1x1.5m table, new machine design

[stef_d]

I am at the design stage of my new machine.
This is what i budgeted so far:
It will have a 1x1.5m table. structure is made of stainless steel tubing (50x50mm) and gantry is aluminium. backbone of the gantry is made of 2 beams, 50x80mm, in 4mm thick aluminium. and spans around 1.2m.

It runs on 20mm Hiwin rails (2 on x, 2 on y and 2 on z)
screws are 20mm 2TPI on x and y and 5TPI on z.
all motors are nema 34 640oz.
Out of this, i should get decent speed and decent accuracy for wood working (production)

I am trying to decide on whether to use twin screws on x, with slaved motors, or use a single screw.
here are the drawings for both approaches.
the 2 screw setup has the gantry runing inside the x rails, where as the 1 screw setup runs outside.
this means that i can have a higher gantry on the single screw setup, with support added at the bottom of the gantry.

what would be the best design/ compromise ?

[hemsworthlad]

Not a long way off my design and size, I would go with 1st pic.
It,s a lot more hassle using 2 screws belive me I know.?

Look here. Started but not finished

[ger21]

I'd go with two screws, but still build it like the first pic, and put the screws underneath the rails, to keep them clean. I've had no problem with my 2 screws driving the gantry, and all future machines I build will be driven from both sides using slaved motors. It's simple, and works well.

[hemsworthlad]

I'd go with two screws, but still build it like the first pic, and put the screws underneath the rails, to keep them clean. I've had no problem with my 2 screws driving the gantry, and all future machines I build will be driven from both sides using slaved motors. It's simple, and works well.

Ger do you have a build i could look at, I did try slaved motors but really struggled keeping them in sync, but i think my computer may be the cause of most of my troubles along with a under powered PS.

Built a new PS and I,m just waiting for a smooth stepper to arrive which should eliminate my pewter problems, so I may give it another go when that arrives.

Cheers.
Dean.

PS: Sorry Stef-D wont high jack your thread anymore. .Lol

[jsheerin]

I'd agree with Gerry and go with number 1. I'd also suggest dual drives for the x axis. I have that on my current router with no problems but am currently redesigning it to be more like your first design (rails above the table, very short gantry legs). However I am planning on rack and pinion drives, as that gives a better force / speed curve using typically available motors and avoids the limitations of screw whipping at high speeds.

I'd also suggest that if you add a plate to the outside of your x axis structure, you'd significantly stiffen it up.

Dean,
It seems odd that a computer power supply would be causing problems... I'd be more inclined to look at the machine, like binding issues causing the motors to miss steps. You could try to isolate the problem to the machine or the computer / electronics by taking your motors off the machine and running them together to see if they get out of sync. If they do, it's the electronics. If they don't, it's the machine.

[hemsworthlad]

Dean,
It seems odd that a computer power supply would be causing problems... I'd be more inclined to look at the machine, like binding issues causing the motors to miss steps. You could try to isolate the problem to the machine or the computer / electronics by taking your motors off the machine and running them together to see if they get out of sync. If they do, it's the electronics. If they don't, it's the machine.

Hi Jsheer,
Dont want to high jack this thread so i'll make it a quicky.
No you miss understood me, it's Parallel Port on the computer causing problems not PS, the power supply in my contol wasn't big enough so motors were down on volts that's what I meant when i said PS.
No binding issue's with machine, smooth as silk.

cheers
dean.

[jsheerin]

Ah, gotcha. I had the same thing - crappy on board lpt port with inadequate current capabilities. I dropped in a pci dual lpt port card ($15) and had to combine pins from both ports in a custom cable to get it run my G540.

John

[stef_d]

Thanks for the input.
I will carry on the design work based on option one then, with two screws..would it make more sense to use two smaller motor for the x ? rather than two 640oz as i planned ?
I guess this would mean higher speeds on x. but then, i am not too concerned about speed. i dont think two 640oz would crawl.

[ger21]

I'd keep the 640 oz motors. However, 640 oz motors come in very different flavors. Be sure to use high current (6+ amps), low inductance motors for maximum speed.

[jsheerin]

To add to Gerry's point, if you pick good motors, you'll have more force at a given speed with the larger motors, or to look at it another way, you'll be able to go faster for the same force. You could also gear the motors differently either through your choice of screws, belt reduction, or rack and pinion. This would change the force versus speed curve so you could have about the same force as the smaller motors but much higher speed. Of course that would come at the expense of resolution. Also, screw whipping might limit your speeds before anything else, so you might want to consider that when making these decisions.

[stef_d]

I had a rethink, and i decided (so far) to design it single screw, but to keep the option of doubling it if required.
Here is my thinking.
I can see a big advantage in having a double screw:
If i am routing something at the far end of the gantry, and i pull in the midle on a single screw, it's bound the rack if my gantry is not rock solid. especially at the connection between the rails and the gantry. in fact, i would need good carriages, good quality linear rails, with little play in them to prevent this from hapening.
hence, the double screw makes sence...
however, the precision of a screw (at a reasonable price) could be say, 0.003in/ft, over 1.5m, that's 0.015in, or 0.381mm...ouch.. that sounds bad, and probably the limiting factor on this type of machines.
now if i take 2 screws, and find myself unlucky enough to have mismatched errors, one of them could be 0.381mm ahead, and the other 0.381mm behind.
that's the same as pining one side of the gantry, and forcing the other one 0.8mm (almost 1mm !) out of alignement.
that's probably ok if your gantry is slightly flexible, but probably not so if it is rock solid. if the gantry cant flex that much ( and i seriously hope mine wont be able to) then it will bind..

solution to this is to buy expensive (high precision) screws, or be lucky and have matching error in both screws.
that's my take on it, considering i am no expert. does it make sence ?

[jsheerin]

Several thoughts:
In some cases screw error is not cumulative. In other words there is an allowable error over one length of the screw, but a different allowable error over the entire length. What this is depends on the grade of screw you're buying. I looked up a C7 spec here:
http://www.kss-ballscrew.com/media/t...KSS_CAT74E.pdf
It looks like they'd allow 0.52mm total error over a 1.5m length if I'm reading it right. However it would likely be less than that.

You can map your screws and input that into your software so the error is compensated. I'm not sure if Mach3 is smart enough to do that for two slaved axes.

You can also mechanically compensate for screw error using various cam type setups.

You can using a moving knot cable arrangement with a single screw to prevent your gantry from racking. Doing a search should find information on that.

Ultimately it probably comes down to which way you want to try to solve the problem.