Monday, February 09, 2015

Major setback on the engine

When torquing the bolts for the connecting rod I felt one bolt became a tiny bit "mushy". Wishful thinking kept me from examining it right away. But the next day I did, and as expected the bolt had shredded the threads.

I contacted Sonex, and they told me to weigh the rod, and they should see if they could find a new one within 1 gram. If they can't find one, I have to get a new set of 4. They would offer me a set for the price of 3 single, hmmm.

Looking a bit closer at this, the rod is modified by Sonex. A rather large part is filed off, probably to prevent interference with other parts in the casing due to the larger stroke. The modification has resulted in a much shorter thread length than originally. The thread length is only about 4.5 mm.

Enough thread length is a difficult subject, depending on lots of factors and the application. Common engineering rule of thumb for steel, is a thread length of 0.8 to 1.5 times the diameter. The reason for this is, in case of overload, the bolt shall break before the threads. Threads in a piece of machinery are very difficult to "replace" once they are shredded, but a bolt is easy to replace.

A nut has typically a thread length of 80% of the diameter (larger for small sizes and shorter for larger sizes), and the reason behind this length is to be sure there are enough threads so the bolt breaks before the threads. You don't want the threads to be the critical part of a design.  More than this length does not improve the thread strength further. A nut and bolt usually have equal material properties. Female threads are stronger than male threads, due to larger diameter and more material for each thread.

The diameter of the threads on the bolt is 9.5 mm (seems to be 3/8 inch bolts instead of some metric dimension for some reason) . This means a minimum thread length is about 7.6 mm if a nut is used. But, since it is not a nut, it could very well be be 9.5 to 14.3 mm according to normal engineering practice. Originally the thread length seems to be about 8-10 mm, but this is difficult to say exact, because I don't know exactly how much material Sonex has trimmed off.

I did a more exact calculation for this bolt. The bolt main diameter (D) is 9.525 mm. I measured a pitch (p) of 1.077 mm/thread.

The tensile stress area for the bolt is:
At = pi/4 * (D - 0.938194*p)² = 56.94 mm²
The thread shear area is:
Ass = 0.5 * pi * dp * Le
where dp is the pitch circle diameter of the thread, dp = D - 0.64952*p and Le is the thread engagement length.

To be sure the bolt breaks before the threads shred, Ass must be twice that of At, which means that:
Le = 2 * At/Ass = 2 * At / (0.5 * pi * dp) = 8.21 mm.

Simply spoken, the correct minimum length for the thread engagement for this bolt is 8.21 mm, or at least very close to this number. With a length of only 4.5 mm, there is absolutely no margin for error of any kind, and the threads have to be 100% exact. This is not the case here, and the original length of these threads are 8-10 mm before Sonex decided to modify them. The original thread engagement length is there for a reason.

I'm not sure what to do about this. I don't think the bolts and threads themselves are too weak. They are probably made for a much higher RPM than 3500. But the torque used to fasten them seems to be too much for their own good, with this way too small engagement length. The threads in all the other bolts could also be weakened, and this is not good for fatigue, not good at all.

No comments: