Vessel.acceleration: what is going on?

[martincmartin]

I did a little test: launch a rocket straight up, and a few seconds later (before drag can become significant), cut thrust (by pressing X).  If I compute acceleration by subtracting consecutive Vessel.velocityD and dividing by time, I get what I would expect: between two frames, the acceleration instantly changes direction, and the new magnitude is 9.799.

However, the "acceleration" field barely changes:  See my log messages below:

accel: (-3.475, 0.034, 2.243)(4.136), computed: (-3.487, 0.037, 2.255) (4.153), vel: (-29.271, 0.098, 19.660)

accel: (-3.477, 0.034, 2.244)(4.138), computed: (-3.489, 0.037, 2.255) (4.154), vel: (-29.340, 0.099, 19.705)

accel: (-2.900, 0.034, 1.854)(3.442), computed: (8.082, 0.016, -5.541) (9.799), vel: (-29.179, 0.099, 19.594)

accel: (-2.323, 0.033, 1.465)(2.747), computed: (8.082, 0.016, -5.541) (9.799), vel: (-29.017, 0.100, 19.483)

accel: (-1.746, 0.032, 1.075)(2.051), computed: (8.082, 0.016, -5.540) (9.798), vel: (-28.855, 0.100, 19.372)

(The single number in parentheses is the magnitude).

This is while Krakensbane.GetFrameVelocity() is still zero, so there's no floating origin problem.  It's only about 4 seconds after lift off.

Is the acceleration field some sort of time-averaged value?  Once I shut off the engines, nothing on the vessel is actually accelerating at that rate.  What is going on?

I'll be happy to update the API docs with any answer.

[Padishar]

4 hours ago, martincmartin said:

Is the acceleration field some sort of time-averaged value?

Basically, yes.  Try Vessel.acceleration_immediate instead...