Mounting linear rails. How precise does it need to be precisely?
What do you guys target for horizontal and vertical parallelism accuracy when mounting (square) linear rails?
I've had mixed success trying to mount my Y and Z rails using digital calipers and a dial gauge to test how parallel they are but I think the quality is a limiting factor. In theory, my digital calipers are accurate to 0.01mm but I have trouble achieving the same measurement twice for a fixed distance. In reality, I only trust their accuracy to 0.1mm. I place even less faith in my dial gauge / magnetic base stand set (which seems to just produce random readings as I slide the bearing block back and forth).
What would you guys consider to be acceptable mounting parallel accuracy for rails over a 30" or 48" slide?
The rails seem to function ok if they are parallel to 0.1mm in that the stage slides up and down without that horrible grinding noise that seems to occur when I get it too far out of tolerance. That doesn't mean it will perform well on a finished CNC machine though.
Do you add any adjustment screws for fine tuning? I find it very difficult to make 0.05mm adjustments to rails by hand. I am wondering if there are some other tools to help me....
I achieved the best results (so far) by just using machinist parallels between the rails as I drill the screw holes in the mounting surface. Still not much better than 0.1mm though.
Re: Mounting linear rails. How precise does it need to be precisely?
bolt both ends of one rail down, push it into place to get it straight. clamp a test indicator on the rail and test the movement against a straight edge.
then bolt the blocks to the moving gantry and get the two ends of the second rail bolted down. push the gantry around while tightening the bolts of the second rail.
rails are usually very stiff in all 5 axis and if your mounting surfaces are not flat then you will be overstressing the bearing blocks without any additional load.
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
Goemon
What do you guys target for horizontal and vertical parallelism accuracy when mounting (square) linear rails?
I've had mixed success trying to mount my Y and Z rails using digital calipers and a dial gauge to test how parallel they are but I think the quality is a limiting factor. In theory, my digital calipers are accurate to 0.01mm but I have trouble achieving the same measurement twice for a fixed distance. In reality, I only trust their accuracy to 0.1mm. I place even less faith in my dial gauge / magnetic base stand set (which seems to just produce random readings as I slide the bearing block back and forth).
What would you guys consider to be acceptable mounting parallel accuracy for rails over a 30" or 48" slide?
The rails seem to function ok if they are parallel to 0.1mm in that the stage slides up and down without that horrible grinding noise that seems to occur when I get it too far out of tolerance. That doesn't mean it will perform well on a finished CNC machine though.
Do you add any adjustment screws for fine tuning? I find it very difficult to make 0.05mm adjustments to rails by hand. I am wondering if there are some other tools to help me....
I achieved the best results (so far) by just using machinist parallels between the rails as I drill the screw holes in the mounting surface. Still not much better than 0.1mm though.
That is why Linear Rails are used on a machined surface, with machined pockets, or a machined edge with tapered wedges or screws the clamp them to the machined surface, they want to be better that .1mm , .025mm would be close to what you should have for a hobby build, this also will relate to how much clearance the Bearing Blocks have, if you have Bearing Blocks that have a preload, then the rails have to be perfect, if they are just general Bearing Blocks with no preload than you can get away with up to .05mm over the complete travel, they wear quite rapidly if Bearing to Rail alignment is not correct
You can not use a digital caliper to set these Rails up, you need to have a minimum of a .0005" indicator or a test indicator
Check the Linear Rail manufactures web site, they have examples of how accurate your Rails have to be mounted, for them to work correctly
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
Eldon_Joh
bolt both ends of one rail down, push it into place to get it straight. clamp a test indicator on the rail and test the movement against a straight edge.
then bolt the blocks to the moving gantry and get the two ends of the second rail bolted down. push the gantry around while tightening the bolts of the second rail.
rails are usually very stiff in all 5 axis and if your mounting surfaces are not flat then you will be overstressing the bearing blocks without any additional load.
I have that part down ok (mostly). The question I have is about how much precision I need to target when measuring. When you mount your rails and then measure, what do you target for parallel tolerances? I.e. Do you consider the mount a success when the rails are parallel to 0.05mm or .01mm or 0.001mm etc? It's the fine adjustments I am unsure about.
