Frame Design

Designing a frame requires maths . . . not much though, that's the good news; it's GCSE maths. You can buy expensive software such as BikeCad to do this for you; this is very good, with a free, on-line version . . . but what do you learn about the design process when you use a black box? So I thought why not design from scratch so I understand how it the parts fit together and interact; and from that understanding I should be able to build a better bike.

I didn't come by this idea through divine inspiration. There is a spreadsheet application out there called BG101 which is a highly functional Excel spreadsheet that calculates the tube angles, lengths clearances, plots the final frame design, templates for the tube mitres etc. etc. . . . it's pretty good. But, as with every spreadsheet I've ever come across which is written by someone else; and that is exaclty the problem . . . it's written for someone else. But then I thought 'this is actually really simple to do for anyone who knows about spreadsheets with simple maths and the ability to plot a graph', so I set about writing my own.

This will of course be useless to you because I wrote it for me. So write your ow - it's not too hard and you will learn a lot. This is what I did:

1. I plotted the front triangle as a series of right-angle triangles using the trigonometry rules of tan, sine and cosine of the ST and HT angles, or their complements, and the c-c lengths of the ST, TT and HT. For example the ST had co-ordinates of [0, 265] and [0 - ST x cos(ST angle), 265 + ST x sin(ST angle)].
2. Once these co-ordinates were calculated the DT co-ordinates were automatically fixed, so the HT-DT angle and DT length were computed.
3. I applied the same logic to calculate co-ordinates for the CS, SS, crank, toe overlap, fork etc. using the appropriate lengths, angles and chosen values for, rake and trail.
4. Then I used these co-ordinates to plot a graph with lines representing the c-c tubes lengths in Excel as shown below.

Once done, I could alter the lengths and angles and the see the effect on the plotted graph instantaneously. After some lengthy playing around in this way I have confirmed some obvious and not so obvious conclusions:

1. Small lugged frames are tricky because there just isn't enough HT to fit the lugs on.
2. These design limitations can be reduced by using smaller wheels such as 650b. This is because you can extend the HT and reduce the fork length, but only if you don't use huge tyres otherwise the wheel radius ends up as big as 700c.
3. Lugs and BB shell only come in a limited selection of angles. You can play arounf with the HT and CS length to solve this but then compromise elsewhere is necessary, or you may have to file them to get the angles you want, but only by so much. Now I know why people life to fllet braze . . .
4. Commercially available forks come in a limited range of sizes, especially so for steel. Make your own!

 

Frame Fit - update

Phil got me thinking about my bike fit. In the last post I described the CONI method, which has limitations in that it is largely based on the dimensions of an average person. So, I downloaded the Kinovea software (Kinovea, 2013) as suggested, dusted off the old family camcorder and (hated) turbo-trainer, and made some dynamic measurements of me actually riding the Tricross. The result is shown in the video above; and what a great piece of software it is, with the facility to apply a 'bikefit' tool which overlays a skeleton to measure body angles. This being the beta version there are some glitches and it is necessary to stick pieces of white tape to your joints to help the software 'pin' the moving skeleton in place - this does not always work, as evidenced by the necessity to pin one point to the front axle and the points on the arms not being fixed. Nevertheless, it is still possible to make accurate measurements at any pojnt in the pedal stroke by taking snapshots. The advantage of making dynamic measurements (i.e. cycling rather than static) is that they are more representative of the actual angles when pedalling.

In order to make sense of this a bit more searching on the web yielded the BikeDynamics website (BikeDynamics, 2013) which has a very useful summary of the expected body angles as follows:

• Saddle height should be adjusted so that knee angle is between a minimum of 65-70 degrees and a maximum of 143-148 degrees, with the caveat that people with tight hamstrings (generally men) might prefer a maximum angle of 138 degrees and women up to 150 degrees because they tend to have looser hams. Snapshots taken of me at 12 and 6 o'clock below show that I am within these bounds.

• Saddle setback can initially be adjusted using the KOPS (knee over pedal spindle) method. BikeDynamic point out that there is no biomechanical reaon for this but it serves as a useful ballpark measurement to start. Analysis of my position on the Tricross seems to indicate that my knee joint is too far forward infront of the spindle. A slight forward position is not, of itself, a problem - it would be if the knee was behind the spindle - but it contributes to my forward position on the bike which feels cramped. The obvious solution is to set the saddle further back BUT there is no more travel on the rails of my Brooks B17 to enable this (note to Brooks to address this failing). Using a seat post with more set back might solve the problem, but since I am customising my own frame I might as well go the whole distance and decrease the ST angle to 72 degrees. With a ST design length of 53.5 cm this will have the effect of moving the saddle back by 1.3 cm and dropping it by 0.4 cm for the same ST length. This will also reduce the the minimum leg angle and increase the maximum leg angle slightly, but this can be dealt with by readjusting ST or crank length if necessary.

• Torso angle is another parameter which requires attention. BikeDynamics recommends an angle of 45 deg for touring and 34 deg for sportive/racing (measured between torso and horizontal) As you can see, I have a very upright position with an angle of 57 deg (not shown on the picture). Moving the seat back will reduce this but I clearly have much room for adjustment - providing my not very flexible lower back can cope that is! Another consequence is that the torso-upper arm angle will increase from 60 deg closer to the 75-90 deg recommended range.
• In conclusion, reducing the ST angle will: move the seat back and bring my knee over the spindle; decrease my torso angle; increase my torso-arm angle; and move my centre of gravity back a bit, so taking weight off the bars. . . that's the logic. Plus some other fiddling around with stem length, bar position and saddle height of course.

