I'm working hard to catch up on my backlog. Carbon has been at the fore of my efforts as noted in the last post. Along these lines, I'm having some fun and showing great progress in a number of dimensions.

I've begun working with Edge Composites for tubing and rear triangles, and they're great. They'll build to my spec, and can turn around custom requests in pretty short order. A very simple example, I can now spec a tube a either plain uni-directional fiber, or any of several weaves of overweaves (from the traditional burlap look to the newer 12k checkerboard looks). So riders get a choice of aesthetics and performance. On the front triangle, plain uni carbon saves about 10% in weight, on the rear triangle it's closer to 20%. Not bad. But, for those who don't worry about fractions of an ounce, there is the option to choose their favorite fiber look. We can even turn the weaves on an angle for one more dimension of customization.

Speaking of rear-triangles, Edge molds the cable casing stop into the chain stay - so that eliminates holes and rivets - which is a good thing.

They've also provided their input into my plans for CF dropouts. This option is especially appealing to me for track frames where no one is manufacturing a carbon compatible dropout. That may be changing as there is another player getting ready to announce some super neat metal dropouts for CF rear triangles. I can't say any more for now, but I'm getting excited about that.

I've also been exploring various options for BBs and head tubes. Eventually, I think these will be pure CF, with the option for BB30 bottom brackets. In the meantime, I will shortly have some nice Ti BBs. These are thinner gauge metal than what you would see on a welded Ti bike, and will be reinforced by the CF over-wrap, rather than having that merely be a surface to which other tubes are bonded onto the BB. I think this will be lighter than my current wrapped Aluminum BBs, and have longer lasting threads. Plus, Ti resists electro chemical interactions with CF better than just about any other available metal. In a related move, I'm moving to CF head tubes with Ti rings bonded into each end. The Ti gets reamed and faced for the headset, but the CF tube provides the structure for the front end. Much less weight than a wrapped aluminum head tube, and again the chemical stability of Ti. So, all of these parts represent steps forward - and will help distinguish my CF frames from the run of the mill

On a different front, I've been constantly refining my vacuum bagging technique. It's easy to shove some parts in a bag, turn on the vacuum, and wait for them to cure. What gets tricky is maintaining the layers of wet CF in alignment and snuggly wrapped around the tube. A loose piece can easily create a bubble between layers that will become a source of failure. Even tricker is doing this in a manner where we get a nice smooth finish on the outside of the part. Many forms of vacuum bagging work against a mold - which is highly polished. The CF surface that lays against the mold (often with a layer of gel-coat between) is essentially finished when done.

I could create molds for my joints - some manufactures do. But, that limits the combinations of angles and tubing sizes I can use (either that or have a nearly infinite range of molds available).

Instead, I'm wrapping the joint with a layer of smooth plastic release layer. The plastic has small holes that allow excess epoxy to weep out into the bleeder felt. This layer of plastic has to be fashioned to follow the contours of the joint - so it doesn't wrinkle and cast the wrinkle into the epoxy. The bleeder felt also has to be arranged so as not to wrinkle. I've taken to fitting multiple pieces of felt to the joint - because it won't stretch to fit. Finally, it's important to arrange the 'bag' around the joint in a fashion similar to the release layer - again to avoid wrinkle. The net of practice is that my joints are coming out of the bag much better finished - and needing much less touch up before they're ready for paint. Cool.

Anyhow, thats enough words for tonight. See you next time.