Got the idea to this question from this one, guiding in framebuilding.

I cannot jump directly to welding before I have a cheap working prototype, due to tight budget. In architecture of houses, they use balsa or bass woods to build scale models. Bicycles are in a way more complicated to houses that you have gears, cogs, anti-vibration-structure and equipment -positions to design. When you suggest prototyping, please, tell me how you use it? For example, how can you kill vibration in the very beginning?

  • Are you thinking of wooden models? Commented Feb 18, 2011 at 5:47
  • 2
    What problem are you trying to solve, or what experiment are you trying to perform, by making a prototype? E.g. architectural models are trying to solve problems like, "What will it look like?" However models, which have a different size and are made of different materials than the finished product, don't necessarily model/simulate the performance of the finished product (e.g. vibrations).
    – ChrisW
    Commented Feb 19, 2011 at 2:12
  • 1
    @hhh I'm not writing it as an answer because it's not an answer: instead it's a question about your question. Your problem isn't (or shouldn't be) irrelevent. Any answer is relevent only to the extent that it addresses your problem.
    – ChrisW
    Commented Feb 19, 2011 at 5:35
  • 2
    @hhh -- Meta (not here) would be the place to discuss this: please see Broad questions or specific?
    – ChrisW
    Commented Feb 19, 2011 at 14:16
  • 1
    @hhh If low-cost is really the only citerion, the lowest cost might be a pencil sketch, or perhaps a model made from clay or pipe-cleaners.
    – ChrisW
    Commented Feb 22, 2011 at 17:31

7 Answers 7


This page says,

Reynolds is also rolling out to its customers who are typically custom framebuilders without access to sophisticated computer modelling tools a programme called eReynolds FEA (Finite Element Analysis) which will enable them to 'build' their frames in virtual form to test the feasibility of tubing and joining options before expensively committing to physical prototypes. Developed in conjunction with Birmingham University, Murphy says "it's a way of helping builders offer competitive frames especially now that cyclists are developing a whole new appreciation of the upsides of Reynolds metals."

  • +1 that is totally right, the cheapest/best way is certainly to start with FEM -modeling (reusable, easy to debug). Now have to look for some software, odd that I did not think this before. I once saw some list of them but they were closed paid-only software, any ideas whether some open-sourced FEM -modelling tools? (btw...FEM is used all-over the industry so it will work here, no idea though whether there is some sort of ArchiCAD -style software with ready parts -- would speed up the designing a bit...)
    – user652
    Commented Jan 8, 2012 at 10:14

What actual problem are you trying to solve, or what experiment are you trying to perform, what question are you trying to answer, by making a prototype? E.g. architectural models may try to solve problems like, "What will it look like?" and "Is there enough space for the plumbing?"

However models, which have a different size and are made of different materials than the finished product, don't necessarily model/simulate the physical performance/behviour of the finished product (e.g. vibrations).

For example, how do they plan weight distribution in their models? How do they beforehand see vibrational problems?

To experiment with physical properties like vibration, I'll guess they use full-scale prototypes, made of actual metal: identical to the finished mass-manufactured article, except hand-made one-offs instead of mass-produced.

You might be able to get by with a scale model (e.g. quarter-size), so long as it's made from a sufficiently similar material (e.g. welded steel). Using anything else (e.g. LEGO blocks or balsa wood) would not, I guess, be a useful prototype, because its mechanical properties are too dissimilar.

Another possibility might be a computer model.

On the other hand (as an example of a different problem being able to use a different kind of prototype) if you were trying to model the wind resistance instead of the vibration: because a thing's interaction with the air depends on its external/superficial shape, more than on mechanical properties like stiffness, for that you could use a different material.

In his autobiography, Freeman Dyson asserted that the reason why motorcycles are reliable is because there have been so many different generations of them since they were invented, with each new one building on and correcting the problems with the previous versions (he contrasted that with nuclear power stations, saying that there haven't been 100s of generations of those).


Echoing ChrisW: What actual problem are you trying to solve, or what experiment are you trying to perform, what question are you trying to answer, by making a prototype?

Your list of goals suggests that you're more interested in attaching computers to the bike than riding it, so it might be better to start with a picture of a bicycle, put the computers in place, then add whatever hardware you need to activate the sensors. That way you can get the computer hardware and software sorted out before you start building the bike.

When it comes to bicycle prototypes, models are rarely used except for areodynamic modelling simply because the cost of a complete bicycle is so low. The cost of a prototype is mostly labour, and that cost isn't affected by the scale (within reason, a 1:1000 or 10:1 model of a bicycle is going to cost a lot more). Most of your prototpre requirements can be addressed by bolting cheap bike parts together.

If you don't want to buy welding or brazing gear I suggest using straps wrapped around the frame tubes, probably metal hose clamps and flattened tube ends bolted together. This will leave you a lot of slop compared to a welded joint but they are cheap to do.

If you want to build a bike that can be ridden any distance you will have to buy something to make proper joints with. Composites are more expensive overall but you can pay as you go - only buy enough epoxy for the current project. Using twine rather than synthetic fibre will be cheaper but somewhat less robust. Oxy-acetelene gives you the option of welding or brazing (or soldering), but makes working with aluminium difficult and realistically limits you to steel. Electric welding is more flexible but the equipment is more expensive. Stick welding is possible but very difficult on the thin wall tubes that bicycles typically use. GTAW/TIG welding is most flixble but also requires the most skill. MMAW/MIG welding uses cheaper equipment (at the low end) and you only need one hand to weld, but switching between materials/changing electrodes is tedious. Cheap MIG gear will not do nice-looking welds, but operating the expensive gear is a highly skilled task.

  • I basically need a breadboard [1] for a bicycle or quantified central parts that are easy to assemble/disassemble, some sort of test bed that I can reuse. Is there anything like that? I need a variety of positions to test where I can put things. [1] ladyada.net/images/arduino/halfbb.jpg
    – user652
    Commented Feb 21, 2011 at 22:09
  • 1
    A bicycle and duct tape. Seriously. Unless you're making major structural modifications to the bike, just duct tape stuff on where you need it. You could possibly make a bicycle out of some sort of full-size Meccano but I suspect it would not be very rideable.
    – Мסž
    Commented Feb 21, 2011 at 22:53

It would be helpful if you outlined what your goals are. Are you trying to create a new style of bike from scratch? Are you tweaking an existing frame/design for comfort? What are expecting to get out our your model? There are also physical modeling tools like Legos or erector sets, and virtual computer tools like AutoCAD or Rhino. The tools depend on what you are trying for.


You could try making a bamboo bike.

The materials required:

  1. Bamboo of 2 different sizes.

  2. Hemp twine.

  3. Epoxy resin.

  4. Donor bike for parts.

You'll need a hacksaw and or a dremel or angle grinder to cut the metal parts you need off of the donor.

I suggest you do some web searches for more information.

  • +1 good opening, it is cheap and reusable. If your bike fail, you may use bamboo elsewhere like in fire. Distadvantage is that everyone has not bamboo but don't worry about it, your answer is perfect. It may help other innovators.
    – user652
    Commented Feb 19, 2011 at 13:41

For Physical prototypes

  • hot glue -gun to attach the elements
  • physical blocks for example by Legos, also other
  • Neodymium magnet toys for example by Magnetix, rare-earth-magnets and industrial magnets
  • breadboards for electric prototyping
  • wooden sticks and wooden elements
  • plastic, Super -glue, plastic bender, --

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.