[mcArch]

This is actually an upgrade . . . or actually a new table that will transfer the computer and electronics from the version 01.01.
X: 122"
Y: 49.5"
Z: 22.5"

A/C head: 380deg. ea - in the design but may have to be added a little later.

Porter Cable 1.5hp router
ESAB PCM1125 plasma
CandCNC MP1000C torch height controller & control electronics
Steppers from Probotix & HomeshopCNC
Gecko 203V drivers.
Rack & pinion on X & Y
Ball Screw from HomeshopCNC on Z.
3/8IDx1.125"x.375" bearings for slides on X,Y,Z.

Website log:
http://www.calvinodesign.com/90037/90037-01.02/

[mcArch]

Table Base & gantry initial assembly.
W6x16's used for rails.
Gantry is trussed made of 0.75"sch40 pipe, 1/2" round bar, 3/8" round bar diagonals, 1.25"sch40 for the main horizontals which have a 3/4x3/4x1/8 angle welded on for the track for the Y carriage.

Horizontal truss of gantry is bolted to the verticals to allow for fine adjustment, leveling, and for removal if required to move the machine.

Bearings are inexpensive route, 3/8idx1.125x3/8 bolted to plasma cut brackets & they ride on the bottom flange of the W6 (dust sweep should really be in order here, I figured the top flange should protect it somewhat, at least from direct contamination & then with a sweep & cover attached, it should do ok.)

The gantry is pretty stiff, deflecting about .029 under a direct 50# load on the spindle tip. It's a little more than I wanted, but in order to get it down much more, the gantry would have to get to heavy to drive it with the inexpensive steppers . . . I'm trying not to blow the budget here.

[mcArch]

A shot of the structural analysis model.

More gantry bearing plate detail photos. The adjustable plates will have internal tooth lock washers to keep them in place once they're tightened up to the best pre-load.

Got the Y carriage mounted on temporarily. The gear reduction/motor bracket is mounted initially here as well. The belt reduction is 3:1 (10 at motor, 30 at pinion) using a 20pitch, 20deg pressure angle, 20tooth pinion gear.

[tmbg]

just out of curiosity, why do you need so much Z travel?

I'm curious because I'm trying to work out how much Z travel to give my machine. It's going to be mostly a plasma cutter, so I don't think I need too much for that, but I'd like to be able to do some router duty as well. Even there I'm not seeing more than a couple of inches of travel as necessary. What are you gonna cut on that machine that can take advantage of 22.5" of travel?

[lukewarm]

Looks great what are the specs on your stepper motors and how much do you think your carriage weighs?

[mcArch]

just out of curiosity, why do you need so much Z travel?

I'm curious because I'm trying to work out how much Z travel to give my machine. It's going to be mostly a plasma cutter, so I don't think I need too much for that, but I'd like to be able to do some router duty as well. Even there I'm not seeing more than a couple of inches of travel as necessary. What are you gonna cut on that machine that can take advantage of 22.5" of travel?

Sometimes I'll have to throw a W-shape on the table & burn a series of holes in the flange or web. I also have some plans for making some 3d molds from foam (EPS most likely) in one milling to either pour concrete over, or lay up fiberglass - some custom roof structural/light monitor panels (these are still in design stage, but coming soon).

If you are just going to do flat stuff, then no, you don't need that much Z travel . . . most people don't. But I have an A/C axis head design for this as well that will enable faster milling of amorphic shapes like the skylight mold and will allow me to cut multiple angles on steel shapes for things like back-beveling for welding, or for pipe fitting.

I also sometimes will put a table top (old table saw top) with some short leveling legs on top of the plasma slats & it's like 3 or 4 inches high. This then requires me to raise the Z up above the leveling table plus whatever thickness of material I'm cutting. & it leaves open the possibility to cut/mill assemblies of material (like the top of a box for example) if needed.

To quote Edna Mode from "The Incredibles" . . . When asked "What on earth do you think he's going to be doing with this suit?" . . . she replied: "Well, I'm sure I don't know . . .!" what I'll be cutting with this thing, that's why I made it with as much Z travel as my roof would allow for now.

But, you're right, most will not need any more than like 6" or so.

Cheers!
-Mike

[mcArch]

Looks great what are the specs on your stepper motors and how much do you think your carriage weighs?

