Frame Testing

[Mark Kelly]

Interesting podcast on CyclingTips today on the subject of testing and custom bikes.

Begs the question: how many builders here do?

[Mike Maloney]

I have only tested one fork in a lab. I am basically a hobbyist so I felt it was a good way to measure durability.
Was set to send a frame in.. then decided to make my own frame tester.

Flickr

[Mark Kelly]

Like the integrated cycle counter. I rely on the PLC control software for the cycle count, it's a PITA. I think I'll add a counter like yours.

Smoked Out on Testing

[Mike Maloney]

Yeah I only use the control box I made.. no computer.

The counter was just a cheap one off ebay. It only goes to 99,999 but, that's my stopping point to restart.

[Mark Kelly]

I've been thinking about the contents of the podcast mentioned in the OP and have come up with a coupleof ideas.

One thing that got a bit glossed over in this discussion is that the fatigue tests specified in ISO4210-6 are a bit of a compromise between short cycle and long cycle fatigue, but this has an interesting consequence that was mentioned: both Nick Crumpton and Ben Shultz say in the webcast that they are still riding bikes that have been through the fatigue test, as am I.

This isn’t because we are crazy or cheap, it’s due to a phenomenon that was mentioned Mike McDermid some time ago, which is that composite fatigue failures follow a fairly extreme version of the bathtub curve: if a composite frame is going to fail in fatigue it will do so in the first few thousand cycles or after millions of cycles, but not between. The “bottom of the bath” is a near zero failure rate.

Another interesting part of the discussion is that there’s an inherent bind faced by all custom builders: where production builders can test a representative sample of each size of each frame and generate a statistical model to predict failure rates and thus safety for the production run as a whole, this is impossible for a custom builder because each frame is unique.

Someone in there mentions that even if we built a duplicate frame and tested that, this would still not work because it could not take into account small variations in layup and possible voids and inclusions.

Putting these two concepts together suggests an out. For those of us with in house testing capabilities, it is practicable to test every frame for a few thousand cycles on the fatigue tests. 10,000 cycles will be enough to get to the floor of the bath but not enough to materially affect longevity. This will only take a few hours per test.

So on a practical note, I am committing to testing all frames using the ISO4210-6 methodologies for pedalling forces fatigue, frontal forces fatigue and the impact tests before delivering them. The fatigue tests will be curtailed to 10,000 cycles.

[devlin]

......BUT......this would only be of benefit if you were pushing the envelope of design. ie. trying to build a super light frame. Yes? No? If you were building with a reasonable built in 'margin of error' ie. going for a balance of strength and weight then would you benefit from going through the whole testing routine for each and every frame? Carbon can already allow you to make some pretty light frames without compromising strength and not being a weight weenie myself Id just make them a touch 'overbuilt' in the first place.

[Mark Kelly]

I disagree completely, most of the reasons for my disagreement are in the podcast. Did you listen to it?

[devlin]

Not all of it. I fell asleep and haven't got back to it. I'll sit down and have a listen when I have a chance.