THE DATA AND SOME CAVEATS

First, there is little to no standardization of the tube lengths. In general, I think chainstay is taken as center of the bottom bracket (or more often the spindle) to the center of axle, but not always. "Size", which is a marketing term is conflated with "seat tube length", which is a design term for which there at least 2 different meanings; center to center and center to top. In general, size (or seat tube length) is presented in geometry charts as being rounded off to the nearest inch or centimeter and so I've used this approach. Top tube length and effective TTL are also both recorded as being the same and generally are assumed to be center to center but who knows. So remember, there is no universally accepted way to measure tube lengths and the resulting noise is present in the data we've collected.

Second, most of the newer frames in the list are from catalog specifications while many of the older frames are, by necessity, from hand measurements. Even if you assume that all specs were correct (they aren't), that bikes were accurately made according to published specs (they aren't) and that all submitters have measured the frames the same ways (they haven't), we need to understand that these are fundamentally different kinds of data.

So, if the data in the spreadsheet needs to be understood as being approximations in all aspects, then his begs the question, does it have any value? I think it does for 2 reasons. First, it's a starting point and better than nothing (I hope). Secondly and more importantly, I think approximations are enough if we remember to think of bike designs as categories with hazy, ambiguous borders instead of strict classes with precise, non-ambiguous borders. Whether a chainstay is 41.3cm or 41.5 cm is nowhere as important a distinction as the recognition that it is a comparatively short chainstay length. Whether a fork rake is 6.00cm or 6.30cm is not as important is recognizing it as a large amount of rake.

A couple of other notes about the data are in order. Wheelbase, top tube length and bottom bracket drop are all interesting aspects of frame design. Unfortunately, the data is not universally available. By all means, if you can help to fill in the gaps, don't hesitate to contact us with the data. We don't list either the front-center or the bottom bracket height. These just aren't listed often enough in the sources we've had access to.

We also don't have much information for frame sizes outside of the 56cm - 58cm range. Believe me, this isn't out of lack of interest as I ride a 64cm frame myself. In fact, the spreadsheet grew out of a personal project to research frames in my own size. What information we have on other sizes is captured in separate tabs within the spreadsheet. If you have a passion for tracking down the data for frame sizes outside of the 56cm - 58cm range, by all means please submit the data you find. We've focused on th 56cm - 58cm range because it is the most commonly published frame data.

THE FORMAT OF THE SPREADSHEET

• Type A (orange): Road Racer: approx 41.5 cm
• Type B (yellow): Audax/Event: between 42 cm and 42.5 cm
• Type C (green): Sport Touring: between 43 cm and 44.5 cm
• Type D (blue): Loaded Touring: 45 cm or more

Next, I further refine these 4 basic categories by sorting them according to their front end geometries, since it's generally accepted that this has the largest affect on the handling style of the bike.

The specific sort order that I current use is:

1. Chainstay Category (ascending)
2. Head Angle (ascending)
3. Fork Offset (ascending)

The spreadsheet also provides the the calculated values for the trail and the "flop-factor". The former is a measurement of how far behind the point of contact of the wheel is relative to the axis of the steerer. The latter is a measure of how fast the top of the headset drops as the bars are turned. Both are calculated based on the assumption that the frame uses 700x23c tires. While the vast majority of the frames listed use 700c wheels, a great deal of them use 27" wheels and a few use either 650B or 26" wheels. The spreadsheet does not record this in anyway, so the trail calculations will certainly be inaccurate for many different models.

Lastly, I've introduced a new column that I call the "Steering Category". This is based on some analysis I did a while ago based on the data in the 56cm tab of the spreadsheet. You will see a tab titled "steer cat" in the spreadsheet. This is a simple matrix of head angles and rakes. The cells in the matrix are color coded indicating the number of bikes in the 56cm tab that have that HA/Rake combination (as of the time of the analysis). This gives a crude visualation of the frequency of the different HA/Rake combinations, which in turn may be a relative popularity of the different combinations. Based on this, I've defined four front end geometry categories based on what I think are four prototypical geometries. They are:

• Race (orange): HA: 73.0; Rake: 4.5 cm
• Cross (yellow): HA: 72.0; Rake: 5.0 cm
• Sport (green): HA: 73.0; Rake: 5.5 cm
• Tour (blue): HA: 72.0; Rake: 5.75 cm

HOW YOU CAN HELP

The most helpful thing to move this project forward would be to send in more geometries. I'll gladly add them to the spreadsheet and republish it. I can be reached at pinnah at comcast dot net.

Here's what we need...

In terms of sizes, I'm primarily interested in frame sizes in the range of 56cm to 58cm (approximately 21"). This size range seems to the most widely published in terms of geometries and thus most obvious in terms of comparisons.

The geometry information that we need to add a bike to the list can be divided into 2 groups of "must-haves" and "nice to haves". Must-have information includes:

• Head Angle
• Fork Offset (Rake)
• Chainstay Length
• Brand Name
• Approximate Year

• Seat Tube Angle
• Horizontal Top Tube Length (Effective Top Tube Length)
• Bottom Bracket Drop
• Model Name
• Exact Year

In terms of brands and countries of origins, I'm interested in getting more Japanese and French bikes documented. In particular, I would like to get more of the Japanese bike boom sports touring and touring from the 1970s and 1980s. I would also like to find geometries for the even earlier French touring bikes from the 1960s and early 1970. I have a good deal of Italian and English bikes already documented but will gladly take more.

