Real Life Trouble with Long Acme Screws?

[adt2]

I recently completed my first CNC build, a large machine with a 24x90 cutting capacity. (Okay, "completed" might be a stretch; but it's completed enough to actually be used for small projects in its current state.)

Wouldn't you know, the first paying request I get to produce a project for somebody else requires a bigger damn machine (specifically, a wider machine; I need to be able to load a full sheet of plywood and cut from edge to edge).

I'd like to leverage as much of the existing machine as possible while "widening" the gantry to accommodate the new material requirements (i.e. go from a 24" Y-axis to a 48" Y-axis). Ideally, I'll only have to replace the 8020 extrusion with a longer piece, as well as the metal plate on which the bearings ride and the Acme screw which drives the axis.

So my question is, what are the real consequences with using an Acme screw that long? To get a 48" cutting width, I figure I'll need about a 60" screw. I don't cut at real high speeds (haven't done anything over 100 ipm so far), and nearly all of my work is in MDF and/or plywood.

One other question for anybody who has some engineering background: Is the 3030 extrusion (with the quarter-inch plate bolted to the front face every few inches) plenty stiff for the 5-foot span? Or should I look at the 3060 extrusion? I'm hoping the 3030 is sufficient; the price difference between the two is astounding (like more than double). If I can stick to 3030, I can do the whole upgrade for less than $200.

As usual, thanks in advance for you input.

[ger21]

So my question is, what are the real consequences with using an Acme screw that long? To get a 48" cutting width, I figure I'll need about a 60" screw.

The 60" screws on my machine whip badly, and are limited to about 600rpm by the whipping.

[Walky]

Download the Tech Toolkit from 8020's site, it's a free software which can calculate deflection and other things.

The main problem with longer screws is whipping, it can be reduced with increased diameter (less flex, but more weight) and lead (less rpm for a given travel). I can't say which ACME screw would be right in your case, since I have almost no experience with them.

[adt2]

I have no idea what the RPM on my Y-axis is, but I doubt it's 600. A 1/2-10 5-start screw takes 2 revolutions to produce one inch of travel, right? If that's correct (and I am doing it from memory on my way out the door at 5pm, so it might not be) then 600 RPMs would equal 300 inches of travel per minute - which is about 3x what my machine works at.

If my off-the-cuff calculations are correct, is screw whip still a problem in my situation?

[adt2]

Also: is screw whip based on overall length of the screw, or the longest unsupported span? Screw length will be about 60 inches, but span will only be about 45 inches max.

[ger21]

A combination of factors, including how it's mounted. Also, if it has any bends in it, it will start shipping sooner.

If I had a 4x8 machine, I'd want it to be able to move a lot faster than 300ipm.

[louieatienza]

My longest screws are 44" from bearing to bearing, and I have the screw under light tension. I'm able to spin them at 570rpm, then the motor starts to stall i guess to to friction/gantry weight. I can spin them at 720ipm on my Y axis, but it's only 34" bearing to bearing, also light tension. I have 8-start leadscrews so my rpm ends up being my ipm, but last time I checked the AB nuts for them are unavailable at dumpstercnc...

You can also have two nuts, at the extreme ends of your carriage, one AB and one plain, that would help a little with whipping. Also getting the screws straightened and alignment as perfect as possible. My bearing blocks on my old machine were off parallel by .5 degrees, but this caused the ACME screw to bow signficantly!

I think Nook's website has a calculator for both ACME and ballscrews, but it would be a best-case scenario.

You might be able to do the long axis inexpensively with timing belt...

[Frederic Gagnon]

My machine has a 24x48" bed and I'm using 1/2"-8 (1 start) acme lead screws. My dual X axises are around 55'' long, and they can be easily flexed. With this set-up, I can hardly go over 90 ipm, this means they're spinning at 720rpm, without too much wobbling. The trick was to be able to adjust the lead screw in 3 points: the stepper end, the screw end position, and the leadscrew nut position, so they are perfectly colinear. Maybe too obvious, but that did it. I've also read some people dampen the vibration with foam pads or other soft material touching the leadscrew, but that remains to be seen!

[TrickyCNC]

Also: is screw whip based on overall length of the screw, or the longest unsupported span? Screw length will be about 60 inches, but span will only be about 45 inches max.

you cant have a 48" cut , with 45" of screw span .

[louieatienza]

I recently completed my first CNC build, a large machine with a 24x90 cutting capacity. (Okay, "completed" might be a stretch; but it's completed enough to actually be used for small projects in its current state.)

Wouldn't you know, the first paying request I get to produce a project for somebody else requires a bigger damn machine (specifically, a wider machine; I need to be able to load a full sheet of plywood and cut from edge to edge).

I'd like to leverage as much of the existing machine as possible while "widening" the gantry to accommodate the new material requirements (i.e. go from a 24" Y-axis to a 48" Y-axis). Ideally, I'll only have to replace the 8020 extrusion with a longer piece, as well as the metal plate on which the bearings ride and the Acme screw which drives the axis.

So my question is, what are the real consequences with using an Acme screw that long? To get a 48" cutting width, I figure I'll need about a 60" screw. I don't cut at real high speeds (haven't done anything over 100 ipm so far), and nearly all of my work is in MDF and/or plywood.

One other question for anybody who has some engineering background: Is the 3030 extrusion (with the quarter-inch plate bolted to the front face every few inches) plenty stiff for the 5-foot span? Or should I look at the 3060 extrusion? I'm hoping the 3030 is sufficient; the price difference between the two is astounding (like more than double). If I can stick to 3030, I can do the whole upgrade for less than $200.

As usual, thanks in advance for you input.

i'd get multi-start ACME, it really doesn't cost much more than single-start. And you can spin them slower and still get good rapids. Though if you plan your cuts well, you can reduce the "on air" time.

The extrusion prices are usually proportional to the amount of material. That said, 3030, barely 1-1/8" over 4-1/2' doesn't seem like a lot. If the 3030 is less than half price of 3060 it might make sense to buy 2.

[adt2]

you cant have a 48" cut , with 45" of screw span .

Duh. I subtracted there, instead of adding. What that should've said was, "the longest unsupported span will be about 52 inches." If I were to completely remove the router carriage assembly (i.e. no lead nut threaded onto the screw) then the unsupported span would be about 56 inches, I think, from bearing block to bearing block. The spindle carriage is ~9 inches wide, with the lead nut in the center, so with it installed the maximum unsupported span can't be more than 56 minus 4.5, or 51.5 inches.

Of course this design is all still in my head, so these numbers are subject to change.

[adt2]

i'd get multi-start ACME, it really doesn't cost much more than single-start. And you can spin them slower and still get good rapids. Though if you plan your cuts well, you can reduce the "on air" time.

I'm using 5-start Acme screws on both the Y- and Z-axes. I'm not terribly concerned with transit speed; right now, the machine gets infrequent use, and it's not like I'm standing there tapping my toe and looking at my watch. If it takes 15 seconds to move from the extreme positive end of the X-axis to the negative end, so be it. If I was doing this all day long for a living, I'd feel differently.

The extrusion prices are usually proportional to the amount of material. That said, 3030, barely 1-1/8" over 4-1/2' doesn't seem like a lot. If the 3030 is less than half price of 3060 it might make sense to buy 2.

I downloaded the Tech Tool from 80/20 and entered my information. Max deflection was less than 0.003", and that doesn't take into account the fact that a quarter-inch-thick steel plate is going to be bolted to the front face of the 3030 extrusion providing extra stiffness. I can live with 0.003" or less.