[Goemon]

I wanted to ask if ball screws have to be positioned between the linear rails or if you can get away with placing them above or on the other side of the gantry beam?

I am paying the price for not listening to Ger21's excellent advice early on about buying components in the right order so you don't end up with problem like getting stuck with the wrong size motors or ball screws that don't fit between your rails....

I went with hefty 25mm ball screws with oversized double ball-nuts because I figured that a guy could always benefit from a little extra girth and stiffness. The problem is that it doesn't fit under the stage that moves across the rails because those large double nuts are huge....

I can make it all fit by adding risers under the rails or on top of the bearing blocks to increase the amount of clearance but that would make the spindle carriage protrude an extra 2-3" in front of the gantry which seems like it would reduce stiffness.

or

I could mount the ball screw on the top of the gantry (or on the reverse side).

Which is a better option?

if I would have seen this issue coming, I would have cast a channel into the gantry surface to position the ball screw lower than the rails but I didn't and machining one in now is not an option for various reasons.

[R0bbie]

Hi,
I placed both rails and ball-screws on the backside of the gantry for two reasons:
1. As you mention, to keep the spindle as close as possible to the gantry.
2. To keep the screw and rails clear from chips flying around.



Regards,
Rob.

[ger21]

Ballscrew on top would be the better option, imo.

[maxspongebob]

Goemon. Good question.

I always try to visualize the forces at work to get a better understanding of where things should go. The way my brain works might be interesting to some, silly to others, but here it is.

Imagine that there was a tray of mud on the surface of your router. Then imagine a stick in the mud that you use to mix it using your hands. If the mud is very thin, mixing is easy. If thick mud, then not so easy.

Now the analogy. The end of the stick in the mud is the cutter. One of your hands on the stick is the ball screw (can only move back and fourth) and one is the rail bearings (can only keep the stick straight). Hold the stick with the rail bearing hand at the end opposite to the mud. Put the ball screw hand in the middle of the stick then move the stick back and fourth through the mud. The closer the ball screw hand is to the mud, the easier it is to move and easier to keep the stick straight with the ball screw hand. Then imagine if the hands were reversed. Ball screw hand is at the opposite end of the stick from the mud and the rail hand is in the middle. In this configuration when you try to move the stick through the mud, you are increasing the amount of twist on the rail hand to move the stick.

The end result is that when the ball screw is between the cutter and the rail, you are reducing the amount of torsional force on the rail bearings depending on the distance between the cutter and the ball screw. Where the rail bearing is between the cutter and the ball screw, you are increasing the torsion on the bearing depending on the distance.

Granted, there are lots more forces at play here, but it always helps me to visualize the situation. The closer you can put the ball screw to the work and in line with the rail bearing, the better.

[Goemon]

Ballscrew on top would be the better option, imo.

I was thinking it would be easier to place the ball screw on top for alignment reasons. It would certainly keep the rails and spindle closer to the gantry beam. I'm glad it can work like this because it will be a lot neater and stiffer than using risers under the rails. Thanks (again).

I can use the long riser plates to enclose the ball screw (with bellows) for protection instead of mounting them under the rails so they won't go to waste.

The only other option would be to mount the rails on top and have the ball screw up front with the spindle traveling on a precision angle plate. Linear rails seem to be made to hold a lot more weight in horizontal applications than vertical ones (even the ones designed to hold weight in 4 directions). I decided against it though because it created other issues.


For future builds, I need to measure how much space I need ahead of time. any machine bases I build in future will all have channels for ball screws.

[Goemon]

Goemon. Good question.

I always try to visualize the forces at work to get a better understanding of where things should go. The way my brain works might be interesting to some, silly to others, but here it is.

Imagine that there was a tray of mud on the surface of your router. Then imagine a stick in the mud that you use to mix it using your hands. If the mud is very thin, mixing is easy. If thick mud, then not so easy.

Now the analogy. The end of the stick in the mud is the cutter. One of your hands on the stick is the ball screw (can only move back and fourth) and one is the rail bearings (can only keep the stick straight). Hold the stick with the rail bearing hand at the end opposite to the mud. Put the ball screw hand in the middle of the stick then move the stick back and fourth through the mud. The closer the ball screw hand is to the mud, the easier it is to move and easier to keep the stick straight with the ball screw hand. Then imagine if the hands were reversed. Ball screw hand is at the opposite end of the stick from the mud and the rail hand is in the middle. In this configuration when you try to move the stick through the mud, you are increasing the amount of twist on the rail hand to move the stick.

The end result is that when the ball screw is between the cutter and the rail, you are reducing the amount of torsional force on the rail bearings depending on the distance between the cutter and the ball screw. Where the rail bearing is between the cutter and the ball screw, you are increasing the torsion on the bearing depending on the distance.

Granted, there are lots more forces at play here, but it always helps me to visualize the situation. The closer you can put the ball screw to the work and in line with the rail bearing, the better.

I understand that, with all else equal, a ball screw is best positioned between the two rails (in single ball screw axis applications) but that is not an option for me unless I add risers or buy a new and far smaller ball screw with low profile pillow blocks. So... all else is not equal.

Now, I could position the ball screw below the rails and that would be even closer to the cutter (than if it was between the rails) but would that give me any advantage over putting the ball screw on top?

I was working under the assumption that ball screws (or any drive mechanism) were not meant be load bearing like the rails. If you are right and they are meant to reduce cutting forces on the bearing blocks, then I am probably still better off with my 25mm ball screws (with oversized anti backlash nuts) instead of using some 15mm ball screws that fit neatly between the rails.

[maxspongebob]

Do you have a picture of your gantry or a design file so we can see what you are dealing with? You are correct, don't load the screw radially, only axially. Also, yes, bigger is usually better unless you are trying to achieve the absolute max speeds.

[genixia]

I'd put the ball screw up top rather than extend the z forward. Just make sure that the nut to carriage connection is rigid - deflection there becomes backlash.

Sent from my SAMSUNG-SM-G900A using Tapatalk

[Goemon]

Do you have a picture of your gantry or a design file so we can see what you are dealing with? You are correct, don't load the screw radially, only axially. Also, yes, bigger is usually better unless you are trying to achieve the absolute max speeds.

It's a relatively standard fixed gantry / portal mill design aside from what it's made from. It's on it's side today as I was doing some work on it and it's too heavy to move for pics so you'll have to turn your head for the correct orientation. It's a little dirty too from where I was tidying up the corners... it's usually jet black without the dust apart from the edges which currently have exposed carbon fiber (for now).

Here is a link to a pic. Sorry something is wrong with the insert image function on this site so it won't let me insert thumbnails for some reason....

https://postimg.org/image/3n4ildb6r/

And from the other side:

https://s5.postimg.org/g1ralouzb/IMG_4202.jpg


The gantry is all ones piece and is a little over 8 1/2" thick. The mounting surface (above the legs) is approx 8" wide. It's 24" tall but the clearance is far less due to the thickness of the gantry and table. It will have 8" of travel on the Z axis and 33" x 20" on the X and Y. The gantry axis is the long one.

The gantry goes on top of a granite surface plate which together makes up the base and that goes on a steel machine table.

Not sure if any of that makes a difference for any advice on ball screw placement but one thing that is relavent is that I made sure that all of the front surface of the gantry beam and legs is perfectly level and flat to at least 0.01mm (the smallest I can measure here). I also made sure that the top of the gantry was square with the mounting surface (as well as flat). The rear surface is not precision flat but nothing is mounted on that side.

[Goemon]

I'd put the ball screw up top rather than extend the z forward. Just make sure that the nut to carriage connection is rigid - deflection there becomes backlash.

Sent from my SAMSUNG-SM-G900A using Tapatalk

I'm going to try connecting the nut to the carriage using steel or aluminum square bars. I don't have the practice experience yet to know how chunky with part needs to be to avoid deflection so I'll have to experiment. The extra distance from ball nut to carriage makes me think it will need to be a fairly substantial piece of metal.

I have some 1 1/2" thick Mic-6 square bars. I'll try them first.

[wizard]

Hi genome. Fell a sleep during the last attempt to answer and couldn't submit the original response.

First off i don't like the idea of shimming up a saddle on the linear bearings. Unless you go full machine shop and accurately mate the parts, pin them in place and then remachine everything i would not trust the shims to remain in place.

From my stand point a thicker plate for the gantry saddle would be a better option. There are several approaches i will try to outline a couple.

1. Machine into the plate the steps that you might have shimmed the plate. You can thus maintain registration surfaces for the linear bearings in a one piece unit. This does require a bit of metal removal depending upon the required clearance.

2. A different approach is to take a thick plate and bore a hole through it for the leadscrew and a pocket for the nut! With a one inch leadscrew and a big nut this implies a thick plate. However there is an advantage in lowering the leadscrews center height. You do this by removing material where the linear bearings mount. You will still have a lot of mass and that might mean a need for pocketing to lighten the saddle.

