My implementation is very simple, creating and initializing them directly in the setup section of the test sequence, and releasing them in the cleanup section. In this way, I have no problem controlling 30-40 various instruments and devices in parallel or batch mode at a testing station (I have actually practiced on multiple projects).
My understanding is that as a VISA handle, whether it is a standard SCPI instrument or a non-standard ordinary instrument, it is an object to me. For the entire testing process, the lifecycle of this object is from the beginning of testing to the end of testing.
If a device experiences an error during testing, the biggest impact is causing the current test to fail. As the current test aborts or ends, this error will also be eliminated within the final section as it is released. Even if I do not close the testing program, it will not affect my ability to switch to manual operation of the instrument or use other tools and software to cooperate with tracking.
