[Closed] crank spider failure - why?

Is this a test? My vote goes to a non standard chainring that is not properly supported by the lip on the rear of the spider. Resulting in the bolts and those little tabs taking most of the pedalling load.

You get the same problem with putting wheels on a car when the centre bore is too large for the hubs. The wheelbolts become load bearing when they are only designed to hold the wheel to the hub.

Its not a test. There is an argument going on about it on the tandem site I pulled it from and knowing that many folk on here have good knowledge I am simply interested in what folk think has happened at least in part so I can avoid in future

They have exceeded the design's limitation. That looks to be the standard cinch 2x spider. I'm betting it's not rated for tandem use.

Bolts must have been clamping reasonably well as the broken bits are still all clamped to the chainring in the normal orientation (they haven't moved or fallen off).

Could also be some weird local stresses due to the additional loads from the timing chain in the granny ring location. The front rider is probably strongest and is applying a force to the spider in an unusual location.

Would be an interesting fe analysis / strain gauge project.

Replace it and move on. It was never designed for a tandem in the first place (although I do like "rhd" tandem drivetrain as they stop you eating the rear bbs so quickly as the chain tensions partly cancel out).

Maybe the ring wasn't seated properly when the bolts were tightened.

Can't be sure from that photo, but also looks like he has bolts on the outside and nuts on the back. Are the nuts protruding through the ring into crank holes to dowel / spigot in place correctly? Or is it only the 8mm bolt in a 10mm hole causing point loading / stress raisers etc?

Doesn't look like there is much clearance between the chain and the top edge of the spider arms at the chain ring bolt locations. Maybe If the chain was contacting the aluminium alloy part of the spider above the bolts then that could have been imparting a point load that could, over time, initiated a crack which then propagated through fatigue leading to eventual failure. It does look like the failure point is at the 12 o'clock position at each spider arm possibly where the chain might have been contacting the arm lug.

It just looks like it has been overloaded to me. It would likely have been a cascading failure where as soon as one bolt hole gives way, the others follow suit.

Not much meat around the chainring bolt holes on those Raceface spiders (assuming this is the right one) to transfer the weight of a couple of hefty riders to that rather chubby rear tyre.

2 x 80kg riders on 175mm cranks = ~5.4kN at 52mm radius (half of 104mm PCD). If that is pure torque, it would be shared equally between all 4 bolts = ~1.3kN tangential force at each bolt. At least half of that would have to be taken by the thin sliver of metal at the outside of each bolt hole.

Yield strength of 6061 (say - I don't know what the cranks are made of) ~270 N/mm^2

Minimum material required to avoid yield is 4.8mm^2 per bolt (in theory shared equally each side of the bolt hole, but won't be perfect).

I suppose there might be 4mm^2 on the outside edge of the holes in that photo.

In reality, it won't be a pure torque - a disproportionate amount of the load will fall on one bolt, and without the ideal conditions assumed, and with dynamic loading and fatigue to consider, it would not be surprising to see it fail in that manner. (IMHO).

One could possibly learn more from a careful examination of the failure faces on each of the bolt holes.

The only way I can see that particular crank failing like that is if the rear rider has tried an absolute herioc standing start and the front rider has stood firm on their cranks and not allowed theirs to start rotating. Judging by the way the tabs have seperated anyway

cheers folks

Looks like the load has exceeded the friction generated by the bolts between the spider and the ring. (Why the bolts are still tight) Once the ring slips the bolts will start transferring load to the spider.

I would assume it was fatigue failure caused by the load cycling between overload and design load. The bolts loading  spider every time the ring slipped then unloading when force dropped back down and the ring was gripped again.

To overcome the friction and break the material in a single load event would require quite a lot of force.

Hard to know without seeing the fracture faces

Potentially incorrect shouldering inner chain ring bolts but more likely the lack of material on the bolt holes and torque issues/ stresses beyond limits.

Very complicated answers. Simple answer- Raceface aren't very good at cranks. I had an old set of Rides fail in the exact same way on my Soul and I have twiglets for legs, and it's not the first time i've seen this failure on newer ones. The switch to spiderless/single ring suited them very well

Difficult to tell without closely looking at the broken bit but possible fretting round the bottom right bolt. Poor fitting or looseness is often the root cause of failure with bolted joints IME and the first thing to look for when presented with a broken machine of any kind.
Chainring is non standard so potentially poor fit but would need checking.
Mainly wrong tool for the job. That is a 2x lightweight mountain bike crankset being used to carry both drive and timing chain. This will change the loading on the chainset. Might have got away with it on a crossover drive but lightweight kit is designed to be used in only one way.
Buy proper tandem parts is probably the moral of the story. Like complaining it snapped after drilling it out to save weight.