3. Similar in concept to a milled out saddle solution would be to do a casting. You can DIY Aluminum fairly easy or have somebody cast one for you. If you have one cast it can also be done in iron. The idea here is far less machining that in suggestions one and two above. You can go crazy with cores putting strength where you need it.

4. I imagine you could do a carbon fiber solution along the lines above. Id go for a design that runs the lead screws through the center of the form. Your the expert as to feasibility here but i could imagine reduced post casting machining. The nice thing here is the centerline for the leadscrew can be put anywhere without a concern about distortion.


In any even ideas off the top of my head. You are likely to get dozens of suggestions because it is fairly common to have to modify machine tool designs like this. Usually this is the result of obsolete parts but also the occasional screw up. So don't let this get you down, in the end you will need to choose what is right for you and your machine.

By the way it might make more sense & cents to consider a smaller leadscrew. It would be less disruptive to your overall design. I don't know how big the nut is thus how drastic the change required but having seen my share of 1" ball screws you might be better off fitting a smaller screw than to blow out the size of many of your saddle components.

[Goemon]

Hi genome. Fell a sleep during the last attempt to answer and couldn't submit the original response.

First off i don't like the idea of shimming up a saddle on the linear bearings. Unless you go full machine shop and accurately mate the parts, pin them in place and then remachine everything i would not trust the shims to remain in place.

From my stand point a thicker plate for the gantry saddle would be a better option. There are several approaches i will try to outline a couple.

1. Machine into the plate the steps that you might have shimmed the plate. You can thus maintain registration surfaces for the linear bearings in a one piece unit. This does require a bit of metal removal depending upon the required clearance.

2. A different approach is to take a thick plate and bore a hole through it for the leadscrew and a pocket for the nut! With a one inch leadscrew and a big nut this implies a thick plate. However there is an advantage in lowering the leadscrews center height. You do this by removing material where the linear bearings mount. You will still have a lot of mass and that might mean a need for pocketing to lighten the saddle.

3. Similar in concept to a milled out saddle solution would be to do a casting. You can DIY Aluminum fairly easy or have somebody cast one for you. If you have one cast it can also be done in iron. The idea here is far less machining that in suggestions one and two above. You can go crazy with cores putting strength where you need it.

4. I imagine you could do a carbon fiber solution along the lines above. Id go for a design that runs the lead screws through the center of the form. Your the expert as to feasibility here but i could imagine reduced post casting machining. The nice thing here is the centerline for the leadscrew can be put anywhere without a concern about distortion.


In any even ideas off the top of my head. You are likely to get dozens of suggestions because it is fairly common to have to modify machine tool designs like this. Usually this is the result of obsolete parts but also the occasional screw up. So don't let this get you down, in the end you will need to choose what is right for you and your machine.

By the way it might make more sense & cents to consider a smaller leadscrew. It would be less disruptive to your overall design. I don't know how big the nut is thus how drastic the change required but having seen my share of 1" ball screws you might be better off fitting a smaller screw than to blow out the size of many of your saddle components.

I like the idea of solving design problems with custom machined parts but the problem is that I just don't have the equiptment to recreate the precision flat surfaces I have now in steel, aluminum or iron. My X-axis stage is an IKO precision ground steel carriage that is over an inch thick with perfectly positioned screw holes for the bearing blocks. It's a nice piece and well beyond my skills to recreate with anything close to that level of precision. It's flat to .0005mm....

Similarly, my X-axis ball screw assembly is an NSK C3 grade precision ground piece with a 5" long double nut, dual precision angular contact bearings at each end and 1.5" thick cast iron end support brackets. Having invested in ground screws for each axis, I would like to use them.

My IKO stage has holes to allow a ball screw to run through the middle but it doesn't solve my issue about not having enough clearance. I would also lose 7" of travel which I can't afford. The ball nut is the main issue. Even if I used custom low profile end supports, the nut flange would not fit under the carriage without risers.

I was looking at how larger ball screws were used on larger VMCs and it seems that they are mostly used on boxed ways which have machined (and ground) channels under the moving table. I have copied that design for my moving table but it's too late to do it on my gantry surface.

I thought about using custom carbon fiber stage assemblies but it would still require precision ground metal surfaces to attach bearing blocks and rails.

If it can be made to work with the ball screw on top of the gantry then that is the path of least resistance for now. A lot more becomes possible once the CNC machine is operational (including your ideas on custom stage assemblies).