I've got a pair of 1280ozin.'s on the X, & they're geared 3:1 for the 20pitch, 20tooth pinion gear, that means I'm going to get 1.0472 (Pi/3) inches of travel per revolution. In theory I should be able to push the gantry with it but I suppose I'll have to see. I have a 400ozin on the Y and the Z. The Y is geared with same rack & pinion as the X, & the Z uses a 5turns/inch, 5/8"dia. ball screw. The whole gantry with Y & Z together, by initial calcs weighs somewhere between 200 - 250#. The Z carriage without the router or plasma torch weighs about 15, maybe 20# . . . the small PC router is 8#. The tricky thing is when I put the A/C head on it, that's going to add like 25 or 30# right there & I may need to go to a higher gearing ratio, or bigger steppers all around, or both. It's a work in progress!

I'm really no expert with this & I've just tried to do as much research as I can, calculate things when I can, & try it . . . this gantry is a little lighter than my current/first one and the bearing system is way smoother, so I'm hoping it will work much better . . . the original has worked fine for me for more than a year, it's just not as accurate as I want & it uses lead screws with brass nuts for the drive system on X & Y & this is really bad news especially with plasma dust because the brass just wears away & you get some pretty bad backlash pretty quickly. Rack & pinion is the way to go.

Cheers!

[tmbg]

you may find the 1280's to be monstrously overkill. I have a pair of 205oz-in steppers driving my gantry at the same gearing (3:1 driving a 1" pinion for pi/3 inches per rev of travel) and I was getting 600ipm rapids with 200in/s^2 accel without breaking a sweat on 24v power. I just got my G540 and 48v supply in last night, and I think that I can run it even faster.

Now granted, my gantry is probably lighter than yours, but 1280oz seems like entirely too big a motor. Remember that you trade speed for torque, so you may well end up with much slower travel than you'd want!

[mcArch]

Hmm, thanks Ian, that's sounds promising, I have those 1280's running my current machine at 24V (they're limited to about 1190ozin by the gecko's only allowing 7 amps wide open), with the lead screws, it travels at 2 revs/inch and I get only about 200ipm on the X & the G203V's show the orange light - meaning they are operating at full capacity if I understand it right.

This old machine doesn't have the greatest slide bearing system, so the new one should be much better. But I'm glad to hear that your thoughts are that they will be plenty of power. I hope you're right. I'm thinking I'll get about 400ipm with them, that would be plenty for me on X & Y. Thanks!
-Mike

[mcArch]

Ian, what are you running for your linear slides?

[tmbg]

You'd be amazed how much force these motors can make through the type of gearing we're using. If you go too big, you just end up slow. I don't have torque curves for your motor, but here's a keling torque curve:

1200oz-in nema34:
http://www.kelinginc.net/KL34H2120-42-8AT.pdf

So, at the 4000 pulse per second data point, we're turning 600rpm (4000 pulses per sec / 2 half-steps per pulse / 200 steps per rev * 60 sec per min = 600rev/min) and making 2Nm of torque. You have two motors, so that's 4Nm torque acting on the gantry. 3:1 belt reduction makes it 200rpm and 12Nm torque. That torque turns the pinion, acting on a lever arm equal to half the pitch diameter, which is 1". So your arm is 1/2", 12.7mm. 12000Nmm / 12.7mm = 945N force. F=ma, and a Newton is a kg accelerated at one meter per second per second, or N = kgm/s^2. Gravity is 9.8m/s^2, so 945N can accelerate 96.4kg at 1G, which is equal to 384 inches per second per second accel.

Now, just plucking another motor off the keling site:
Nema23 185oz-in. MUCH smaller than you're running
http://www.kelinginc.net/KL23H256-21-8BT.pdf

at the same 4000pps half-step, we'd be turning the same speed. The little motors make 0.6Nm each for 1.2Nm total, through 3:1 gearing gives us 3.6Nm. 3600Nmm/12.7mm = 283N force, or 29kg at 1G. So we can run a 64lb gantry at the same snappy 1G accel.


Sorry, I got off on a tangent, but the point is if your motors are too big, you have torque that's not actually doing anything useful for you, since there's only so fast you need to accelerate the gantry. However, you'll severely limit your top speed with the bigger motors. For plasma, you are gonna want a max cut speed of around 400ipm.

Do you know how much your gantry will weigh?

[tmbg]

Ian, what are you running for your linear slides?

THK SSR25 rail for Y, THK HSR12 rail for X (along gantry), and HiWin MGR9 rail for Z and floating head.

