New Machine Build- Fixed gantry router

[KTP]

I started on my new machine yesterday, a fixed gantry router. I made pretty good progress and hopefully will be cutting parts with it by the middle of this week. As you can see in the pictures, it is comprised of a NSK robot module for the X, and a similar one will be bolted onto the fixed gantry for the Y. A smaller module will work as the Z. The clearance between the gantry supports is 24.5 inches and the clearance under the gantry with the mill surface bolted on is 6+ inches. The gantry is further supported by two 18 inch angle struts, making the whole thing quite rigid. The 20" x 20" x .5" mic6 cast AL plate mill table should be rigid enough for light cuts over the whole surface, and actually should be capable of supporting heavy cuts in the center 10" x 10" area. Sort of a dual purpose machine here I guess..if you want to mill something out of aluminum on a vice, just plop it down in the center of the table...at least that is the idea.

Speeds are going to be insane. Rapids should be in excess of 1000 IPM on all three axis.

Here are the first pics, more to follow after todays work is done:

[ger21]

Can you actually get to 1000ipm and back to 0 in 18"? Looks good.

[KTP]

"Can you actually get to 1000ipm and back to 0 in 18"?"

No, probably not I'll test it later today when I have the X axis table surface bolted to the robot module. The actual module itself can do 1 meter per second, which is almost 2400 IPM. The table surface weighs about 20 pounds, so that is probably going to reduce the acceleration by quite some bit. Maybe something like 500 IPM is more realistic..we shall see!

A video of X and Y moving as fast as possible would be interesting.

[ger21]

Originally Posted by KTP A video of X and Y moving as fast as possible would be interesting.

And mandatory of course.

[KTP]

A little bit of progress yesterday. I made 2 brackets for the NSK robot module that is the gantry axis of the machine and mounted it to the fixed gantry frame. It is VERY rigid, but I am probably going to add one more bracket near each end to fasten it to the gantry upright supports (you can see similar brackets on each end of the moving table slide). I ran out of 1/2" alluminum stock though, so will have to do those a bit later.

I did a test of the Y axis and indeed it can move at over 2000 IPM. The acceleration is fast enough that it gets up to speed in a very short distance (1 or 2 inches?) Obviously there is not the mass of the z axis yet, but at any rate, it is going to impressive (I hope).

Currently working on the z axis mounting plate. Hope to get that done and mounted today.

[rwaudio]

That's a nice machine, the t-slot extrusion is great to work with. I like the fixed gantry design, I'm thinking of making a pcb mill of a similar style when my first cnc machine is done. It just needs to be very rigid and accurate, with maybe an 8x8 to 12x12 cutting area. Where did you pickup those NSK robot modules? Those are pretty pricey devices, but solves alot of design work with linear bearings and screws. Can't wait to see more!

[KTP]

Ok, so I am going to do a bit of calculation here on what a realistic IPM might be for the gantry axis.

The 300 watt brushless servo and 90% efficient ballscrew produce 300 Newtons of force at the slide carriage.

Let us take a guess that the mass of the z axis will be 15kg or less.

F=MA, or 300=15 x A

So the max acceleration of the gantry axis would be 20 meters/sec^2

So we want to know how long it will take for the carriage to reach 1000 IPM from 0 IPM.

Vfinal = Vinital + A* t (where t = time)

so plugging in 0.423 m/sec for Vfinal, we find that t = 0.0212 seconds.

So a 15 kg mass (about 33lbs) will reach 1000 IPM in 0.0212 seconds.

Now how far has the carriage travelled in 0.0212 seconds?

Assuming that the initial velocity and initial position are zero, we can use:

X = 0.5 A * t^2

plugging in our values for A and t, and we find that X = 0.0045 meters

So the carriage has travelled 4.5mm or about 0.177 inches to reach 1000 IPM velocity.

Really? I would have guessed at least an inch or more...anyone want to recheck my figures?

[Gir]

No, that looks correct. Most people underestimate how fast the machine can accelerate. Obviously assuming 300N of force is the very upper limit, which may or may not be reachable. He most likely had his acceleration set much lower.

[KTP]

Yes, two thousand inches per minute (at the end of the video).

The z axis is finished, and weighs 28 lbs without the brushless servo motor that will be driving the Taig spindle

[Regnar]

Now that is cool!

[KTP]

rwaudio: sorry I missed your question. I purchased the NSK (and adept, which seem to be exactly the same) robot modules on ebay over the past couple of years. Once in awhile a good deal on them will pop up, but you have to be patient.


Here are some pictures of the z axis, which is almost finished except for adding in the brushless 500 watt servo to drive the Taig spindle. It currently weighs 28 lbs without that motor, so the total will be around 33lbs..pretty dang close to my guess of 15kg. I was not happy with the roll, pitch, yaw moment specs (10 Nm) of the z axis adept robot slide (sz series) so I added a beefy linear rail and carriage to significantly increase the moment capacity. The thing is rock solid, no flex.

[KTP]

I just set up a dial indicator on the gantry axis and measured the backlash.

My dial indicator only has 0.0005" graduations, and I couldn't tell if it was off by any amount, certainly less than half a graduation.

I think we can call less than 0.00025" backlash good enough for wood routing

As another note, when I pressed on the side of the z axis with about 10 pounds of force by hand (estimated), the dial indicator moved just a little less than one graduation (0.0005"). Of course some of this was probably the servo motor not being tuned perfectly, but also it would show that there is very little flex in the z axis assembly.

Quite happy so far.