It's all part of the learning experience. I am sure that most people's second builds go smoother than the first.... it never even occurred to me to ask the diameter of the ball nut before I bought my screws. Next time I'll ask....

[wizard]

Is the problem the entire nut or just the flange?

If it is just tge flange a 4.5" grinder with a cut off wheel can solve the problem. This assumes that only about a quarter of the flange is an issue.

Yes this is a crude solution but ive seen nuts so adjusted run for decades. Take plenty of percautions to keep grinding dust off the screw and nut internals.

[maxspongebob]

I agree, you probably should go with the screw on the top of the gantry. Once you have your machine up and running you can use it to make other parts to improve your design. Sometimes you just have to compromise to make headway on the build.

BTW, on the ball screw nuts, are they the one's with the circular flange, or with 2 flats? I have always been fond of the design of the Bridgeport leadscrew assembly design. Something about having both X and Y screws use a single nut bracket. One day I will design a table that has a bracket like this so that there is only a couple mm of clearance between the screws.

Leadscrew Assembly

[Goemon]

I agree, you probably should go with the screw on the top of the gantry. Once you have your machine up and running you can use it to make other parts to improve your design. Sometimes you just have to compromise to make headway on the build.

BTW, on the ball screw nuts, are they the one's with the circular flange, or with 2 flats? I have always been fond of the design of the Bridgeport leadscrew assembly design. Something about having both X and Y screws use a single nut bracket. One day I will design a table that has a bracket like this so that there is only a couple mm of clearance between the screws.

Leadscrew Assembly

This is what it looks like:

https://s5.postimg.org/endqp4puf/IMG_4019.jpg

The flange is round with one flat edge. I really don't like the design of most ball nuts or end supports I have seen. I think they should be designed so they can be easily mounted without an additional bracket and I would prefer a rectangular shape on the outside. Long ball screws like this are clearly made for horizontal, not vertical, travel so a flatter and wider design with side mounting holes would be better imo. Every extra part you need to buy is a potential source of innacuracy (as well as mounting costs for no good reason).

Maybe I am the only guy on the planet who wants square nuts with precision flat mounting surfaces....


Wizard, the flange is the only part of the nut with mounting holes. If I removed it, there is nowhere else suitable for mounting with any accuracy.

It looks like the double nut can be separated and I briefly considered removing the larger front half to increase travel but then I reminded myself that I bought this double nut precision screw for a reason. I was initially a little irritated by how much travel was lost with a double nut though. You end up with 32-33" of travel from a 45" OAL screw....

[maxspongebob]

I think what WIZARD is saying is to only remove part of the nut flange. Like this.

[Goemon]

I think what WIZARD is saying is to only remove part of the nut flange. Like this.

I don't think it would fit even if I did it that way. It might if I used larger rails. I have been waiting for a while for my 25mm THK rails to arrive to replace my 20mm rails for that axis. They would naturally offer a little more clearance but I still doubt it would fit. The nut would need 2" of clearance even without the flange.

Still, the mod in your pic illustrates nicely why flat square ball nuts would make more sense than round ones.

Given the choice between mounting the ball screw on top or taking an angle grinder to the ball nut, I think I would rather mount it on top. If it was a cheap Chinese screw it would be different but I can't bring myself to mod a nice NSK one.

Once I mod the nut and replace the precision cast iron end supports with homemade low profile ones, I can't see them being that precise anymore.

[wizard]

I think what WIZARD is saying is to only remove part of the nut flange. Like this.

Yes that is exactly what I was referring to. I've seen it done many times, sometimes on purpose, sometimes the result of a screw up and sometimes to get around a casting defect.

Unfortunately it sounds like Genome doensn't have the clearance to go this route.

[Goemon]

Yes that is exactly what I was referring to. I've seen it done many times, sometimes on purpose, sometimes the result of a screw up and sometimes to get around a casting defect.

Unfortunately it sounds like Genome doensn't have the clearance to go this route.

I was never expecting the flange to fit under the stage. I was planning to bolt the top edge of the flange to the side of the stage to push it (instead of bolting the whole thing under with a seperate ball nut bracket). It was not even close to fitting though.

My new rails arrived today so I am going to check how much difference it makes. Just my luck, I ordered 43" long 25mm rails and the guy sent me 53" 25mm rails instead. Normally, I might have been happy about the extra travel but my gantry is only 48" wide and I can't stand the idea of having 5" of rail poking out the end...

[maxspongebob]

Just cut them to length with an angle grinder. Also send pics when you have it together. Can't wait to see it.