[mcArch]

My gantry right now is at 210#, without the router, the torch, cables, the X motors, the B/C swivel head w/ 2 more motors later, etc, I figure all told you can add about 70# or so for all of that, so I'm closer to like 280#. I think that's pretty heavy don't you think?
-Mike

[mcArch]

So are you saying that we can just take
1190ozin x 3 (3:1gear) =3570ozin
then:
3570ozin*1/0.5in(the rad. of the pinion)=7140oz
then:
7140oz / 16oz/lb = 446.25# of force for each motor?
So that's theoretically 892# that these motors (at 1190ozin) could accelerate at 1G?

If that's the case, then my current machine has way way too much friction in the bearing system & those lead screws, because those 1280's have missed steps before . . . maybe something got in the bearing path or something, but they don't seem to be able to move it that well, & 400rpm is all I get out of them, but I read that's about all I can expect. They're from HomeshopCNC.
-Mike

[tmbg]

one thing that may be worth trying is to get a spring scale, hook it to the center of the gantry, and see how much breakaway force it takes to start the gantry moving

[Feckless]

Some of the inaccuracy may be due to the amount of z travel as it will multiply a small amount of backlash on the rails to something unacceptable, just like a large lever, does that make sense or have I lost it. I think you may be diagnosing the result rather than the problem.
Nice build by the way and great drawings Mike.
Gordon

[mcArch]

It's not really a question of inaccuracy (yet) it's more of a question of the theoretical calculation of the torque required to move the gantry. Or rather the size of the motors required. From what I understand, the critical value with sizing motors is the rotor inertia. The ratio of the required to the actual should be about 6:1 max. I use the Parker Automation Sizing Software & come up with a reflected rotor inertia for each of the X axis motors of 0.014487764 in-lb-s^2. And the 1290's (they're 1290's as opposed to 1280's) have a rotor inertia of 0.002654867 in-lb-s^2. This gives a 5.46:1 ratio which is not far off. This is all academic of course and there is no substitute for tests with the actual motors. I plan on having the motors on this week and can post some results of how they work under router(on wood, maybe some alum & steel too if I have time) conditions.

As far as the deflection of the Z carriage, I understand that it is a vertical cantilever susceptible to lateral milling forces, but by my calculations and structural analysis models; with a 50# lateral force at the tool tip (say a 1/2" endmill with a pretty aggressive pass on maybe mild steel) the machine would see a cumulative deflection at the tool tip of 0.019" - that's only 19 thousandths of an inch. A little excessive maybe, but for a finish pass, one would not make as deep a pass and the milling forces would be greatly reduced. For example, for the same lateral force but at 15# instead of 50# (shallower pass and slower speed, more akin to a finish pass for milling something), the machine would see a cumulative deflection at the tool tip of 0.006". Not too shabby I'd think.

Of course this is all academic as I have said but I hope to have the machine running this week and can post some real test results.

Cheers!

[mcArch]

Some more recent photos. The rack & pinion are from McMaster.com: 20 pitch, 20deg pressure angle, 20 tooth pinion . . . that's 20 teeth in a 1"pitch dia. so it's PI inches per revolution.

The racks are screwed to the web of the beams with #10 machine screws at 4" o.c.

The trussed gantry has proven to be very lightweight and very stiff so far. I can push the gantry from one side with my hand and it barely twists.

[mcArch]

Ok, so all up, the table will run 600ipm in both X & Y (lost a few steps once on the Y at 600)

http://www.youtube.com/calvinoarchit.../2/njuhavmt7g8

The 1290's don't have any trouble, I figure they're pushing about 240# or so at the moment & seem to have plenty left in them to push the A/C swivel head when I put it on. The little nema23 400ozin is what lost the steps on the Y, but maybe there was something on the bearing track . . . I still think it should be at least a 600ozin, if not a 740ozin.

I have not checked the signal light (orange) on the Gecko203V's to see if they're tuned/matched right yet. But will check that out.

I still have to check them under some roughing pass router loads and that will tell me more. So far, so good. I still have to re-do the electronics enclosure, the cable carriers are on the way, and the plasma unit & table slats are coming soon. I have to get this project out the door first, then get the plasma & slats on the machine so I can get the next proj. in here.

I need to re-design the Y carriage because I think I'm getting too much resonant deflection in it, the design is really a little weak, I have another idea. I'll post some drawings & the re-build when I do it.
--Mike

[mcArch]

http://www.youtube.com/calvinoarchit.../2/uMwdwqepxSY

Plasma cutting 3/16" plate. The Plasma runs pretty well, only thing is that I need a little bigger motor on my Z becuase it loses steps once in awhile & I can only get about 60ipm out of it before it really loses steps. I have to redo the ball screw mounts also & that may help.