Forum Migration Notice
Update (2026-01-21): The user forums are now in read-only mode pending the data migration.

Update (2026-01-12): The user forums will be put into read-only mode on the 21st of January, 00:00 CET, to prepare for the data migration.

We're transitioning to a more modern community platform by beginning of next year. Learn about the upcoming changes and what to expect.

Force offset after FCCalib

grth46
grth46 ✭
in Robot Controller
Hello together, I am working with an IRB6700 and IRC5 Controller. Currently we are using this robot for stiffness testing applications with automotive parts. For this kind of testing there is a customer defined regulation for the measurement procedure which states, that the tool must achieve contact with the part while not surpassing a force of only 1 Newton in tool Z direction.

We have already found some ways to actually get closer to this target, but we have noticed, that our force sensor from ATI always has a force offset after calibration close to the measuring position. So the static noise from the signals as seen in TuneMaster (207 to 209) are not symmetrical to the baseline. I know we will be limited by the accuracy of our sensor at some point, but from my understanding this offset should not remain after the calibration.

We already contacted the ABB support for this, but they only mentioned the LoadIdentification routine or FCLoadID(), but running these has not solved the problem.

I'd be happy if anyone shared their experience or some thoughts to this.

Tagged:

Comments

  • grth46
    grth46 ✭
    Has anyone an idea why this is happening? We've developed some workarounds for all the other problems. Sadly I can't figure out why this is happening... For context: It is not sensor drift after calibration. Directly after calling FCCalib and FCAct there can be force offsets up to 4 N while robot stands still. Does anyone have insight on how exactly the calibration is done internally? Like does it just take the current reading once and use that as adjustment value or does FCCalib take a mean value of a specific number of sample values?

Categories

In this Discussion