Maximum Axes Acceleration

[ThatguyDave]

What would be an advisable amount of axes acceleration for a 96 Fadal 4020?

Without opening up a huge can of worms discussing the merits of control retrofits, look ahead, ect, just how fast should I program small circles on my machine?

My understanding is that when running under a G8 code, all feed G codes are performed at the specified feedrate with no hesitation between moves so it it should be simple to calculate just how much axis acceleration is occuring as the machine negociates around a corner.

FYI Centripetal acceleration is calculated by (velocity of machine^2)/radius of move.

I am assuming that while my legacy control has inherent limitations when processing point to point programs, the inherent arc commands in the control are processed sufficiently fast so that the control's own capabilities will not hamper the cutting of even the smallest arcs. Am I correct in this assumption?

I program using Onecnc which like many cam softwares offers pocketing functions which generate programs in which whenever possible arcs are connected in tangent to either other arcs or to lines which minimizes wear to the machine and in theory allows you to calculate exactly how much the table is being accelerated in individual moves.

I'm half asking this question out of curiosity and half because my machine sort of shutters when negociating tight toolpaths into corners. The shuttering certainly doesn't seem like a good thing but I'm not sure that the "impact" of changing directions is any more severe than would be encountered during traditional slow speed movements under a G8 command.

Thanks in advance for the feedback. Am I the only one pondering this? Dave

[iess525]

Is it an AC or DC machine?

[ThatguyDave]

Last year of DC drives. To open up another can of worms how much better would axis acceleration be with AC drives. I know AC drives have higher top ends and therefore can be pared with tighter threaded ballscrews but I still can't image a 97 4020 being that much better than my machine.

[Delw]

I have wondered this as well. I know I run 150 IPM back and forth and corners etc etc using g8. I dont like it , dont like the hard thumps from one direction to the other. what I started doing was running the program in more of a circular then zigzag and it helps at least in sound.
I never had a size issue except in the "Z" and that was a rebuilders screw up.
if I have alot of x- x+ lines I take the g8 out, it does slow the program up quite a bit.( alot meaning .0005 step overs and a y distance of 5+ inchs. hogging out 1/4" or more)


Have your program start in the middle and work out in a circular motion, or start from the out side. its much better for the machine, on your roughing passes put rads on every corner( biggest you can)

There are times mine sounds like a jackhammer.

Delw

[smithgrind]

Originally Posted by ThatguyDave I program using Onecnc which like many cam softwares offers pocketing functions which generate programs in which whenever possible arcs are connected in tangent to either other arcs or to lines which minimizes wear to the machine and in theory allows you to calculate exactly how much the table is being accelerated in individual moves.

I just recently bought Onecnc and am still learning it right now, I haven't even posted a program over to the 4020 yet. Is there a setting in there for setting the acceleration?

I have never liked the way the Fadal is so clunky changing dirctions compared to my linear machine. I'm not sure if it's the box ways on the fadal or if it's common to many box way machines. I know I've got about .0005" of backlash and I've never adjusted my backlash or gibs since I got the machine a few month ago. I always hoped it would get better with some tweaking but apparently not.

I've got the AC drives and can't imagine that it's drastically different between DC or by now you would hear "Whatever you do stay away from the DC Fadals"

[iess525]

Are the axis motors 4050 or 6000?

[ThatguyDave]

I'm not positive but I believe they are 6000 on the x and y. The p/n on the x is illegable but the Y is 6M60008C.

Smithgrind, congrats on the new mill and Onecnc software. What version do you have. I agree before sending cam programs to my Fadal it seemed extremely clunky, particularly when watching it rapid the table around. Its not. I have no bearing against which to compare it to but now that its running is seems much more capable.

Delw, I agree about the 150ipm back and forth moves seeming more like small crashes than fluid motions. We can only wonder how bad this is for ball screw and thrust bearing life. It seems like I'll be trying to keep my feeds under 100ipm when such motions are unavoidable.

On my version of Onecnc, advantage I believe surfacing is done by following a series of points so there is no reasonable way to keep the program from stoping and starting abruptly at the end of its reciprocating motions. Mill professional and expert may have the capability to sufficiently improve surfacing toolpaths so that faster feed rates can be employed without jarring the machine. Thanks guys.

[SBC Cycle]

There are some special codes that can control the ramp (Accel/Decel) on the fly if the angular difference between the current move and the next vector exceeds a certain value but I've never been able to get the hang of how it works. Maybe it's because my CAD spits out too many short moves to allow it to work.

I would love to hear from a guy/gal that has this figured out.

BTW, I have a 2 4020s with Box ways and a 5020 with linear ways. The linear way machine sounds just gawd awful when I try to do surfacing on it. The box way machines are also clunky but nothing near the 5020. I guess their is a reason they quit making them.

[ThatguyDave]

Funny, I thought that light surfacing jobs were where linear machines were supposed to excel. I'll look into those functions when I get some time.