I just started testing my build, and after verifying squareness with a pen, I created a simple gcode program to flatten the work area.
The router starts close to the origin, carves the pocket, and finally performs a contour pass - so far so good.
The movement sequence does the main cutting x-wise, while y-wise the movements are smaller steps.
The problem is that cutting depth is inconsistent depending if the router is going left to right, or right to left, so that in the end the spoilboard is not flat at all, but instead a succession of low passes and high passes.
Cutting depth seems to be consistent if direction does not change.
I can add as many information as needed, but I’ve got the feeling that I am missing something very basic here.
In red the x axis, in blue the y axis - green is the tool path. Please note that the areas where the tool actually moves in the y direction are outside of the picture - I connected the green segments for clarity.
As you can see, there is a quite prominent step. The high part is less wide than the low part because I have some amount of overpass.
There are two things (at least) that can contribute to this type of result. One is if the axis of the router is not perpendicular to the table. This would make an effect that looks a bit like blinds.
The other contributor is if the router mount is not very stiff, then the load from cutting can cause deflection, and the angle of the router is not the same when cutting in the +X direction compared to the -X direction.
First I would suggest shimming the router so the cutter is as close to perpendicular as possible. Then if that is not enough, you could try cutting in only one direction so the load on the tool is relatively consistent. This is the motivation behind the surfacing mode of this test pattern generator. Also if the tool is not quite perpendicular, it’s advantageous to choose the cutting direction so the deeper side is the leading side of the cutter.
Good info from Jamie. You can take a length of metal lose hanger and make a Z like shape, put that in your router and manually spin it around. Start with then end/point touching and as you rotate it, it will reveal if the router is not perpendicular to the board. Using an know flat surface such a pane of glass or mirror, use a level to check how true the glass is to level. I use this method on my drill press. This will give you and indication of which direction to shim the router. Also check to see if the router is secure and not flexing in the mount. I had a similar problem and found out that my mount was a little loose, and that the mount had actually broken, allowing the router to flex in one direct but remain vertical in the other.
Good suggestions and I will add one more. Check the core bearings are touching the x and y tubes. Recently had the x bearings on my core develope a little play and and produced similar issue.
Based on what I see, I would say it is flex at least… maybe a combo of bad tramming too, but definitely flex. If it was not flex, the blinds pattern would be even both climb and conventional cuts. With flex, the router will tilt away from the cut during the climb pass, and pull in to the cut on the conventional pass.
First thing I’d do is move the bit with my hand and try to observe where most of the movement comes from… work on those bearings or whatever. If everything is as mechanically good as you can get it, then 2 things, a) change your cam so it does only climb cuts, and/or b) use a smaller step down with more passes. Generally speaking, for the higher speed routers we use, you should always be climb cutting anyways (I do for all materials). Sometimes a smaller step down is also needed, if you can’t get the router to low enough RPM for a lower feed rate and reasonable chip loading.
first of all - thanks for the great and useful feedback. It’s really something to have several people jumping in willing to help a total newbie - speaks loud for the community in my opinion.
To sum it up, I might have two problems, or a combination of both:
I certainly did not tram the spindle; I will do that with the “coat hanger” hack or maybe splurge on a chinesium dial comparator
flex is different in the two directions
I am using a Makita RT0700 - very similar to the US version - and RPMs are not settable. For these test cuts I am using a 1 mm depth of cut so I am guessing I am pushing 15kRPMs - feed rate is 15 mm/s.
I believe that with the setup above I will need to climb cut - not sure if feed rate & depth of pass are ok though - any resource I can use to learn?
Regarding the two problems: I had some bearings loose before I did the first cut, I tightened some bolts and now I have no more play on the gantry. I tightened the Z axis brass nut down as well, two screws only as suggested on the assembly instruction - I still have a bit of play there, but I carefully observed the area during surfacing, and I see no movement at all, I would expect to be able to appreciate such a jump.
I will definitely tram the router, recheck for stiffness, and try to surface the waist board again. I’ll keep you guys posted on the outcomes.
Again, thanks a lot to everyone! You really made me feel welcome, and provided useful insights.
