by **cncdrive** » Fri Dec 22, 2017 10:40 pm

I have also never used G41/G42 myself, but now that we implemented it into the UCCNC I fully understand it's working and also it's limitations.

Because it is clearly defined how it has to work mathematically and we wrote the algorithms I see are limitations. The limitations comes from that the codes are based on one movement look ahead and therefor there can be cases especially on path with lots of short and different direction movements and especially with large offsets compared to the movements dimensions then the offsets will be not perfect.

It is then not perfect because of how it has to work.

CAM softwares calculating the offset with totally different mathematical algorithms which algorithms looking at the path as one big piece and doing the calculation on all the path instead of just looking one movement ahead and going along the path.

Ofcourse such algorithms can't work on a CNC controller like how the G41/G42 can, because while the G41/42 needs only a small amount of time to calculate that one movement at a time, the algorithm in a CAM software may require a serious amount of time to complete, because it looks the whole path and so the calculation is time costy and so it can't work in "realtime" is why the simple and fast G41/G42 still remained the standard on CNC controllers eventhough it has lots of limitations because how it works, how it has to work.

So, I'm also prefering regenerating the path with the CAM over the G41/G42 in most cases, what I think the G41/G42 is perfect for simple paths and/or small offsets for example tool wear size offsets.

Maybe in the future when computers will have a so high computing power that advanced ray tracking algorithms which are implemented in CAM softwares for offset calculations can be computed in a short enough time then the G41/G42 will be replaced with these codes, but I also think that it will probably take a few decades if it will ever happen.