This bug is pretty obscure, and involves the speed chop with multiple samples, as well as pulsing a pattern chop, that feeds the reset input.
If I hook up a speed chop with one sample only and with the speed per sample toggle turned on and pulse the amplitude of the second chop input I get a sharp and consistent ramp over time:
Now I can take that exact same network, and just change the two inputs to the speed chop to 2 samples (or more) instead of 1 and it looks something like this:
If you look closely, the bottom of the ramp is flattened out, and the width of this flat area is equal to the duration of the parameter’s pulse - so it has something to do with that.
Here’s a file with 3 different configurations of this topic, the middle one being the bug:
speedChopBug.3.toe (4.4 KB)