Rolling around in the mud doesn't solve problems. Mike and I had what I consider a nice chat on the phone today and we will see if we can figure out what happened.
I am not aware of anyone making a rod out of a square tube but in the 1920's, Packard made rods out of round tubes that had pin end welded on one end and a housing bore welded on the other.
I have an example of why an I-beam is not the hot ticket in high power applications and will try to up load it.
Tom
That's what I like,intelligent dialogue.
Glad you talked to Mike.
He is an insightful guy and a gentleman.
He does scare me sometimes though.
I'm looking at the pic's presented so far and I see a common problem:
Parallel beam sides.
I don't know if you remember your wave theory from physics class,but you have the start,the peak,the crossover point,the trough and the end at zero.
It's known that metals do tend to have a resonance "Q" or natural vibration.
Put a series of sensors along the beam of a set of rods with parallel sides and you will find peaks and troughs around 1/4 & 3/4 marks along the length of the parallel sections,when the natural resonance is excited.
The higher the inpulse applied,the greater the excitement or motion.
Looking at where Mike's rod failed in the pic',I would say it was at the 1/4 point of the most highly stressed end of the rod.
Shouldn't be hard to excite that resonance somewhere in the RPM band.
Couple that with the extreme loads .......
Two things might have saved that rod [and Mike's wallet]:
1.Greater broad face width on the big end [to withstand the swinging motion].
2.Non parallel sides to minimize any standing waves [lower Q].
I realize the wave issue may not sink in with most people,but the issue has it's roots in structural designs such as buildings and bridges,especially in earthquake zones....think Tacoma bridge.
Structural reinforcement doesn't work too well when it's wobbling around like a drunken sailor.