Cenerline Gear Rack?? (First time poster)

[duce00]

Greetings from newbie-ville,
I looked around for a posting on this issue and didnt see it addressed directly. I apoligize in advance if Im beating a dead horse here because I cant believe Im the first on this topic.
The question is: How effective is (or has anybody ever tried) mounting a gear track to the centerline of a CNC router gantry's long (X?) axis. Ive read about people mounting gear racks on both sides and running a shaft accross (dubious results aparently), and also using 2 steppers ganged at the controller to avoid the shaft idea (very costly solution). I havent seen any posts about using a gear rack to basically replace a ball screw.
The obvious goal is reduced build cost. The obvious effect is reduced accuracy due to backlash. Ive read about gear mesh and spring tensioner solutions to decrease backlash and it seems very workable. Ive seen posts quoting .005" results for rack backlash and I could easily live with that. Especially considering the cost of a 48" ballscrew vs. 48" of rack on McMaster Carr.
Has anybody done this? Ive got a basic layout but Id really like to see other plans if they are out there.
Thanks!
Duce

[duce00]

Well I think I answered my own question. The simplicity and cost of a lead screw with anti-backlash device seems to solve the problem better than a gear rack with pinion tensioner.

I figured I was missing somthing because ball screw costs were far too high for entry level machines.

Did I miss somthing on the linear motion basics FAQ?? you know the part at the top that says "read this first". I dont recall seeing this issue raised and Im wondering why. It seems that it could be explained in more detail considering the quantity of machines being built like this.

duce00

[LeeWay]

A cnc machine is very broad. Mills, routers, engravers, plasma cutters etc. all use different design considerations and materials.
It all depends on what you are making and what the expected results might be.
There are some basics of course. I did build my router using a single ball screw for the X. I was lucky in that it is a small machine and I used over sized steel to make it. I was able to tweak the gantry to load the sides so there would be no racking of the gantry. I pay for that though in cut speed.
A bigger machine will certainly benefit from driving both sides. My second machine is a plasma and I am driving both sides with a single motor. A shaft runs accross the gantry. It would cost more, but be a better machine to drive both sides and use two motors to do it.
This is a plasma though and the gantry at nearly 6' across may only be half as heavy as the little 30" gantry on my router.
The info you seek may not jump out in a search, but may be buried in comments and experiences just like I a posting. It may be that yours isn't a straight forward answer, but you have to read between the lines and there are a lot of lines at the Zone.
Read up on some of the project logs. Guys usually have comments and reasoning behind the designs they ultimately wind up with. They may have started with some different than the result.
It takes a while to read through them but the info I gathered was very helpful for me when I started down this slope.

[duce00]

Thanks for the input Leeway,

I think you pinned where Im at right now pretty well. Your sooo right about digging deeper into this site. I have learned quite a bit just today. The posts and designs of the other members have many of the answers but there are soooo many posts that key word searches became far more benificial than just wandering through forums. Today was kinda a breakthrough though. Im totally stung by the CNC bug and its great except for needing to get my schoolwork done (doh!).

If anybody has any pointers on my basic ideas please chime in. Im all ears

My goal is to build a CNC router that has:

34" min. X travel
20" min. Y travel
6" min. Z travel

Accuracy of +or- .005

Capable of routing/machining maple and mahogany guitar parts (multiple passes ok)

---------------------------------------------------------------
Design concepts as of now are:

-Use steel rail bearing for X and Y axis similar to Mechmate
(considering modifications for greater stability and drawing up a few new designs that improve basic idea)
PLEASE!! ANY INPUT ON BETTER SOURCES OR OTHER INFO

-Ball screws for all axis
5/8" for X axis to avoid whip/flex
Mcmaster Carr #5966K262
PLEASE!! ANY INPUT ON BETTER SOURCES OR OTHER INFO

-Part steel part wood (mdf or melamine) construction
(steel can be fabricated and welded at school, wood can be done at home, cost vs benifit and hassle will determine which is best)

-Probotix kit or similar for steppers, driver power etc
http://www.probotix.com/multi_axis_s...axis_probostep
(Im sure there are more cost effective choices but it fills the spec sheet for now.)
PLEASE!! ANY INPUT ON BETTER SOURCES OR OTHER INFO

--------------------------------------------------------------------

I am a proficient Acad user and midlevel Solidworks user. The resources are there to get quite elaborate but the design must be VERY cost effective so maintaining the origional goal is the bottem line not re-inventing the wheel (as I sometimes find myself doing

I am trying design out the use of expensive linear bearings and other factory components. I have a full machine shop for use and a welding shop too (although its been a long time since I welded).

My timeline is to start construction over summer break. I will have a new garage (woooohooo) and more time. Design is starting now and Im looking for some good input.

Thanks

duce00

[Zumba]

Ballscrew setups are not nearly as trivial as you've made them out to be. A double anti-backlash nut is required, because single nut will eat up your entire tolerance window by itself. Expect to spend $120-250 for the nut, depending on size.

You also need proper thrust bearings to support the ballscrews (duplex angular contact bearings in a well made housing). If you don't have access to a machine shop, expect to spend nearly a grand for them.

If you drop your accuracy requirement to about +/- .020, I think this project will be much more feasible with your budget constraints. Most commercial routers under $40,000 can't even hit +/- .005. If they say they can, they probably have "accuracy" confused with "repeatability", and are testing in ideal conditions.. that is, the moving at a very slow speed, low-G accelerations, and cutting air with the spindle off.