One other thought: When in doubt, go for more bandwidth. Massive data
storage is cheap, and when things go wrong in a rocket engine test, they
often go wrong FAST.
Sampling a channel at 100 hz can be fine when things are going right,
yet totally miss key impulse characteristics and timing relationships
when things go wrong. *After* the system goes "bang!" is a really
expensive time to realize that cheap faster sampling would have helped
you figure out what happened.
On 1/11/2017 8:52 AM, Edward Wranosky wrote:
Something else that I realized early is that start recording data on
all your channels well before you intend to test. That way if you have
an anomaly you have some data to go on.
Edward