If you have a moving gantry machine, I imagine it is even harder to measure how parallel the mount is as most calipers and dial gauge stand / arms don't stretch that far. One of the reasons I chose to make a fixed gantry machine and have the long axis on the gantry is that I wanted all of the sets of rails to be mounted close enough to use precision parallels and close enough that both rails could be mounted to the same plate to avoid vertical parallel issues. I couldn't (and still can't) figure out how to measure the distance between rails to 0.01mm accuracy if the rails are 36" or 48" apart....
All of my mounting surfaces are flat to a minimum of 0.005" on the specs and they all measure the same to 0.01mm on all edges with my digital calipers, so.... any issues are almost certainly created by me...
Aside from not being sure about much parallel precision to target, the other (and more frustrating) issue I am finding is that if I unscrew the rails after the initial mount and reattach them, they don't always end up in the same position even though they are going in the same place using the same screw holes as before...
I thought that maybe there might be some products available to help with fine tuning but I can't find any. I am just about to try my diy fine tuning screw design (unless anyone here has a better way). If my diy fine adjustment device works, I will be able to make fine adjustments with the turn of a screw when I find mounting errors. I figured that would be easier than trying to correct 0.1mm errors with my fingers.
Re: Mounting linear rails. How precise does it need to be precisely?
The stiffness of the system the rails are bolted to are what define how straight the rails need to be relative to each other, simply because you can over stress the bearing blocks by twisting the rail as its clamped down to a non flat surface. but this isn't an exact science without a lot more information because you also have to factor in the stiffness of the bearings as well as the entire system.
so if you look up the data sheet for your bearings they will include moment loads. apply similar loads to whatever structure the moving bearing block is bolted to and measure how much it bends.
the distance between the linear rails is perhaps the easiest problem to fix, simply by bolting the moving bearings to a piece of metal, slide them along both rails and bolt the rails progressively as you slide the bearing. if you want to, you can intentionally make both rails a curve if you want.
the axial twist of the linear rail mounting surface is perhaps a much more difficult problem to fix because a low tensile 6mm bolt can provide 2000 pounds of axial force on the rail before the bolt stretches. if the rail is 20mm wide then it will take about 20 pound-meters of torque which is approximately 60 foot pounds of torque to stretch the bolt,
if i'm not mistaken a lot of 20mm wide, profile rails, even a short bearing block can handle 60 foot pounds of torque applied, so you're sort of safe there, but only if you use low grade 6mm bolts. you can also not torque them beyond 5 foot pounds.. anyhow its no surprise to me that you notice some changes in noise as you slide the bearings across the rail, but that change in sound is not evidence that you're actually at the limits of the bearings.
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
Eldon_Joh
The stiffness of the system the rails are bolted to are what define how straight the rails need to be relative to each other, simply because you can over stress the bearing blocks by twisting the rail as its clamped down to a non flat surface. but this isn't an exact science without a lot more information because you also have to factor in the stiffness of the bearings as well as the entire system.
so if you look up the data sheet for your bearings they will include moment loads. apply similar loads to whatever structure the moving bearing block is bolted to and measure how much it bends.
the distance between the linear rails is perhaps the easiest problem to fix, simply by bolting the moving bearings to a piece of metal, slide them along both rails and bolt the rails progressively as you slide the bearing. if you want to, you can intentionally make both rails a curve if you want.
the axial twist of the linear rail mounting surface is perhaps a much more difficult problem to fix because a low tensile 6mm bolt can provide 2000 pounds of axial force on the rail before the bolt stretches. if the rail is 20mm wide then it will take about 20 pound-meters of torque which is approximately 60 foot pounds of torque to stretch the bolt,
if i'm not mistaken a lot of 20mm wide, profile rails, even a short bearing block can handle 60 foot pounds of torque applied, so you're sort of safe there, but only if you use low grade 6mm bolts. you can also not torque them beyond 5 foot pounds.. anyhow its no surprise to me that you notice some changes in noise as you slide the bearings across the rail, but that change in sound is not evidence that you're actually at the limits of the bearings.
It sounds like there is no widely used specific target for how parallel a mount needs to be other than making it as precise as the person is able. I was assuming that, as people were using dial gauges and calipers to measure their success, they would have an idea of what they were targeting or what "good" looks like.
I have looked at the specs of my rails many times but I find them too confusing when it comes to mounting tolorance requirements. If it really does depend on the broad variables you mentioned, I can't imagine that many people here are being that scientific when it comes to calculating target tolerances. Math was never my favorite subject though...
Maybe I am overthinking this bit. The reality is that my (lack of) ability to measure accurately beyond a few decimal points is going to be a limiting factor regardless of mounting technique.
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
Goemon
It sounds like there is no widely used specific target for how parallel a mount needs to be other than making it as precise as the person is able.
correct, because the rails usually set the global accuracy of the machine.
Quote:
I have looked at the specs of my rails many times but I find them too confusing when it comes to mounting tolorance requirements. If it really does depend on the broad variables you mentioned, I can't imagine that many people here are being that scientific when it comes to calculating target tolerances. .
the mounting tolerance requirements in the documentation may be specified for a situation where two parallel rails are bolted together on the same, very high stiffness structure. the tolerance will likely be less than 20 microns in this case. but the reason why is simply because misalignment will load the bearing.
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
Eldon_Joh
correct, because the rails usually set the global accuracy of the machine.
the mounting tolerance requirements in the documentation may be specified for a situation where two parallel rails are bolted together on the same, very high stiffness structure. the tolerance will likely be less than 20 microns in this case. but the reason why is simply because misalignment will load the bearing.
I guess that any target stated in microns will be of little help to diy builders here. It's well beyond what people can see or measure with digital calipers. I have some 50 micron micro-lenses and they are too small to see individually without a microscope.
I had a (very brief) look at the sort of equiptment the pros use to measure rail mounting accuracy and it's definately not something anyone is going to buy for a diy build unless they planned on starting a CNC router manufacturing business. Even a pair of digital calipers that goes beyond .01mm is getting close to $1,000.
If rail mounting errors are too small to detect with digital calipers then, hopefully, the effect of those errors will also be too small to see.
I just invested in better calipers with carbon fiber pinchers so hopefully there will be less flex which should make it easier to achieve consistent readings at least. Somebody needs to invent some double calipers which allow you to measure the distance between rails at two points at the same time. It would make this job easier..,.
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
Goemon
I guess that any target stated in microns will be of little help to diy builders here. It's well beyond what people can see or measure with digital calipers. I have some 50 micron micro-lenses and they are too small to see individually without a microscope.
I had a (very brief) look at the sort of equiptment the pros use to measure rail mounting accuracy and it's definately not something anyone is going to buy for a diy build unless they planned on starting a CNC router manufacturing business. Even a pair of digital calipers that goes beyond .01mm is getting close to $1,000.
If rail mounting errors are too small to detect with digital calipers then, hopefully, the effect of those errors will also be too small to see.
I just invested in better calipers with carbon fiber pinchers so hopefully there will be less flex which should make it easier to achieve consistent readings at least. Somebody needs to invent some double calipers which allow you to measure the distance between rails at two points at the same time. It would make this job easier..,.
You can't use digital calipers to set up rails, most Hobby users are buying indicator's to set them up correctly, they are not expensive and get the job done, you might get by with something like this which is under $10, you also get what you pay for in instruments like this, ideal you would be buy one for $120 and up, but this is the type of instrument you need to set up the rails, you will need a magnetic stand or some way to mount it to use it correctly
https://www.ebay.com/p/0-01mm-Accura...tor/2119185824
Re: Mounting linear rails. How precise does it need to be precisely?
Goemon,
I'm curious as to what you found that the pros use.
A tenths indicator and adjustable mag base are all you should need. Set your mag base on a truck and your needle on the side of the other rail, adjust to 0-0, set your needle on the top of the rail, adjust to 0-0, check the side again and adjust as necessary.
