1.7*0.74*0.4m Mill / Router building...

[Jonathan]

There's a big thread regarding this idea on CNCZone... seems some have used it successfully and theres a commercial offering based on the idea.

Yes I spotted that. There's two threads, one quite short one about the method I'm using and one huge thread about roller screws which is another story entirely. I may well try making a roller screw if this method doesn't turn out so well.

These photos wore taken yesterday:
Pretty self explanatory - Z-rails are bolted down and really smooth.


I milled the alu plate square at school, shiny! Never seen such a good finish on aluminium - must be the grade.


Today I've milled the gantry sides flat at either end and drilled them. Also had to nip a few mm off the end of one of the Z rails. I was a bit worried with it being hardened, but it milled OK taking small cuts. Plenty of smoke...

Also started machining the M16 rod for the Y axis.

I'll post some more photos tonight.

routercnc:
I see what you mean with the bearing jumping now - that is a bit of a worry.

[Jonathan]

Some more photos ...

This is how the gantry fits together.


Detail of gantry sides. I've milled them flat etc. Unfortunately I drilled those 5 holes (for M8 bolts) in exactly the same place on both sides, forgetting that the other one is flipped over! Looks like I'm going to have to mill those 5 holes on one side to move them across by 6.5mm...dow!


Just after milling block to fix box section to gantry sides:


It fits!


Pretty accurate too I sawed the plate to roughly 80x80mm then used a 20mm 4 flute endmill to cut it to size. I used 2mm cut depth and 1000mm/min - no problem at all! I'm sure I could have gone faster or cut deeper since it didn't get hot. I then did a single finishing pass taking off 0.2mm at full 20mm depth and 300mm/min.

I'm very pleased with the accuracy seeing as this is with trapezoidal screws and using backlash compensation. I guess the brand new cutter and having set the backlash compensation values to the nearest 0.002mm a few days ago helped.


Shiny :)


Just a shame I didn't measure the box section beforehand and relied on the sellers dimensions. Turns out it's more like 74.6mm, not 74. I think it'll be ok though...

I'll post a vid of that latest bit of milling when I've compiled it.

[Jonathan]

http://www.youtube.com/watch?v=iC7lXNTPYrg

Some progress today. I made another of the above part and assembled the 50kg(!) gantry, then assembled it:

[Jonathan]

Success!

I made the mount for the angled bearing thingy on the Y-axis. I'm not using springs/belville washers at the moment since I don't have any...yet. Currently it's just a few bolts pushing the bearing onto the screw at the right angle, but it seems fine!

So about an hour ago I clamped the motor mount and just the bearing for the Y screw at the motor end onto the gantry using G-clamps and mole-grips :whistling:

Motor is half stepping, 42T pulley on motor and 28T on the 2mm pitch screw. Currently I can just get 3600mm/min rapid feed but I'm sure it'll be more if I use a smaller than 28T pulley and mount the bearing on the floating end of the screw. Still a respectable amount though I think?

Anyway....backlash: 0.006mm :exclaim:
I spend about an hour checking that figure so I'm pretty confident it's right. I was using a 0.0001" indicator and the reading averaged 0.25 thou.

I think I'll settle for that :rofl:

[Jonathan]

Got the motor and bearings fixed down properly on both ends of the Y screw now.

Been experimenting with the pulleys. With 42:13 I could get 6000mm/min, but that's above the critical speed of the screw as calculated by Irving's spreadsheet, and yes it started whipping at that. So I tried 28:13 and got 4800mm/min (maybe 5000), which seemed pretty smooth. Acceleration currently 800.
I don't think I'll use the 42:13, Irving what do you think would be best if you don't mind working it out? I've got plenty more pulleys to experiment with. The motors are the 3N.m ones from Zapp moving an 18kg Z-axis.

I know it's a balance between high rapid feed and having sufficient torque for cutting, but it's difficult to work out where this point is since there's so many variables!

I measured the backlash again, and got 0.01mm this time...

I'm not sure wether to tension the screw or not. It wouldn't be too difficult to tap a hole in the end and use that to tension it with a thrust bearing, but is it going to make much difference if I stick with less than the current critical speed?

[irving2008]

Jonathan,

I'd try the empirical approach.. decide on your required rapid speed. Setup for that. Then put a rope and pulley on the router so that movement of y lifts a weight... add some weight (stones in a bucket or something), about 4 - 5kg and see how fast it'll go without losing steps (use a DTI to measure that it returns to the same point after moves). I'd start with the 28:13 ratio as this gives a nice rapids at 4800-ish and see what you can manage with load... 3000 would be OK, 4000 would be good... then turn accel down a little and speed up... you'll soon find the sweet spot... plot a chart (speed along bottom, accel up side) and you'll get a line for a given weight... change the weight, different line. That way you can tune it for different materials as you'll know what its capabilities are.

[Jonathan]

Irving,

Thanks for those instructions - I'll do that at some point when I've got the whole router together. Thought I'd done my last physics practical :lol:

Some photos...

'Nut' I may well add springs/Belleville washers.

[Kenneth Paine]

Irving2008,

There's a big thread regarding this idea on CNCZone... seems some have used it successfully and theres a commercial offering based on the idea.

Would you be so kind so as to post the link to this thread? I am unable to find it.

Many thanks,

Kenneth

[Lee Roberts]

Kenneth,

Irving is going to be absent for the next 3 weeks as he is on a trip to the states.

Lee

[Jonathan]

Irving2008,



Would you be so kind so as to post the link to this thread? I am unable to find it.

Many thanks,

Kenneth

Here it is...

http://www.cnczone.com/forums/showth...=13593&page=26

I can't find the link to the shorter thread where the method I use is dicussed, but it's linked off that thread somewhere!

I may use the method discussed in the above link for the Z axis as I'm running out of space.