[jntambs]

Hi group,

I have been lurking for a couple months now. I have been putting together stuff for an aluminum constructed wood/aluminum router type machine. I have debated about driving the gantry and can't decide on two screws vs one. I have access to a couple large diameter lead precision ground ball screws and nuts, one is 1.5 inch diameter thread length 48 in (y axis), and the other is 2.25 inch diameter with 67 in thread length (x axis).

The motors I have are the reliance electric from ebay currently listed as ebay 3800676277. These are DC brushed servos partial specs (from listing) 120 v, torque constant is 5.5 k=lb in/amp, back emf 65v/kRPM, 1750 RPM. No mention of amps. NEMA 34, 500 line encoders.

I know these are big screws but I have them , are they too big for these motors? Are there problems with using screws this size as far as 'momentum' and damage to the motors?

The table will be somewhere in the neighborhood of x=6 ft and y= 4 ft.

I thought about the 5/8 in screws and using two with a single motor but was worried about whip, however after looking at these screws I am concerned they may be too big.

Also I have not decided whether to use geckos or rutex.

Any thoughts about the screws or drivers would be appreciated.

Thanks,
Jim

[Mr.Chips]

Look into rotating ball screws, here is a link http://www.nsk.com/eng/journal/mc08/mc8_09.pdf
this would eleminate the whip of a rotating screw.
Hager

[jntambs]

Hegar,

Thanks for the reply. I saw that post and checked out the PDF last night. I have thought about rotating the nut instead of the screw and that may be the way to go, just not sure yet.

I am still interested in hearing thoughts about spinning a screw that size with the motors I have. Would the torque requirements be outrageous? With gearing down because of the pitch would I be setting myself up to destroy belts and pulleys?

I have considered the possible problems with racking because of using only one screw on X and wonder if beefing up the bracing to the nut in an all aluminum machine would be acceptable, or would I just be asking for trouble. I plan on using hardened thompson shaft, supported, and open super pillow block bearings for X. For Y and Z I haven't decided between a smaller diameter thompson set up or thk linear rails. For X I was able to acquire a couple 6 foot lengths of 1.5" shaft and bearings, will have to fabricate the supports but that is not a problem. If I can find some 3/4 or 1" shaft for Y I may go that way, if not I have the linear rail which I am considering for a plasma table down the road.

Looking for thoughts about these screws with these motors. Also which would be better gecko or rutex .

Thanks,
Jim

[arvidb]

The problem with large diameter screws is that their rotary inertia rises apprx. proportional to the screw's radius^4! This means that if you double the width of the screw, you will increase the rotary inertia of the screw 16 times!

Rotary inertia is directly proportional to the acceleration and deceleration times; thus, if you double the dia of the screw, you will get 16 times slower acceleration and deceleration! Your top speed will be the same though.

The inertia of the gantry itself is actually quite small compared to the inertia of the screw. As an example:

700 mm long, 16 mm dia lead screw (SKF PVK16x5) would have a rotary inertia of ca 32 kg*mm^2.

A 20 kg gantry coupled through the same screw would add some 12-13 kg*mm^2. The gantry contribution varies linearly with weight of the gantry and lead of screw, so if you have a screw with 10 mm lead instead of 5 mm, the gantry's inertia contribution doubles to about 25 kg*mm^2.

Suppose we take the screw above, and make it 2.25" thick and 1800 mm long (about 70"). This would increase the rotary inertia >400 times, for a total of 13200 kg*mm^2! (= 0.0132 kg*m^2)

If your motors are capable of 5 Nm peak torque (corresponding to maybe a 400W servo motor) your acceleration would be 5 [Nm] / 0.0132 [kg*m^2] = about 379 rad/s^2, or 60 rev/s^2. So it would take you about 0.42 seconds to reach 1500 RPM. To me this seems like a sluggish performance, even with this quite powerful motor.

Arvid

[jntambs]

Arvid,


Thanks for the reply. I was thinking this would be tough on the motors and pulleys. Although 0.4 sec to get to 1500rpm doesn't sound too slow, I guess if you consider a lengthy program with alot of starts and stops that could add up to an annoying amount of time.

Thanks,
Jim

[arvidb]

Hi Jim,

Also remeber that the 0.4 secs are just for accelerating the screw itself. All cutting resistance, drag torque, the inertia of the gantry itself etc will have to be added.