If your mag base can't reach to the other rail, set a large parallel or square between the two and adjust the parallel/square until 0-0 and repeat the above procedure on the side of the parallel.
A tenths indicator is $50 from Shars. An adjustable mag base is $60. If necessary, a 24" parallel is $232.
And now that I think of it, a set of 3-4-6 blocks can be used in place of a parallel if the highest accuracy isn't needed.
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
footpetaljones
Goemon,
I'm curious as to what you found that the pros use.
A tenths indicator and adjustable mag base are all you should need. Set your mag base on a truck and your needle on the side of the other rail, adjust to 0-0, set your needle on the top of the rail, adjust to 0-0, check the side again and adjust as necessary.
If your mag base can't reach to the other rail, set a large parallel or square between the two and adjust the parallel/square until 0-0 and repeat the above procedure on the side of the parallel.
A tenths indicator is $50 from Shars. An adjustable mag base is $60. If necessary, a 24" parallel is $232.
And now that I think of it, a set of 3-4-6 blocks can be used in place of a parallel if the highest accuracy isn't needed.
The pros use expensive tools like a laser inferometer and others to allow them to measure how straight, level and parallel a mount is with sub-micron level accuracy. They also use precision machined bases with a pair of pre-machined parallel datum planes.
Obviusly most people here won't have access to measuring tools that allow that level of precision. Even if we use the best mounting techniques possible, the cheap eBay gauges and calipers most of us use struggle to measure with accuracy much better than .1mm.
But... none of that is my main issue. The part I don't understand and that hasn't been addressed with any of the responses so far is how adjustments are made and what level of accuracy people target. Apparently, even quality brand name rails, are not straight to begin with. So, to mount them on a flat surface (as opposed to machined channels) there needs to be a way of making small adjustments to the make the rails parallel from end to end. Moving rails by hand won't work unless you are strong enough to bend 1" steel bars.
I have mounted rails that are parallel at both ends but not in the middle.... all my rails are THK, IKO and Misumi - I.e. Good brands....
Anyway, I have given up on this problem for my build. I bought 3 old actuators so I can use their precision machined bases and carriages with my rails. It saves a lot of time. I now have rails mounted with greater precision that I ever could have achieved on my own (or been able to measure).
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
They also use precision machined bases with a pair of pre-machined parallel datum planes.
Exactly. Commercial machines don't require fancy tools to install rails. They get mounted against a machined flat and straight surface.
This website describes one way to set rails, using a piece of music wire stretched tightly as a reference.
Machine frame - MadVac CNC
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
Goemon
So, to mount them on a flat surface (as opposed to machined channels) there needs to be a way of making small adjustments to the make the rails parallel from end to end. Moving rails by hand won't work unless you are strong enough to bend 1" steel bars.
as long as the two rails are mounted on flat, parallel paths, it doesn't matter if they are co-planer, as all you need to do is shim two of the bearing blocks. axial twist of the mounting surface is perhaps more important to get right, but one rail tilted relative to the other doesn't matter provided you shim the blocks accordingly.
anyhow the way i did it was to bolt the two ends of the rails the same distance apart, then temporarily bolt two blocks to each other, slide them down the rail progressively tightening the bolts. yes you really can push the rails quite a bit. the hardened rails that have bolt holes every 30mm might be quite a bit more work to get straightened (because the distortion may be significantly higher, hence double the number of bolt holes), but the ones with bolt holes every 60mm are relatively easy to straighten out.
one of the difficulties of bolting is tightening the bolt will slide the rail one side or the other unless the bolt hole is tapped vertical and centered in the hole. mine weren't, so i had to use a C clamp to hold the rail while tightening the bolt.
30mm rails will be 16 times stiffer than 15mm rails, so it quickly gets out of hand. for larger rails you may need to drill and tap some adjusting bolts to push the rail around if you don't have a guide milled into the frame of the machine to push the rail up against.
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
ger21
Exactly. Commercial machines don't require fancy tools to install rails. They get mounted against a machined flat and straight surface.
