Re: [CAD_CAM_EDM_DRO] Re: Climb Milling (again)

I've been looking at some of these responses for "software compensation for
backlash".
Real HOGWASH IMO. If the machine "jumps" due to excessive load under
backlash, ho the H is the software going to know how much "jump" to
compensate for? Furthermore, how's the software going to know how much
backlash you have in your machine...or how much extra backlash to compenste
for due to wear or poor machine adjustment?

Smoke

>Ron,List,
>
>Your snipped quote below brings up a good point. Most software
>backlash comp is done at the "mathematically correct" point.
>
>I have turned off b/l comp in most pkg's I have used, but I got to
>looking HARD at the problem when writing my own code.
>
>I came to the conclusion that one of the problems is that the b/l
>comp was being applied at this "mathematical" point, when in fact,
>the axis would not ACTUALLY change direction for at least a few more
>steps. This often happens in arcs, since the finite set of points
>"available" to the system do not often coincide with
>the "mathematically correct" set of points for the curve.
>This results in the backlash comp INTRODUCING a "bump" since it
>applies the correction before the axis changes direction.
>
>NOTE:If the cutter forces are strong enough, when compared to the
>forces required to move the table; then the cutting forces WILL "take
>advantage" of any backlash available in the machine. However, it is
>good to remember that cutting is a dynamic action with many forces
>and resistances at work, including mass, stiction, inertia and
>momentum. In more than a few cases, these factors will(or can be made
>to) work in FAVOR of the desired toolpath, even where backlash is
>present on the mill.
>
>Back to my point: One easy way to see this effect is to draw circles
>arbitrarily on graph paper, then use the vertices of the graph paper
>as "achievable" points in the system, and draw the "best"
>approximation of the circle (the one first drawn) that you can.
>
>Not only will this be an eye-opener for many OUTSIDE of the b/l comp
>implications, it will show that the mathematically correct point
>(where the s/w typically inserts b/l comp) is NOT the same place as
>the change of axis direction!
>
>One way to "fix" this is to set a flag (in the s/w code) when a
>quadrant boundary is crossed, but NOT to APPLY the b/l comp UNTIL the
>dir bit changes... I can state that this DOES do a better job on
>arcs; compared to simply adding steps in at the quadrant boundary.
>
>Admittedly, there are still going to be problems with any "loose"
>machine. But we need to look at how we can BEST compensate for such
>limitations (including using the s/w), rather than just calling the
>whole thing "bogus", IMO. And as I mentioned previously, not all
>machines controlled by these s/w pkgs. are "milling", and/or working
>against(relatively) high cutter loads...
>
>Hope this helps.
>
>Ballendo
>
>>--- In CAD_CAM_EDM_DRO@y..., ron ginger <ronginger@r...> wrote:
>>After all the good explanations and horror stories of the results of
>>climb miling and backlash let's hear from the guys that think
>>software compensation for backlash is a good, or even useable thing!
>
>>Sure, the software can cause the motor to backup to the right
>>mathematical spot<snip>
>
>>A great bit of proof why Ive always maintained software compensaton
>>for backlash is bogus.
>
>
>
>Welcome to CAD_CAM_EDM_DRO@...,an unmoderated list for the

discussion of shop built systems, for CAD, CAM, EDM, and DRO.

>
>Addresses:
>FAQ: http://www.ktmarketing.com/faq.html
>FILES: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO/files/
>
>Post messages: CAD_CAM_EDM_DRO@yahoogroups.com
>Subscribe: CAD_CAM_EDM_DRO-subscribe@yahoogroups.com
>Unsubscribe: CAD_CAM_EDM_DRO-unsubscribe@yahoogroups.com
>List owner: CAD_CAM_EDM_DRO-owner@yahoogroups.com, wanliker@...
>Moderator: jmelson@... timg@... [Moderator]
>URL to this page: http://groups.yahoo.com/group/CAD_CAM_EDM_DRO
>bill,
>List Manager
>
>FAQ: http://www.ktmarketing.com/faq.html
>
>Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>