Is it a good idea to use ChatGPT to optimize feeds and speeds?

Large language models are simply going to extrapolate based on examples in the training data and whatever the “temperature” and similar settings cause them to output.

That said, the feeds and speeds in Carbide Create are quite conservative, and a spindle affords a lot of torque which machines running a trim router lack, and the HDZ and larger V wheels and wider belts of an SO4 are quite the rigidity improvement over an SO3 w/ a belt-drive Z-axis.

There are number of videos at:

and we’ve had quite a few discussions about feeds and speeds here:

One technique which I used successfully back before Carbide Create was available is that from:

Note that we had a pair of blog posts from the late Bob Warfield:

https://carbide3d.com/blog/feed-and-speeds-part-1/

https://carbide3d.com/blog/feed-and-speeds-part-2/

and his “CNC Cookbook” site has been preserved:

1 Like