References

BikeDynamics (2013). Homepage. Available at http://bikedynamics.co.uk/index.html [accesed 19/11/2013].

Kinovea (2013). Experimental version 0.8.21. Available at http://www.kinovea.org/en/downloads/ [accessed 19/11/2013].

 

Frame Fit

One of the major reasons for buliding a custom frame is so that it can be measured to fit. Commercially produced bikes are, understandably, produced for the 'average person', with the ST length being the de facto criterion used for sizing. In the past, when most bikes had horizontal top-tubes, this measurement was comparable between bikes, so standover height for the prospective buyer was a simple calculation. Now, however, with sloping TTs this measurement is practicaly meaningless so any realistic idea of fit usually requires a trip to the LBS to try out a prospective purchase. In reality, ST height is not the most important measurement; provided there is sufficient standover height, to ensure the safety of the family jewles in the event of an unexpected forward dismount, the ST length can be easily changed by changing the saddle height. More important is the TT length combined with the stem length, because this determines the reach, which has a major influence on upper body position and comfort. However, upper body measurements are not considered when selecting a bike unless you pay for a professional bike fit which might cost upward of £150.

So what is the secret? There are several methods of bike fitting which have been used, mostly based on the Italian C.O.N.I. method (C.O.N.I., 1972). Many framebuilders developed their own methods based on this, or by combination of their own experience as cyclists and customer feedback.

In my search for the perfect algorithm I have come to the conclusion that there is no such thing because everyone is different (surprise) and has slightly differing requirements from their bike, so I tend towards a method based on some general principles allied with previous riding expiecnce and an understanding of the purpose to which the bike is to be put - this seems to be the approach adopted by custom framebuilders to greater or lesser extent, with the following general principles seeming to apply:

1. Take measurments of the inside leg, sternum, and arm length.
2. Decide on the ST length based on this measurement.
3. Use tables and/or various formulae to get an estimate for the TT and stem length.
4. Measure one of your bikes that you already own and feel comfortable on, and compare to the measurements above.
5. Decide on HT angle to suit the type of riding.
6. Decide on ST angle to suit the type of riding and your physiology.
7. Arrive at an optimal solution using the above data and experience based on riding thousands of miles under a variety of circumstances.

As you can see, this is not a terribly helpful set of instructions, but is an accurate reflection of my current state of knowledge. In summary, I have concluded that there is no magic formula and experience of bike riding plays a not insignificant role. That said, for this build, I followed this procedure:

1. Use 'Method 3' in the C.ON.I. manual
2. Take anatomical measurements: Inside leg, A = 83.5 cm; Height, A1 = 176 cm; Arm length, A2 = 63 cm; Foot length = 26 cm, as shown in the image above.
3. Calculate ST length using the formula F = A - (C + D), where C is the saddle extension and D is the crank length. Using the relevant values from the table in th manual this computed to a ST length of F = 83.5 - (14.5 + 17.5) = 52 cm.
4. Calculate the TT length using the formula H = F +\-G, where G is a factor to be added or subtracted depending on the size of the frame. Hence, H = 52 + 1.5 = 53.5 cm.
5. This ensures that the ST and TT lengths are within proportional limits for the frame. The advice is to then adjust the effective TT length for large or small riders by adjusting the length of the stem (B) and saddle position to obtain the correct reach.
6. Choose a ST angle to suit the type of riding and rider physiology - I am larger in the upper body than the pelvis so decided on a shallower ST angle of 72 deg to bring my centre of gravity back a bit.
7. Choose a BB height to suit the type of riding - I wish to use the bike for audax/touring so chose a BB heigth of 26.5 cm which is at the lower end of the limit (26 to 30 cm) because a great deal of clearance will not be required.
8. Choose a HT angle to suit the type of riding. I chose a compromise of 73 deg because I want the bike to have responsove steering but also to be comfortable. Fork rake, trail and tyre size will also play an important part in this, but more of that later.

Having done all this, for comparison, I then measured the bike which I currently use for touring, a Specialized Tricross, and got the following measurements (taking into account sloping TT to obtain the effective horizontal TT length): ST = 53.5; TT = 53.5; BB height = 26.6; Stem length = 9.7; ST angle = 73.5 deg; HT angle ~73 deg. I have used the Tricross for a number of years and understand pretty well how my body feels on it. First, the ST length is longer, but this is not that important because it is well within the range of saddle height adjustment. Second, the TT length is exactly the same, but I find that my body feels too far forward with too much of my weight on my arms, so the intention to reduce the ST angle from 73.5 to 72 deg would seem sensible, with the option of changing the stem length and moving the saddle to adjust for reach. Bearing in mind that the Tricross is a cross bike and not really desogned for touring, these conclusions seem to make sense, to me at least.

In the next post I will describe the full design using my 'homemade' spreadsheet.

References

C.O.N.I. (1972). Cycling, Central Sports School - F.I.A.C., Rome.

 

Wisteria

The wisteria framed door

whence came the coffin;

a top-hatted gentleman

of undertakers, and journey

to that final place.

‘Do you want to be buried

or cremated, Dad?’

‘Surprise me,’ I said.