Hello Guest it is February 27, 2021, 08:22:57 PM

Author Topic: Axis length movement profiling / mapping ?  (Read 3459 times)

0 Members and 1 Guest are viewing this topic.

Axis length movement profiling / mapping ?
« on: October 11, 2010, 07:26:06 PM »
Hi,

I am in the build phase of a DIY wood router.
- Sized for 4 x 8 sheet
- model airplane wings
- sheet plywood
- Various other hobby things.

This is for hobby use for routing wood and foam. I have an idea to also use it to improve my drilling results as it also incorporates a drill in the Z axis. (not a spindle, a hand drill, so not perfection use, just better than I am by hand)

Drive Method
- Tried to do them all the same, but in the end, they might all end up different, but most likely belt and rack and pinion.

Motors
- At this point, steppers running open loop. I have some linistepper driver kits already, but am open to options if other methods are significantly better, and I have the budget.


Question

Regardless of what method is used, there are always imperfections and non - linear areas of a drive mechanism. For example, even high end ball screws have a rating for "accuracy per foot or meter", and a belt has even larger errors, most likely, not consistent over a 8 foot distance.

I read the mach III manual again last night (yes, I am a beginner) and found a setting to "fine tune" a local section of motion for a movement distance, but I wondered if there is a method to "fine tune" sections of the axis. Example, if I can manage to borrow a very accurate meter, can I somehow profile the axis for these imperfections continuously, or over say 10 regions of the axis ?

I realize that my work may or may not need this, it is partially for my own learning.

I am sure this has been asked before, I just could not find it in the many  pages of threads or the manual.

Feel free to just point me to the right section of the manual or an existing thread.

Thanks a lot

Harry

Offline RICH

*
  • *
  •  7,419 7,419
    • View Profile
Re: Axis length movement profiling / mapping ?
« Reply #1 on: October 11, 2010, 08:43:19 PM »
Harry,
There is screw mapping under the Function Cfgs tab. Never used or tried it out so would be interesting to see if anybody else has played with it.
If you have encoders then seems rather easy. If you need to manualy put in the numbers then the task gets somewhat more difficult. Depends on how accurate you want to be in profilling the axis.
RICH
« Last Edit: October 11, 2010, 08:46:52 PM by RICH »
Re: Axis length movement profiling / mapping ?
« Reply #2 on: October 12, 2010, 01:28:22 AM »
Hi Rich,  Thank you for that information, that is exactly what I had in mind.

Since it is perhaps not such a commonly used feature, I guess that means that an experienced machinist probably has a better method of improving their positioning than just relying on the open loop programmed steps on a rack, belt or screw ?

Does that mean that high precision work is all performed with linear encoders along the length of the axis with some kind of feedback to the software ?

Thank you,

Harry

Offline Hood

*
  •  25,838 25,838
  • Carnoustie, Scotland
    • View Profile
Re: Axis length movement profiling / mapping ?
« Reply #3 on: October 12, 2010, 04:28:58 AM »
Please dont take this the wrong way but do you actually need such accuracy with a router?
Hood

Offline RICH

*
  • *
  •  7,419 7,419
    • View Profile
Re: Axis length movement profiling / mapping ?
« Reply #4 on: October 12, 2010, 06:45:54 AM »
Harry,
Your positioning will only be as accurate as the compnents that make up the axis. Suggest you have a look at some commercial units to see what they
publish on accuracy. Also see the posting on this site for what others have made. There are a lot of folks that use router tables, i don't do wood working anymore ( never did any with cnc).  A big table like yours would probably use servos since your after speed and stepper rpm max's out around 1000 rpm or so.Also look at the pro's and cons of the different components. A high precision ball screw 8 foot long will cost a small fortune but a rolled one say .003 to .006" per foot would give you say 1/16" but i think manny users would go with a rack and pinion or belts becuase of the distance.
 
There is a big difference between metal and wood machining in terms of accuracy as Hood questions.

As far a profiling an axis, well i have done a few using  equipment which most folks don't have . It's time consuming.All depends on how accurate
you want to measure.

RICH
Re: Axis length movement profiling / mapping ?
« Reply #5 on: October 12, 2010, 09:01:23 AM »
Please dont take this the wrong way but do you actually need such accuracy with a router?
Hood

Hi, you are absolutely right, that a wood router probably does not need higher precision than I could achieve with good quality components.  I am looking into this area for really several reasons:
- I am already well into a belt drive approach, and the belt needed is nearly 12 ft long.  The price of belt with "average" results is reasonable, and the price of belt for "good results" is much higher.  Belt for "great results" (better than I need) is not made.  I hope that make sense, but the point is, if I can profile the axis, I can use cheaper belt and pulleys.
- Over short distances, the belt specs are fine, but over longer distances, they are kind of marginal.  In an open loop, non profiled axis, a simple thing like drilled hole spacing could be off by 1/8 in over the length of a board, based on "specs".
- One of the goals of this "router" is to be able to drill fairly accurate hole and positions, using a drill, on 20 ft long boards.  Over 20 ft, there could be a rather large placement error, at least in theory.
- Longer term, I would ideally like to use the same ideas I have learned from this project to put cnc control on a mill, which of course has higher accuracy requirements.

Thanks for looking,

Harry

Offline BR549

*
  •  6,944 6,944
    • View Profile
Re: Axis length movement profiling / mapping ?
« Reply #6 on: October 12, 2010, 11:20:15 AM »
Remember the error is not always accumalative(;-) It can swing both ways (+/-) and end up very close.

Just a thought, (;-) TP

Offline ger21

*
  • *
  •  6,289 6,289
    • View Profile
    • The CNC Woodworker
Re: Axis length movement profiling / mapping ?
« Reply #7 on: October 12, 2010, 11:36:44 AM »
As Terry said, in reality, you should be far closer than 1/8". I'd guess better than 1/32". What you really need to consider is the belt flexing. No mapping can fix that.
Gerry

2010 Screenset
http://www.thecncwoodworker.com/2010.html

JointCAM Dovetail and Box Joint software
http://www.g-forcecnc.com/jointcam.html

Offline BR549

*
  •  6,944 6,944
    • View Profile
Re: Axis length movement profiling / mapping ?
« Reply #8 on: October 12, 2010, 01:47:57 PM »
As everyone has said. IF you need accuracy you have to BUILD it into your design. Trying to add it in later rarely works and usually is EXPEN$IVE.

If you run belt drives then make sure the width of the belt and the core material ( I prefer Kevlar) is strong enough to prevent the rubber band effect(stretching). Also use heat treated steel drive pulleys to prevent pulley wear over time. A simple 1 strand over loop design is very cost effective and works well in most applications.

Racks can be very cost effective IF you control backlash (splitdrive gears) and keep up with the wear factor of running in the same spot every time. rotate your work across the bed to spread out the wear.

(;-) TP
Re: Axis length movement profiling / mapping ?
« Reply #9 on: October 12, 2010, 02:34:00 PM »
Hi, Thanks for the replies and info.

I ran the numbers on the belt drive - surprisingly strong belt needed actually.  I was kind of surprised at the width needed compared to what people commonly use, so I am still double checking my calculations to see if they are reasonable.

I guess the alternative to the screw mapping approach is closed loop positioning from a linear encoder, but I have no idea how to approach that beast, or the budget impact.

Thanks

Harry