This website describes one way to set rails, using a piece of music wire stretched tightly as a reference.
Machine frame - MadVac CNC
The fancy tools are not to install them, they are for measuring that you got it right. But, that should be a formality with an accurately machined datum.
Now that I have bought 3 precision actuator bases with carriages, I am hoping that my rail install will be equally easy (and precise).
It should be easier for my second build because I'll have a CNC machine to make my own datum plane bases with precisely drilled mounting holes.
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
Eldon_Joh
as long as the two rails are mounted on flat, parallel paths, it doesn't matter if they are co-planer, as all you need to do is shim two of the bearing blocks. axial twist of the mounting surface is perhaps more important to get right, but one rail tilted relative to the other doesn't matter provided you shim the blocks accordingly.
anyhow the way i did it was to bolt the two ends of the rails the same distance apart, then temporarily bolt two blocks to each other, slide them down the rail progressively tightening the bolts. yes you really can push the rails quite a bit. the hardened rails that have bolt holes every 30mm might be quite a bit more work to get straightened (because the distortion may be significantly higher, hence double the number of bolt holes), but the ones with bolt holes every 60mm are relatively easy to straighten out.
one of the difficulties of bolting is tightening the bolt will slide the rail one side or the other unless the bolt hole is tapped vertical and centered in the hole. mine weren't, so i had to use a C clamp to hold the rail while tightening the bolt.
30mm rails will be 16 times stiffer than 15mm rails, so it quickly gets out of hand. for larger rails you may need to drill and tap some adjusting bolts to push the rail around if you don't have a guide milled into the frame of the machine to push the rail up against.
Mounting the rails parallel was / is my issue because long rails are not particularly straight. Mine bend in at the middle which either requires machined channels or the ability to make horizontal adjustments at specific points before the rail mounting screws are tightened. I devised my own adjustment blocks with horizontal screws that allowed such adjustments. I got them close to being parallel but not close enough.
Everything becomes easier when you have a professionally machined actuator base to work with. That's what I should have started with from the beginning. There are tons of good deals on used actuators. I could have saved myself a lot of time and money but you live and learn I guess.
Re: Mounting linear rails. How precise does it need to be precisely?
Re: Mounting linear rails. How precise does it need to be precisely?
Quote:
Originally Posted by
rowbare
I saw that video a while ago. It's informative but it wasn't helpful to me for mounting rails on a flat plate from scratch. The guys in the video are mounting the rails on a pre-machined base with milled channels with perfect datum planes so no adjustments are necessary. Having a base like that changes everything.
I have two of the three actuators I bought in hand now so I have bases like the one in the video. As soon as I saw them and felt how smooth the motion was, I knew that there was no chance that I ever would have been able to do as good of a job myself without using a (decent) CNC machine.
It's not just the bases either. The design of some of the professional, grade actuators is just awesome. The ones I bought are totally enclosed so no dust and chips can get into the bearings. The ball-nut-brack is milled into the table and the ball screws end support blocks are milled into the base. The motor is also fully enclose inside the base housing. It's very slick and very quiet when it moves.
Re: Mounting linear rails. How precise does it need to be precisely?
It can still be done without an accurately machined ledge to push the master rail up against (you do need a precision square though and straight edge though.) One way is to drill/tap 2 sets of holes along side the rail mounting holes, far enough out that flat head screws in the holes will just clear the rail. You then lightly mount the rail, put the flat heads into their holes and put a piece of round rod (could probably use steel tig wire,) between/under the flat heads and the rail. You can then torque down the flat heads to push the rail as necessary to make it straight (measured relative to your straight edge.) Lastly torque down the rail.
The method is similar to what is often used to push the rail up against it's reference edge, just you use the screws on both sides (and have to measure a ton.)
As for measuring, get a test indicator and good magnetic base (I'd splurge on the base before the indicator. Cheap bases will cause no end of headaches, cheap indicators are usually just less accurate than they say.)
As far as how precise you need to be, at least as precise as the datasheet for the rails says. Beyond that it's how precise you want your machine to be.
