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).