Some of you might find it interesting to see accelerometer data from a high-speed failure
When a flew a minimum-diameter rocket on a J420 with a scratch-built nosecone on Sunday morning, the nosecone failed right around burnout. The front end of the nosecone broke off and took my apogee deployment charge with it. The apogee charge was there to blow off the nosecone to give the rocket a drogueless descent, but with no apogee charge and with the back end of the nosecone stuck in place, the rocket came back down fast, nose-down, until the main fired the chute cannon at 700 feet when the rocket was going 260 feet/second, which blew out the chute, causing a pretty hard landing.
The red line is the axial acceleration, and the purple line is the lateral (sideways) acceleration. Just when the motor was starting to burn out, but while it was still putting out enough thrust to keep the rocket accelerating, you can see a little sideways force where the purple line starts to go down to about 4 Gs sideways. The speed then was 2515 feet/second (Mach 2.2) Then bam, the lateral acceleration goes to 100 Gs one direction and then the other, and the axial acceleration went to 60 Gs of drag, which is a little more than the acceleration at peak motor thrust. This is when the smoke trail took a sharp jog sideways.