search

aesim-tech / simba-project

Roadmap & issue tracking
7 stars 0 forks source link

Possible bug with ZOH and fixed time-step #341

Closed guillaumefontes closed 11 months ago

guillaumefontes commented 11 months ago

Example with a sine source @50Hz with a probe. Let's try to sample the output of this probe at 1e-3s.

With predictive time-step, it is OK. With fixed time-step @1e-7s, it gives strange values for the first periods of the sine.

image

You can find simulation file below. Results considered with beta version of 14-09. possible-bug-zoh.zip

BR

manu-aesim commented 11 months ago

This bug was caused by a numerical rounding error and should be solved in this release:

https://aesimtech.s3.us-east-1.amazonaws.com/artifacts/AeCrxoW3vjiMF3wakZNSLJSalO2bGwVO/simba.exe