The brands of current interest include:

• Japan: Fuji, Miyata, Univega, Nishiki, Panasonic and Centurian
• France: Peugeot, Motobecane, Herse, Singer and Gitane
• Britain: Raliegh, Carlton, Hetchins and Bates

In terms of the kinds of road bikes I'm interested in documenting, first and foremost, I'm interested in documenting Sports Touring bikes, by which I mean light weight bikes with chainstays in the 43 cm to 45 cm range and Head Angles at or near 73 degrees. I'm also interested in expanding the number of entries for Cyclo-Cross and Loaded Touring bikes since these designs are similar to Sports Touring designs but more optimized for specialized purposes. Lastly, I'm would also be interested in receiving data on bikes that have been sold as Hybrid bikes. I'm pretty ignorant about Hybrids and would be interested to see how they compare to these other designs.

Again, I can be reached at pinnah at comcast dot net.

SUGGESTIONS ON HOW TO MEASURE A FRAME

The most important tools are a good eyeball and patience. With these you can make consistent measurements with low-cost tools, even it you don't have the accuracy available like at a good frame shop, Fattic's for instance. If you are going to measure seat tube length, it's not obvious how to see the top of the top tube or the center of the top tube. I like to use the reflective highlight from a flashlight to see the contour of tubes, then mark the center of the seat-tube/top-tube joint with a non-permanent marker, like a fountain pen (one of my other hobbies). Same issue at the BB. If your crank has an exposed bolt or endcap center, you can see where to place the tape measure.

I usearoofer's angle finder (Sears Craftsman or Ace Hardware) to measure frame tube angles. It isn't a very accurate instrument, so I tend to use it only as a difference measuring tool. So to get the seat tube angle, I measure the top tube, then the seat tube, then subtract. I try to get the dial or needle to indicate in the same portion of the angle finder's scale for each set of measurments. I think this technique is consistent to perhaps half a degree. You have to wait before taking the reading, to give the needle time to stop oscillating. You also have to make sure the reference surface, that is pressed up against the frame tube, is clean and free of debris, which can introduce an angle offset. If I see one at a flea market, I'll buy a vernier angle measuring caliper.

Measuring the head tube presents a special problem on my little (52-53) frames, that the angle finder reference surface is usually longer than the exposed portion of head tube, with the headset cups pressed in and the fork installed. If you are building up from a bare frame, measure the head tube while it is bare, before you install the headset. If I am assessing a built-up bike, I use a short wood block that has been miked to bear against the front of the head tube to space out the reference surface so the angle finder can lie against it properly and measure the correct angle. The block needs to be thick enough that the angle finder clears the headset parts. Here patience and stillness are critical, because bike work stands vibrate and the angle finder needle vibrates.

I measure BB height by the vertical clearance from BB center to the concrete basement floor. BB drop is the difference between wheel radius and BB height, so I measure wheel axle distance to the floor. A plumb line is useful to see what is really vertical. Chainstay and wheelbase can be confusing because of the variation possible within a horizontal dropout. I try to see where the chainstay center and seatstay center intersect, and measure from there -- eyeballs again!

Front-end geometry is the hard part. Here I like to put the bike on the ground with the front wheel on a piece of white paper, sitting as level as possible and blocked so it does not roll - little pieces of wood again. The front wheel contact is vertically under the QR skewer, so you can mark the contact point on the paper. I use a yardstick (aluminum for stability) to project the head tube/quill center line down to the floor, then I need an assistant (fellow cyclist/geometry dweeb is best, wife is second best, cat is worst) to mark the contact point. Then you can measure trail directly. Armed with math skills, the head tube angle, the trail, and a calculator, you can compute fork offset. You can also measure it directly from the yardstick with a second ruler, if everything is held steady and your eyeball is good, to approximate the perpendicular distance. If the fork is off the bike, you can block the steer tube so its level on a level table, then measure the height of the steer tube to the tabletop and the height of the dropout center to the tabletop. This is a direct measurement of the fork offset.

I like doing this. Maybe someday I'll buy better tools. I have some confidence in these methods, because I have access to a couple of early-80's Treks. I've measured them and compared to the website values on Vintage Treks, and found pretty good agreement. Allowing for several possible error sources (frame not matching the "spec," frame sizing ambiguity (c-t, c-c, c-lug top???), and of course my home "levelness" and my own inaccuracy (reading these cheap tools correctly!), I got ballpark agreement. I know Doug Fattic could do better because I've watched him, but I'm happy. I can compare frames.

Copyright 2006 by Ken Freeman

A VOCABULARY FOR BIKE HANDLING CHARACTERISTICS

This version of the vocabulary is based on discussions on the Internet Bridgestone Owners Bunch [iBOB] mailing list. It reflects the input and suggestions of many iBOB members. This is not intended to be read as a final product. Instead, it is a work in progress. We expect that many of the definitions should be merged, split apart, removed or entirely rewritten.