In our prior installments on the subject of John Castellano, we learned a lot about his upbringing and the events leading up to the wonderful collaboration Ibis enjoyed with John.
Today, we’ll get into a bit more detail about two of his most famous creations: the BowTi and the SilkTi.
Before we even get going, our first detour. In reading through the last installment, we realized an unpardonable omission; we forgot to talk about the shoe tree. The prior story was about a trip to Moab, along the “Loneliest Highway in America”, also known as Highway 50 across Nevada. Out there, in the middle of nowhere, sat a large cottonwood tree along the side of the road, covered with shoes.
The shoe tree was always a nice site along the road. It didn’t mean you were getting close to anything, but it indicated that you were making progress. We always stopped to stretch our legs and make sure we didn’t have any errant pairs of shoes lying around in the back of our vehicle.
The shoe tree was one of those memories of road trips that draws you back time and time again to roadside America. That and the fact that you can drive really fast out there.
On December 30th of 2010, vandals cut the tree down. It made national news. Maybe not front-page above-the-fold national news, but we did hear about this quaint icon going down out here in egocentric California. We'll miss the shoe tree in future trips to Moab. Using the immortal words of Walter Sobchak we say to the tree: “Goodnight, sweet prince”.
During this trip to Moab, John convinced us to take a huge gamble and build the BowTi, a pivotless titanium suspension bike with 5” of rear wheel travel. The bike was one of the most expensive at the time, $3500 or so for the frame in the mid 90’s. It was the most technically difficult bike we’ve ever made, by a long way.
Here are a couple of questions (reused from JC part 1), where John talks a bit more about the BowTi.
Q: Where did the idea for the Bow-Ti come from?
A: I had always been intrigued by the idea of a pivotless suspension system, so that was percolating in my head. Then after I built the prototype Sweet Spot bike [the Ibis Szazbo], I was trying to envision a way to utilize that suspension concept with a flex pivot. When I worked at Hughes Aerospace, we used flex pivots often because they don’t wear out, they don’t require lubrication, they’re simple and super lightweight. I mean when you’re 23,000 miles from the car, you don’t want something to break. So I was drawing sketches and I came up with what we now know as the Bow-Ti. I made a little brass model of it and it seemed to work - stiff laterally and soft vertically. So I patented it.
Q: Tell us about the development of the Bow-Ti.
A: Well, really the first prototype Bow-Ti was a computer model that I “rode” on screen for about three months, playing with tube diameters and wall thicknesses. Then we built one in actual titanium that we could ride and test and measure stiffnesses on. Then some more time on the computer doing Finite Element Analysis, or FEA. This gives pictures showing predicted stresses on the tubing and the gussets, so I could refine the shapes. Then we started cranking them out!
Here are a few pictures of the BowTi, click the first two images for a larger view.
John was a featured superstar (JC Superstar) in our 1998 catalog. Click on the image for a purty and readable pdf.
We also made some Titanium Bow Ties, not to be confused with our Titanium BowTi. They looked like something the Joker would wear on Batman, they could probably be used for a weapon too. There were two sheets of formed titanium, bolted together with an Ibis headbadge out front, and a velcro closure (none of us ever learned how to tie a tie).
And here's Scot modeling the Tie with none other than Phil Ligget.
After the success of the BowTi, we decided that we wanted a bike that was more accessible to the masses from a price standpoint. We'd seen the success and simplicity of the Moots YBB bikes, and figured the softail was the direction we wanted to explore. So John got to work on a the SilkTi.
Like the BowTi, the SilkTi was also pivotless and made of Titanium. It utilized a flat plat chainstay, that was waterjet cut out of a solid plate of Titanium. John had purchased "Pro E" a professional engineering (hence the name Pro E) software program that allowed him to build bikes on the computer and test the stresses using Finite Element Analysis (FEA). He learned a lot doing the BowTi, and took it a few steps further with the SilkTi. There were several other unique features on the bike. We conducted an interview with John back then, and it's appropriate to repeat it for you right here:
Is the softtail a new idea? The softtail’s actually been around for 100 years. The first one I know of was patented by Charlie Travis back in 1899. He was way ahead of his time. His bike was pivotless, with a flat plate flexure. It had an air shock made like a tire pump--very advanced! I think that if Charlie Travis had elastomers, titanium, and a waterjet machine, he might have built this new Ibis.
