Date: Mon, 30 May 2005 23:39:15 -0500 (CDT)
Subject: front end structural failure
X-UID: 55
Content-Type: IMAGE/JPEG; NAME="img1591.jpg"
Content-Type: IMAGE/JPEG; NAME="img1592.jpg"

The epoxy failed on one of the front end kingpin assemblies. I had picked up 
the robot by the front wheels to get a feel for it. The rear wheels remained on 
the floor with maybe half the weight off the floor. This was enough for the 
epoxy bond to yield on one side. I didn't actually notice until moving the 
robot into the kitchen when one of the wheels was at a strange angle.

I don't feel bad about manhandling the robot like this. If it didn't fail now, 
it would have failed later in the field.

Here are two lessons about mechanical design.

1. If you take care to fabricate parts very carefully, then they may just
    turn out to be good enough. But they will still be off.

2. If you think that parts are overbuilt, in reality they will be just
    strong enough.

Another lesson is that epoxy, even metal epoxy composites, are not very strong. 
I did everything right and the bond just could not hold under a twisting shear 
load. My guess is that the shear yield strength is much lower than the tensile 
strength.

I'm going to make the kingpin assemblies out of single pieces of aluminum now.

I did learn some more from mocking up the steering before the kingpin assembly 
failed. There will be play in the wheels. This could cause oscillations. 
Anytime a controlled system has a feedback delay, there is the potential 
(inevitability) for this. Without experience in mechanical design, I didn't 
realize how much oscillation might arise.

Normally, I'd be very down about this. But this robot is kind of like a job 
now. So it's just one of those things that happens, especially when you don't 
know what you are doing. But I am going to sleep now instead of working all 
night.