Leveling the gantry

[miguelito]

Just wanna make sure, we’re on the same page here:
With the indicator in the collet, you will only check the parallel of your table to the “X” and” Y”
movements. I would say, if from one end to another you have the difference +/- .015”, you’re in
pretty good shape. If it’s more then that, the easiest way to fix it - to skim the sacrificial board
with the biggest tool bit you got and level the steps it made with the sand paper by hand.
AFTER YOU DONE THAT, you can check the perpendicularity of the head by offsetting your indicator from the center of the spindle with the “L” shape arm (the farther the better), touch the table (which is already paralleled and leveled), rotate the arm, and check the reading of the indicator in the extreme positions. But If the setting of the head is less then perfect, I wouldn’t lose to much sleep over it:
Perfect is an enemy of the good.
Unless you’re carving very deep 3D relief or making a perfectly squared, very tall cube, it wont give you much of a headache. Unlike the industrial CNC machining, the tolerances we’re dealing here are very liberal, so we can overlook some small imperfections.

[SimRacn]

Here is an example of what am making with the shark, starting out with one inch plex. So I do need a bit of accuracy and straightness. I understand what your saying about skimming the spoilboard. Already did that, and even with the steps it still is flat as its all on the same plain.
My concern is when raising and lowering the Z axis, i get .015 to .020 runout on a 2" drill rod chucked up in the router, this is what am trying to get a bit closer to perpendiicular

If i have the indicator on the collet and extended out about 2", when i sweep the table, i get over .040 in the x and y axis due to the mounting plate not being square with the table.

[saramos]

When I bought my Colt and brought it home, the Bosch had a very noticable runnout. I returned it to Rockler and they gave me a replacement that was much better. Then I bought a 1/4" and a 1/8" collet from precisebits.com. Though they give a warning about a problem with some Bosch Colts, with the one that I have, the results were better than the factory supplied collet. Runnout now is about .001.

Scott

[jeb2cav]

A really good post on the router aspect of squaring things up is - http://www.cncsharktalk.com/viewtopic.php?f=9&t=785

[SimRacn]

Thank you folks. I did do the truing up of the colt taper, and it is better as far as the router runout. I also shimmed up the clamp so it is better. Now on my 2" drill rod i have about .005 on the Y axis and about .010 on the X axis. And about .0015 runout when revolving the drill rod at the top and about .006 at the bottom. I am also going to get the collet set from precise bits and cutters, so should be good enough for what am doing.

[chiefdan1]

Mine had same problem. I used a dial indicator mounted on an adapter that goes into the collet. Found my bed out about .030. I use automotive feeler gauges as shims under work piece, works great. I use dial indicator on all my projects to zero the project to the cutter.