What are the advantages of flat stays? You’d like the chainstays to be stiff laterally for good handling, yet flexible vertically for better suspension. Round tubes can’t do both of these, precisely because they’re round. They flex as much side-to-side as up and down. The chainstays on this new bike are designed as a planar truss, triangulated for lateral stiffness, but flat for vertical flexibility. There were a couple of bicycle mechanics named Orville and Wilbur who also used a planar truss in the wings on the Wright Flyer. The wings were designed to flex in certain directions in order to steer the plane. Now, with computer controlled waterjet cutting, the planar truss makes its way back to bicycles. These new chainstays are stiff as a bridge laterally, yet provide twice as much travel as any other softtail can. Stresses are reduced in the chainstays and the welds, too.
How do you use your FEA in product development? Finite element analysis lets you try out and compare new shapes, for example, comparing tube chainstays to truss chainstays, or evaluating a new shape of treadblock. Then if you come up with a good basic shape, FEA lets you optimize and refine the shape. FEA stress plots tell the designer where the stresses are in a proposed part. You want to beef up areas that are highly stressed and shave down areas that aren’t taking any load. This puts the material where it’s needed. On these new chainstays, you’ll notice all the material is on the sides of the stays for maximum lateral stiffness. The perimeter has been shaped so that the webs are aligned with the load paths predicted by FEA. There’s still some trial and error involved, but the errors are on the computer model.
What design challenges are unique in building a softtail? One of them is to make the bike stiff laterally. It’s challenging to design chainstays that are soft vertically, so they can flex into an S-shape and replace the usual pivots at the BB and the dropouts. Another is to make a simple, high-performance shock. That’s the new CDE shock.
What is a CDE shock? CDE stands for critically damped elastomer. It’s simple but very highly engineered--and it’s ideal for softtails.
How long has the CDE shock been in development? About five years. It has evolved into a design that performs well, and is lightweight and reliable. It is user friendly, providing more travel with minimal maintenance.
Why is an elastomer ideal for a softtail vs. a coil spring? The problem with a coil spring, aside from excess weight, is that the stiffness does not increase as it is compressed--it’s not progressive. A suspension designer using a coil spring on a short-travel bike has to make some serious compromises. A coil spring must be stiff to avoid harsh bottoming, but then it’s too stiff to absorb even small bumps well. So it doesn’t work well on big bumps because of limited travel, or on small bumps because it’s too stiff. Sometimes a small elastomer is used inside the coil to keep it from bottoming harshly. This helps, but it’s still not as smooth as a pure elastomer spring.
An elastomer spring is ideal for a softtail primarily because the stiffness is progressive, which means it is soft in the first part of its travel, then becomes progressively stiffer. The soft initial travel makes it much better at absorbing small, high-frequency bumps. On big bumps, the stiffness ramps up for bottoming resistance. Plus, it’s lighter than a steel coil and has almost double the travel.
How is this elastomer different from any other elastomers? The elastomer used in the CDE system is a precision machined, highly engineered part. In addition to serving as the spring, the one moving part also functions as bump stop, negative spring, linear bearing, and most important, the damper. The purpose of the damper is to absorb bump energy to keep you on the bike and the tire on the ground. Many other softtails leave out the damper to save weight and cost. But a good damper is critical to performance.
How can the elastomer act as a damper? The elastomer is precision ground on the outer diameter, then lubed. It slides inside a polished sleeve when it is compressed. This gives a small bit of damping for small bumps and a lot of damping on large bumps--like an oil shock, but without all the seals, shims and pistons. We can completely control the compression and rebound curves without oil. All this with only one moving part!