I don't think you will have a problem with pulley wear - the motor's peak torque decides how much torque they see, and that would have been thesame with a smaller screw. Of course, they'll see it longer with the bigger screw, but with correctly dimensioned pulleys my *guess* is that it won't matter.

Since the acceleration and deceleration times are longer, your motors will have to be able to develop the peak torque for a longer time than with a smaller screw. This might lead to a problem with overheating if you do not dimension correctly.

Also, a bigger problem from the motor's (and amplifier's) point of view might be that the amount of braking energy that has to be absorbed/dissipated is directly proportional to the rotary inertia of the system. So you might need bigger braking resistors.

I'm not saying it can't be done or that a bigger screw is inherently worse than a smaller screw, but I do think it requires a more careful design!

Arvid

[jntambs]

Arvid,

Hi and thanks again for the response. I was thinking the decelleration may be a strain for the motors, probably moreso than accelleration, and that concerned me. Regarding the breaking resistors you speak of......are they in the driver or in the motor itself?



Thankyou once again for your input,

Jim

[balsaman]

What is the pitch of the ball screws?

E

[jntambs]

Balsaman,

1/2 inch pitch

Thanks,
Jim

[balsaman]

In that case at 400 rpm you will get 200" per minute. That is lots. I would go with a big ratio (5:1) and I imagine all will be well, including good accel and decel rates.

E

[arvidb]

Jim,

the absorption of braking energy (by resistor and/or capacitor) should be taken care of by the drivers. In this case you have to look up the specs of the drivers and make sure you don't feed them more energy that they can take care of.

It is also possible to have drivers that act as generators while braking, and in that case you need some kind of overvoltage protection between your power supply and the drivers. This depends on the driver - you have to check the documentation.

Deceleration is done with the same peak torque as acceleration, so should not be harder on the motors.

As balsaman says, gearing down the motor will help. A 2:1 reduction doubles your acceleration, 3:1 triples it etc. Still, to get the same acceleration as a direct driven 16 mm lead screw would give you, you would have to gear down the 2,25" screw with a 67:1 ratio! This is because of the radius^4 relationship between screw radius and its rotary inertia. Gearing down only effects rotary inertia linearly.

My advice is: if you know what acceleration you need, calculate the motor torque you need to get it (together with suitable gearing ratio), and see if it's feasible. If you don't know what acceleration you need, just gear down as much as you can while still getting enough top speed, and try it!

Arvid

[jntambs]

Balsaman,

Thankyou for that info. I need to search these archives and start saving the formulas used to make the various calculations I've seen on this site. Is there a file area with the various valuable calculations in relation to these machines on this site? I would imagine many of the formulas, at least mechanical, are in Machinery Handbook. I have also seen much in relation to calcs regarding the electronics and motors here, just can't remember which posts!!! This site is outstanding!

Thanks,
Jim



Arvid,

Again I thank you for your input. I am still debating the one screw vs two for X axis issue, I suspect I will probably give it a try with the single screw as that is what I have and see if it works. I am sure I will be pestering the experts with many more posts after I make some definitive choices on specifics.
I got my aluminum sheet (1/2") and am now trying to firm up how I want this to look and function. Unfortunately I do not have any experience with using CAD software so I am trying to get that down as well.

Thanks,
Jim

[arvidb]

It seems I made a pretty serious mistake about how gearing affects rotary inertia. It is proportional to the square of the gear ratio, so with a pulley ratio of 2:1 you get a 4:1 reduction in the rotary inertia seen by the motor.

So you would only need something like an 8:1 reduction to make the 2,25" screw look like a 16 mm screw to the motor.

This also applies to the "gear ratio" of the lead screw, i.e. its lead. The gantry inertia => rotary inertia conversion performed by the screw is proportional to the screw lead squared (rotary inertia ~ gantry inertia * lead^2).

Sorry about this. The advice above still holds, though.

Arvid

[jntambs]

Hi Arvid,

Hey, no problem here man! I appreciate all of the discussion and cannot tell you how much I have learned at this site and how much it is appreciated.

Sounds like this will even be better (easier on the mechanicals) than I was thinking, now I gotta start thinking about putting some metal together and get going on this contraption.

Again, I thank you for your input and advice!

Jim