PDA

View Full Version : linear rails



darryl
12-26-2009, 09:22 PM
This is just an idea- there have been many sliding mechanisms made using round rods and separate leadscrew actuating mechanisms. Hand crank or motor driven, it's the same- a lead screw is turned which moves the table one way or the other.

What about using acme threaded rod as both the leadscrew and the rail? A typical table would have two, and they'd have to be synced of course. The nuts would be a casting or molding that goes the width of the table, thus giving a large surface area supporting the table and interacting with the acme threads to drive the table back and forth.

I don't know if this method has been used anywhere, but I'd like to see some examples of it, and discuss the benefits and drawbacks.

luthor
12-26-2009, 09:36 PM
This method is used in lots of situations,eg. Back gauges on Brake presses and Guillotines etc.

fredf
12-27-2009, 12:12 AM
the Ridgid woodworking thickness planer has 4 lead screws, one in each corner. They are turned by a common crank via a timing belt

fred


This is just an idea- there have been many sliding mechanisms made using round rods and separate leadscrew actuating mechanisms. Hand crank or motor driven, it's the same- a lead screw is turned which moves the table one way or the other.

.

darryl
12-27-2009, 01:28 AM
That example came up just a while ago too. That's a good example of driving all the lead screws in sync. My biggest concern is flex. I'm just wondering what the limits would be. It's obvious that a threaded rod is weaker than a solid rod of the same diameter, so it's possible (or probable) that for any machine that required rigidity in a direction that is going to flex the rods, it would not be a suitable method of construction. Better to have solid surfaces carrying the weight of the machine parts, and leave only the motion to the lead screws. If there needs to be more than one lead screw to position the table, so be it, and sync them of course.

There's another design coming about in my head, so I'm just considering the options. I'm far from needing another project right now anyway.

whitis
12-27-2009, 12:47 PM
Rods which are not supported along their full length, even without being threaded, have lousy rigidity unless the diameter of the rod is a substantial fraction of their length. With a threaded rod, you cut away the very outer surface that gives the most rigidity.

A 1" rod theaded 1/8" deep, is effectively a 3/4" rod. A 3/4" rod 48" long will deflect 248mils with a force of 50lbs (machining force from 1/4 HP spindle). Static. Due to resonance, deflection can be much worse. At 24" deflection is 31 mils static. And at 12" length, deflection is 4 mils static. Again, under vibration the deflections can be much worse. It takes a 1-7/8" minor diameter to get the static deflection down to 0.1mils over a mere 12" and that will be much worse with vibration or longer lengths. At 18" you need a minor diameter of 2.5" to get the same deflection. At 24", 3-1/2". At 36", 4.25". At 48", a whopping 5.25". And this is just the static deflection of a single part of the machine, various sources will add up, and the dynamic performance will be worse.

Rotating shafts and lead screws also suffer from whip above their critical speed.

A very long nut such as you describe is also asking for binding. Two nuts spaced with a significant distance apart can also be prone to binding due to differential thermal expansion.

And if you were going to do it, it would probably make more sense to use one threaded rod and one solid one instead of two threaded rods.

Black_Moons
12-27-2009, 05:14 PM
woah I never knew deflection had such a high exponent for deflection vs length.

oldtiffie
12-27-2009, 06:16 PM
Check Machinery's Handbook under "Beams" (page 260 in MHB 27).

In particular see/consider "combined loading".

There are many other texts, tables and calculators on the web.

whitis
12-27-2009, 09:10 PM
woah I never knew deflection had such a high exponent for deflection vs length.

Yeah, it is the cube of the length.

However, increasing cross section in the direction of force also has an exponent of 3 but increasing perpendicular has an exponent of 1. I.E. if the force is downward, the resistance to bending is proportional to the cube of height but only proportional to width. If you increase width and height by the same amount (or in this case diameter), you are looking at the 4th power.

It isn't just how much metal you have, but where you put it. A tube (provided the walls aren't too flimsy) is stronger than the same amount of material (but not the same dimensions) in a solid. However, tubes can be expensive.

If you want to double the length of your machine tool on just one axis but keep the same rigidity, you need to multiply both the width and height of the structural beams by 1.6817928. 2.8284271 times the cross sectional area. Or 5.657 times the total weight of material, and cost, for each beam.

lwalker
12-28-2009, 12:35 PM
There's an MIT professor (Slocum I think) who published some designs of machine tools that could be built cheaply in third world countries. At least one of his tools used two parallel screws (ordinary 1/2-10 bolts) coupled together with a timing belt instead of separate ways and leadscrew. It seems that the thread pitch of off the shelf nuts & bolts is usually accurate enough for most uses. The table sat on nuts instead of linear bearings.
I don't have a link, but if you search the pergatory.mit.edu under his name it might turn up something.



This is just an idea- there have been many sliding mechanisms made using round rods and separate leadscrew actuating mechanisms. Hand crank or motor driven, it's the same- a lead screw is turned which moves the table one way or the other.

What about using acme threaded rod as both the leadscrew and the rail? A typical table would have two, and they'd have to be synced of course. The nuts would be a casting or molding that goes the width of the table, thus giving a large surface area supporting the table and interacting with the acme threads to drive the table back and forth.

I don't know if this method has been used anywhere, but I'd like to see some examples of it, and discuss the benefits and drawbacks.

CountZero
12-28-2009, 03:15 PM
This one for example: http://pergatory.mit.edu/perg/presentations/Mechanics%20of%20Designing%20Precision%20Machines. pdf