You mentioned a progressive spring? Yes, the spring curve is carefully designed. It’s actually a two-stage spring: the first stage is soft and lightly damped for good high frequency isolation, unlike oil-filled shocks which transmit high frequencies to the rider. This first stage is very active when you’re standing and gives a lively feel to the bike. The second stage spring is heavily damped and only comes into play on bigger bumps. This gives high quality, progressive travel, which is essential for good performance with a short travel bike. Of course, the SilkTi also has longer travel than any other softtail. It’s the softest tail. And very fast.
We also uncovered some more technical info on the SilkTi, written by John, this time with some nifty annotation in the form of pretty stress plots, one is even animated. Take it away John.
For a given amount of deflection (travel) in a beam (the chainstays), the stress is proportional to the thickness of the part. We've reduced the thickness from the typical 3/4" (round stay) to 1/4", thereby cutting the stress to 1/3 of a round stay, for a given deflection. Roughly doubling the travel to 1.75" brings the net stresses to 2/3 of what a round stay sees--less stress with more travel.
Secondly, we're using 6-4 plate which has over 20% better properties than 3-2.5 tubing. It's waterjet cut, which leaves no residual stress like bending a tube does. This adds to fatigue resistance. The thinner stays are much more limber, vertically, which lessens binding loads on the shock for less stiction.
Here's a worm's eye view of the SilkTi flat plate chainstay.
Note that the windows are carefully shaped to distribute the flex uniformly around the two pivot areas (near the tire and near the dropout). The stress falls off at the weld because the cross sectional area increases there. Also the weld configuration is very favorable, with much less of a stress riser, compared to the sharp corner inside a welded tube. We've got a nice fillet all around.
Here's an animated illustration of the chainstay moving through its travel. The red indicates high stress, the blue no stress.
In lateral mode, we get about a 50% improvement in lateral stiffness, and again, much less weld stress. The left chainstay widens out in front of the tire and spreads to nearly the full width of the BB. The right stay does the same. This cuts weld stress bigtime. The material is almost all on the sides of the stays, and the webs are carefully angled to follow the natural load paths.
For comparison, check out the round tube model:
This shows the high stresses (red) in both lateral and vertical modes, especially at the weld. This is at .8" deflection, vs 1.2" deflection in the plate plot. You want the stays to be stiff laterally, yet soft vertically--a round tube cannot do this.
If we pushed the plate stays as hard as "others" are pushing their tubing, we could get about 3" travel. The shock limits the travel to 1.75", which should make for infinite fatigue life
The SilkTi was a smashing success (we still see them being ridden all over the place today). John wasn't done fiddling around with materials or pivotless bikes yet. And we wanted to take the pivotless concept to the masses, by lowering the price even more.
Enter the Ripley, a pivotless Aluminum softail. That's right, Aluminum believe it or not, the material you've all be told will eventually fail if put under even moderate amounts of stress. Which is true. A couple of things need to be remembered though, engineers design aluminum structures so that they don't fail during their useful life. They can predict this type of thing. Take for example airplane wings, which are aluminum, and which flex a lot, 26' at the tip of a 747 wing under extreme loads, several feet under normal loads.
John did the same thing for the Ripley. It's designed to go for years and years, ridden every day and have no failures at the chainstay joints. Or anywhere else for that matter.
Here's a picture of a Ripley, at the grave of its namesake.
What's Johnny up to these days? Lots and lots it turns out. He's keeping people supplied in elastomers and shocks and derailleur hangers for the various bikes he designed over the years. He's also retrofitting disc brakes on many of the old bikes. He's hopefully getting a new website launched soon, the current one is very out of date: http://castellanodesigns.com/.
Here's what John says he's doing:
I'm mostly making custom SilkTi's with Steve Potts, mostly 29ers. Sometimes with Paragon Sliders, or S&S Couplers.
The Zorro, son of Szazbo is great, but a lot of work to produce. I've made a folding one, a couple with Hammerschmidt mounts, a recent one with belt drive, pictured here: