Maybe I'm missing something, with all the negative comments on this, but your method seems quite cunning to me. Maybe people dont really get that you are clamping the block and hence measuring the error to the base including all the other errors and bits you can't correct or easily measure. I think this is a really neat way of measuring them. Really liking this little build series so far. Some really sensible decision making and working.
Going to build a similar machine. What are your plans for the spindle? I've found spindles with sub micron runout, but even the best ER20 collets still have 5 microns of runout. I've heard about hydraulic workholding options, but haven't found anything on eBay.
CNC_channel has good video about fixing runout ruclips.net/video/Xd8Ea7NRc4Q/видео.html at the moment i'm considering this spindle, check on google: JGL-80 2.2kw 30000rpm 500HZ iso20 Water Cooled. it has a tool changer and cost about 2k usd. i have to say that tool changer is very tempting :) not sure yet. maybe i just go with cheap chinese spindle and change tools by hand. i have one er13 not 20 but still.. it has 1 micron of runout. more than enough. work holding, hmm. so far i thought about vacuum plate and vices, and maybe palettes. but it will depend of wype of work i will be doing.
I'm not sure how important measuring twist and bow is because once the rail is bolted down to a flat surface, it'll straighten out. Of course there's also the opposite case,: say the rail has no bow or twist, so the builder bolts the rail to a granite surface that's twisted or bowed - same result. My point is that it's not so much about what the individual pieces do, but rather how the entire assembly measures. (An analogy is indexing an ER collet in its chuck. Both measured separately will have its own error, yet *as an assembly*, can be set up so that the overall error is less, or even zero.)
I am planning an epoxy granite CNC machine. It is my understanding, and I may be wrong, that the rail manufacturer expects that the rails will essentially be mounted to take up the accuracy of the underlying flat surface they are lying on. (?)
Why not use air bearings, like Dan Gelbart? You already have the granite base for it. And the bearings themselves shouldn't be that much more expensive than the linear rails...... And in theory could also be made by yourself. You should also consider using an air bearing spindle. Either one used from hard-disk test stands found on eBay, or some from jig grinders. Are you protecting your bearing blocks from getting grinding dust in the ball races when you are grinding on your rails like this? I have seen linear rail systems being ruined from not preventing debris from entering in to the ball circulating path. I do really like your ideas and skills in building this, and will continue on to watch the rest of this build series. Thanks for sharing!
What do you think about hydrostatic ways? I have had some experience with an old Brown and Sharpe Hydrotape that used them. Amazing machine! Reading up a bit on hydrostatic and air bearings, Google Books preview: books.google.com/books?id=9EKsLj-nj7UC&lpg=PA1&pg=PA1#v=onepage&q&f=false Wondering whether coolant would be good enough to use as the working fluid? That would simplify the design, to not have to keep the coolant separate from the ways oil. Just playing with ideas at this point, not saying it's practical.
i'm not that troubled if they are straight, but my problem was that they are different thickness in every possible way. extruded aluminium profile is more consistent than those rails...
@@PiotrFoxWysocki I understand why you use the grinding stone to level the rail. But why not fix the block on a surface grinder and push the rail under the grinding stone instead.
Hey Piotr, is that a fiber laser in the background? Can you give more information on your experience with owning one? Thinking of getting one myself. Thanks!
Amazing work sir. Diy builders usually dont understand just how hard it is to make machine that will work in range of 0.01mm on 300x300x300mm. Moving axis back and forth with guage on it, and showing how it each time comes to zero is far from actuall accuracy. You are on great way to make one of very few acccurate diy machines. I used epoxi granite fillament for my machine, but in future I would also like to use natural granite for some of the parts. I saw datron using granite for their machines, for tables. Could you give me some toughts on natural granite vs epoxi granite? Once again, absolutly great work!
i would like to thank you for understanding what i'm trying to do here :) I was afraid people wont get the idea why i'm doing it etc. granite vs epoxy.. i don't know where to even start :) I will have to make a separate video on this. but i will try to explain key points here. 1. you need something flat for the cast. long story short nothing is flat :) and even if it it is, it won't be because you have to put it on something. or it will have to be so thick that it won't deform under its own weight or added weight. best thing is a granite surface plate. 2. every resin shrinks. when it does it will deform your mold or if the mold is to ridgid it will separate. so even if you will cast on a perfectly flat surface it will came out deformed anyway. 3. when you use aggregate its impossible in home workshop to remove all the air bubbles. end result your cast looks like sponge. 4. its messy and more expensive than granite surface plate. plus takes way more time. those are the biggest problems to overcome. but almost everything is possible right? :) casting smaller parts is possible with vacuum pump and vacuum chamber. deformations shouldn't be as big and casting bigger parts,. but if you really want to cast bigger parts and make them as flat as a surface plate, you will need vacuum chamber with build in 100 ton press so you can remove bubbles and avoid deformations caused by shrinking. ( over the time it might change shape a bit anyway.) By choosing granite i tried to avoid all those variables, plus some i haven't even think of and i manage to make it faster than casting. plus i got a really good reference surface to start with. I hope i roughly answered your question.
Thank you for answer. When I asked about natural granite vs epoxi I was thinking more about differences between vibration dampening and strenght. Looks like epoxi granite dampens vibrations even a bit better than naturall granite. About cast, my idea is to sink into mold some metal parts which are later machined. One option is also grinding cast, but I havent seen anybody offer such process here in my country. I use epument from company rampf, you can check it here: www.rampf-gruppe.de/en/products-and-solutions/machine-systems/machine-beds/epument-mineral-casting/ They also offer grinding of casting etc.. But I dont want to send my casting to Germany and pay some crazy price for work. So I would rather sink in metal parts, so local CNC shop can mill them and grind them. You can check this guy: ruclips.net/video/bqGzbKFMhSg/видео.html You can check in teil 2 how he glued two blocks together. It could be also used for you, instead of drilling and bolting, maybe you could glue side blocks This is way I would go with, but with epoxi granite, not UHPC. I know UHPC is cheaper and easier to cast, but I already have about 2tons of epument. I like the way you fixed your rails, great process. Have you tought about investing in autocollimator? This way you could eliminate error from base plate. Do you use grade 00 granite plate?
@@PiotrFoxWysocki you make some excellent points, and I appreciate the opportunity to learn about the techniques you have developed! I especially appreciate the use of the granite blocks as inexpensive precise and rigid frame members. I am still at the thinking stage on a mill. I absolutely want to do better than MDF or plywood! Would a hybrid approach make sense? Use granite frame members plus some cast portions when embedded parts or complex shapes are desired.
@@scottwillis5434 in my example i wanted the most rigid machine i could make without spending a fortune on it. so it did it in the most optimal way i could came up with. because i haven't seen anyone sharing anything about such a build i thought i could document it if this is a good or bad idea. I guess you have to ask your self what will be the purpose for your future machine, than make design and material choices within given budget and manufacturing capabilities. In my opinion mdf or any form of wood is not the best material for cnc machine. but there is still plenty of wooden cnc machines anyway :)
@@PiotrFoxWysocki I want to be clear that my comments are not intended as criticism. I am very impressed by your ideas and design, and plan to incorporate many of them in my mill design. I am simply playing with ideas, thinking out loud, and offering ideas or brainstorms for discussion. These are not fully thought out, just ideas. One comment mentioned air bearings. I had some experience with a machine that used hydrostatic ways, and am exploring the idea of hydrostatic ways, perhaps using machining coolant as the working fluid. Not sure it would be practical. I enjoy learning about possibilities!
same here, but i haven't got much measuring tools. few calipers, one digital height gauge, one dial indicator, 2 small surface plates. with those things i can do all whan i need.
8:02 wow look whats end of linear rail. wow, but what you measuring is not even your rail could be 1 degrees tils if you look same point. are you try cut atom with that machine?
for the same price of a chinese rail and lead screw you can buy a pulled set from a pro shop. they have to replace their rails and screws so often and replace them long before they go bad to keep everything perfect they replace them at specific hours used instead of when its out of spec. and unless your running production steel milling they would still have a life expectancy of the life of your machine. this is the route i intend to take when i build my machine just because i could get better parts for a lower price. I dont trust the steel alloy properties being correct from Chinese parts with all the fraud arrests that went down a few years ago i know that low grade steel is still out on the market. and then theres the balls versus rollers for the linear bearings
high end production machines don't normally use this type of removable guide rail at all. You could certainly buy used rails from a Haas or similar, but they're not particularly good to begin with. Did you find parts for your machine?
5:36 if your rail would be 10 km circle you still see same result its straight enought. measuring it same point looks like straight even its full circle lol
it is not about measuring the straightness of the rails itself, but those notches on which it rols - you can have a perfectly straight rail but if those are cut as some wiggly curved path then straightness of the rail does not matter. and rail straightness itself - you can measure it on the table, but more important is the consistency of distance between the bottom and that sliding part top, as when you clamp rail down it will be strengthened by granite.
12:58 you get more error on 90 degree rails. how in earth you measure they are 90 degree 2 rails paraller you can measure but cross slide. its say 0 if you do it like that 1 point lol
This is all wrong. You are measuring based on the rail floating in space. When you tighten the rail down onto a surface it distorts. You need to measure the accuracy of the rail when it is tightened. Plus it it my understanding these rails are designed to absorb minor variations and still run straight. Hiwin rail documentation states the further out of tolerance the shorter the rail life not the less accurate. For example, What you are doing is similar to say measuring the height of a car when it is on a lift. When you put the car on the ground it changes. You need to measure in context. I.e. tightened down. A measurement out of context means nothing
its not wrong... this is the simplest way to measure rail vertical deviations. I guess my setup was't very obvious. to simplify it: i put a dial indicator base at the top of the carriage than i put the needle of dial indicator from the bottom of the rail just under the carriage. analogy - bottom of the rail is resting on the flat surface and carriage is riding on it with gantry attached or the table, etc. so i'm measuring the distance between top and the bottom. i will use your car analogy. im not measuring the car distance from the floor when its on the lift. im measuring from the top of the roof to the bottom of the wheels. it doesn't matter if the car is on the lift, sideways or upside down. the distance from the roof to the bottom of the wheels is the same. I guess i will have to make a separate video about it because somehow is not so obvious.
@@PiotrFoxWysocki Youre not understanding what im saying at all. the dimensions you are measuring change when you tighten the rails to its base and that is all that matters. your response to my car analogy shows you dont understand the problem. If a car is on a lift it will ABSOLUTELY measure different than on the ground when the car is put on the ground the springs compress, the tires load and sag, metal bends, gravity effects etc. I understand the car analogy is different than these rails im just trying to point out that you need to measure in the components final intended setting. In this video you show that the edges of the surface are higher than the center. How do you know this isnt intended to be that way? you can think measuring your parts to be high precision independently from each other is better but that does not mean they will be good at all when assembled and tightened. n real world applications everything moves and flexes depending on the forces that act upon it. This is important. What I am saying is the same as others have said in these comments. We understand what you are doing perfectly fine. I do not think you understand the concept we (acme GP and Chris Ranyer) are trying to tell you.
@@bobc.2810 if the thickness of the rail (top of the carriage to the bottom of the rail) on one end is 35.090mm than in the middle is 35.000mm and than at the other end is 35.070mm. and you bolt it to a flat surface. no matter how hard you will bolt it down. it still will be higher at the both ends, and lower in the middle. side walls in my base are with in 1 micron. so lets say the rail will confirm to the curvature of my 1micron-off surface. so the exaggerated hypothetical end result is: 35.091mm, 34.999mm 35.071mm. total variation of 92microns thank lord i got a replacement rails which are below 1 micron. so if the rail with carriage is 50mm. With both worst-case combined errors, will be: 50.001mm , 49.999mm , 50.001mm total variation of 2 microns i can check and correct if i can before i put it in. right?
You are literally working on a surface plate! Why no simply blue it up and reference the bottom surface of the rail of the surface plate? You would get it as flat as your surface plate is! You could even reference it directly off the mounting point!
The rails will be pulled straight when mounted to a flat surface. Meanwhile, the critical dimension is the top of the carriage relative to the bottom of the rail.
Без перевода понял что даже если сделать все плоскостя- то все равно точности не будет из за кривых рельс .Я в шоке .Они же денег стоят и все равно кривые.
The reason your rails are not measuring out perfect is simply this. The exact part is the ball bearing contact point - NOT TOP OF RAIL. Measure rails like this. Install them and tighten down to specific even torque - tightness. Install bearing block, next place measure tool on top of bearing block and measure the height of bearing block every 2 or 3 inches - 50 to 75mm. Now if there is error, shim up contact point of rail or grind rail bottom - not top. I am betting your "bearing block" top measurement will be vary good - much better then top of rail itself - bearings DO NOT RIDE ON TOP OF RAIL - say this ten times to yourself so it sinks in to your brain. - Also - important don't start with used junk rails. If you buy used rail first measure fit of bearing block. If there is too much movement return it - only buy rails from sellers that offer return. Bolt on a long bar that is strong to the top of each bearing. The bar should be as long as the rail. Place bearing on one end of rail with bar sticking out parallel to rail. Clamp rail down and with clamps on each side of bearing block soit cannot slide. do not clamp bearing block itself. Now clamp dial indicator on rail so it can measure movement at end of bar. - Next gently move bar end up - down and back and forth. Be aware of bar stiffness so you are not creating flex in bar. Move bar gently. How much movement is there in bearing fit to rails? If much then find better rail and bearing blocks or buy slightly larger balls for bearing block. Best bearing blocks have what is called pre-load. Do enough research to know about rails and bearing blocks so you know ahead of time what is the right ones and wrong ones to buy. Large cnc mills and lathes use these same rails and bearing blocks but they use type with preload so there is nil or almost no play. Then the design mill bed to take advantage of rail and block design so it places load properly. - Now I know this is lot of text but this is important info if you are building this machine. Here is very important detail for you. With this machine design there is really no need to attach granite side plates like is done in this design unless you wish to expand usable bed area. Simple as that?? yes simple as that - read it again and let this sink in to your brain. Place rails directly on top of plate along outer top edge. Want larger base and usable area for milling? buy larger granite plate. Next make sides and top out of granite for low twist and stiffness. - Another good hint for you? Don't use larger granite surface plate - use single thickness or maybe double thickness granite countertop glued and screwed together. Not as stiff?? Yes it will be stiff enough and lighter. What are you doing? Milling with 20 hp spindle at top speed and making airplane wings that are huge? No your not so be more reasonable and use counter top granite. Odds are you have counter top company local in your area? Just go to them and ask them to cut you any size base you need and maybe they can even drill it for you?? they have carbide and diamond drills so they can drill it easy and they might even like your idea very much and give you good deal and help you a lot. If you build it with counter top granite make a video and share it.
Maybe I'm missing something, with all the negative comments on this, but your method seems quite cunning to me. Maybe people dont really get that you are clamping the block and hence measuring the error to the base including all the other errors and bits you can't correct or easily measure. I think this is a really neat way of measuring them. Really liking this little build series so far. Some really sensible decision making and working.
Thank you ;) i thought it is obvious but perhaps i haven explained it well enough. But you got the idea anyway ;)
Amazing work man
thank you, i'm glad you like it :)
Most good way to fix these rails. Never expected that doing this is also quite important.
Going to build a similar machine. What are your plans for the spindle? I've found spindles with sub micron runout, but even the best ER20 collets still have 5 microns of runout. I've heard about hydraulic workholding options, but haven't found anything on eBay.
CNC_channel has good video about fixing runout ruclips.net/video/Xd8Ea7NRc4Q/видео.html
at the moment i'm considering this spindle, check on google: JGL-80 2.2kw 30000rpm 500HZ iso20 Water Cooled. it has a tool changer and cost about 2k usd. i have to say that tool changer is very tempting :) not sure yet. maybe i just go with cheap chinese spindle and change tools by hand. i have one er13 not 20 but still.. it has 1 micron of runout. more than enough.
work holding, hmm. so far i thought about vacuum plate and vices, and maybe palettes. but it will depend of wype of work i will be doing.
Awesome! Thanks for sharing!
pro tip: watch movies at Kaldrostream. Been using it for watching loads of movies recently.
quite smart about measuring with carriage on the rail stuff and that way figuring out the situation
I'm not sure how important measuring twist and bow is because once the rail is bolted down to a flat surface, it'll straighten out. Of course there's also the opposite case,: say the rail has no bow or twist, so the builder bolts the rail to a granite surface that's twisted or bowed - same result. My point is that it's not so much about what the individual pieces do, but rather how the entire assembly measures. (An analogy is indexing an ER collet in its chuck. Both measured separately will have its own error, yet *as an assembly*, can be set up so that the overall error is less, or even zero.)
I am planning an epoxy granite CNC machine. It is my understanding, and I may be wrong, that the rail manufacturer expects that the rails will essentially be mounted to take up the accuracy of the underlying flat surface they are lying on. (?)
Why not use air bearings, like Dan Gelbart? You already have the granite base for it. And the bearings themselves shouldn't be that much more expensive than the linear rails...... And in theory could also be made by yourself.
You should also consider using an air bearing spindle. Either one used from hard-disk test stands found on eBay, or some from jig grinders.
Are you protecting your bearing blocks from getting grinding dust in the ball races when you are grinding on your rails like this? I have seen linear rail systems being ruined from not preventing debris from entering in to the ball circulating path.
I do really like your ideas and skills in building this, and will continue on to watch the rest of this build series. Thanks for sharing!
What do you think about hydrostatic ways? I have had some experience with an old Brown and Sharpe Hydrotape that used them. Amazing machine!
Reading up a bit on hydrostatic and air bearings, Google Books preview: books.google.com/books?id=9EKsLj-nj7UC&lpg=PA1&pg=PA1#v=onepage&q&f=false
Wondering whether coolant would be good enough to use as the working fluid? That would simplify the design, to not have to keep the coolant separate from the ways oil.
Just playing with ideas at this point, not saying it's practical.
most rails wont be strait untill they are fitted to a way. You cant measure them lose like this. Even high end big 30mm thk rails wont be strait.
the slider should track equidistant though.
and the play from ball tracks
it can be adjusted while fixing. the screw torq should solve this ,
i'm not that troubled if they are straight, but my problem was that they are different thickness in every possible way. extruded aluminium profile is more consistent than those rails...
@@PiotrFoxWysocki I understand why you use the grinding stone to level the rail.
But why not fix the block on a surface grinder and push the rail under the grinding stone instead.
Hey Piotr, is that a fiber laser in the background? Can you give more information on your experience with owning one? Thinking of getting one myself. Thanks!
This is a CO2 laser not fibre.
Amazing work sir. Diy builders usually dont understand just how hard it is to make machine that will work in range of 0.01mm on 300x300x300mm. Moving axis back and forth with guage on it, and showing how it each time comes to zero is far from actuall accuracy. You are on great way to make one of very few acccurate diy machines.
I used epoxi granite fillament for my machine, but in future I would also like to use natural granite for some of the parts. I saw datron using granite for their machines, for tables.
Could you give me some toughts on natural granite vs epoxi granite? Once again, absolutly great work!
i would like to thank you for understanding what i'm trying to do here :) I was afraid people wont get the idea why i'm doing it etc.
granite vs epoxy.. i don't know where to even start :) I will have to make a separate video on this. but i will try to explain key points here.
1. you need something flat for the cast. long story short nothing is flat :) and even if it it is, it won't be because you have to put it on something. or it will have to be so thick that it won't deform under its own weight or added weight. best thing is a granite surface plate.
2. every resin shrinks. when it does it will deform your mold or if the mold is to ridgid it will separate. so even if you will cast on a perfectly flat surface it will came out deformed anyway.
3. when you use aggregate its impossible in home workshop to remove all the air bubbles. end result your cast looks like sponge.
4. its messy and more expensive than granite surface plate. plus takes way more time.
those are the biggest problems to overcome. but almost everything is possible right? :)
casting smaller parts is possible with vacuum pump and vacuum chamber. deformations shouldn't be as big and casting bigger parts,.
but if you really want to cast bigger parts and make them as flat as a surface plate, you will need vacuum chamber with build in 100 ton press so you can remove bubbles and avoid deformations caused by shrinking. ( over the time it might change shape a bit anyway.)
By choosing granite i tried to avoid all those variables, plus some i haven't even think of and i manage to make it faster than casting. plus i got a really good reference surface to start with.
I hope i roughly answered your question.
Thank you for answer. When I asked about natural granite vs epoxi I was thinking more about differences between vibration dampening and strenght. Looks like epoxi granite dampens vibrations even a bit better than naturall granite. About cast, my idea is to sink into mold some metal parts which are later machined. One option is also grinding cast, but I havent seen anybody offer such process here in my country.
I use epument from company rampf, you can check it here: www.rampf-gruppe.de/en/products-and-solutions/machine-systems/machine-beds/epument-mineral-casting/
They also offer grinding of casting etc.. But I dont want to send my casting to Germany and pay some crazy price for work. So I would rather sink in metal parts, so local CNC shop can mill them and grind them. You can check this guy:
ruclips.net/video/bqGzbKFMhSg/видео.html
You can check in teil 2 how he glued two blocks together. It could be also used for you, instead of drilling and bolting, maybe you could glue side blocks
This is way I would go with, but with epoxi granite, not UHPC. I know UHPC is cheaper and easier to cast, but I already have about 2tons of epument.
I like the way you fixed your rails, great process. Have you tought about investing in autocollimator? This way you could eliminate error from base plate.
Do you use grade 00 granite plate?
@@PiotrFoxWysocki you make some excellent points, and I appreciate the opportunity to learn about the techniques you have developed! I especially appreciate the use of the granite blocks as inexpensive precise and rigid frame members.
I am still at the thinking stage on a mill. I absolutely want to do better than MDF or plywood!
Would a hybrid approach make sense? Use granite frame members plus some cast portions when embedded parts or complex shapes are desired.
@@scottwillis5434 in my example i wanted the most rigid machine i could make without spending a fortune on it. so it did it in the most optimal way i could came up with. because i haven't seen anyone sharing anything about such a build i thought i could document it if this is a good or bad idea.
I guess you have to ask your self what will be the purpose for your future machine, than make design and material choices within given budget and manufacturing capabilities.
In my opinion mdf or any form of wood is not the best material for cnc machine. but there is still plenty of wooden cnc machines anyway :)
@@PiotrFoxWysocki I want to be clear that my comments are not intended as criticism. I am very impressed by your ideas and design, and plan to incorporate many of them in my mill design.
I am simply playing with ideas, thinking out loud, and offering ideas or brainstorms for discussion. These are not fully thought out, just ideas.
One comment mentioned air bearings. I had some experience with a machine that used hydrostatic ways, and am exploring the idea of hydrostatic ways, perhaps using machining coolant as the working fluid. Not sure it would be practical. I enjoy learning about possibilities!
What brand rails are these?
those are brand-less from ebay. I wasn't prepared for how bad they where.
@@PiotrFoxWysocki Thanks for the reply. I wonder how they compare to Hiwin rails.
what in the world made you to expect that those rails to be accurate in the sub micron range?
Hope i guess ;)
Can I ask what size rails those are?
20mm
How much does this weigh? Im curious LOL
Around 300kg
turns out i spend as much on measuring tools as end mills, so much for plug and play.
same here, but i haven't got much measuring tools. few calipers, one digital height gauge, one dial indicator, 2 small surface plates. with those things i can do all whan i need.
8:02 wow look whats end of linear rail. wow, but what you measuring is not even your rail could be 1 degrees tils if you look same point. are you try cut atom with that machine?
for the same price of a chinese rail and lead screw you can buy a pulled set from a pro shop. they have to replace their rails and screws so often and replace them long before they go bad to keep everything perfect they replace them at specific hours used instead of when its out of spec. and unless your running production steel milling they would still have a life expectancy of the life of your machine. this is the route i intend to take when i build my machine just because i could get better parts for a lower price. I dont trust the steel alloy properties being correct from Chinese parts with all the fraud arrests that went down a few years ago i know that low grade steel is still out on the market. and then theres the balls versus rollers for the linear bearings
high end production machines don't normally use this type of removable guide rail at all. You could certainly buy used rails from a Haas or similar, but they're not particularly good to begin with.
Did you find parts for your machine?
5:36 if your rail would be 10 km circle you still see same result its straight enought. measuring it same point looks like straight even its full circle lol
it is not about measuring the straightness of the rails itself, but those notches on which it rols - you can have a perfectly straight rail but if those are cut as some wiggly curved path then straightness of the rail does not matter.
and rail straightness itself - you can measure it on the table, but more important is the consistency of distance between the bottom and that sliding part top, as when you clamp rail down it will be strengthened by granite.
ahem, buy nsk linear rails no regrets.
12:58 you get more error on 90 degree rails. how in earth you measure they are 90 degree 2 rails paraller you can measure but cross slide. its say 0 if you do it like that 1 point lol
This is all wrong. You are measuring based on the rail floating in space. When you tighten the rail down onto a surface it distorts. You need to measure the accuracy of the rail when it is tightened. Plus it it my understanding these rails are designed to absorb minor variations and still run straight. Hiwin rail documentation states the further out of tolerance the shorter the rail life not the less accurate.
For example, What you are doing is similar to say measuring the height of a car when it is on a lift. When you put the car on the ground it changes.
You need to measure in context. I.e. tightened down. A measurement out of context means nothing
its not wrong... this is the simplest way to measure rail vertical deviations.
I guess my setup was't very obvious.
to simplify it:
i put a dial indicator base at the top of the carriage
than i put the needle of dial indicator from the bottom of the rail just under the carriage.
analogy - bottom of the rail is resting on the flat surface and carriage is riding on it with gantry attached or the table, etc.
so i'm measuring the distance between top and the bottom.
i will use your car analogy.
im not measuring the car distance from the floor when its on the lift.
im measuring from the top of the roof to the bottom of the wheels.
it doesn't matter if the car is on the lift, sideways or upside down. the distance from the roof to the bottom of the wheels is the same.
I guess i will have to make a separate video about it because somehow is not so obvious.
@@PiotrFoxWysocki Youre not understanding what im saying at all. the dimensions you are measuring change when you tighten the rails to its base and that is all that matters.
your response to my car analogy shows you dont understand the problem. If a car is on a lift it will ABSOLUTELY measure different than on the ground when the car is put on the ground the springs compress, the tires load and sag, metal bends, gravity effects etc. I understand the car analogy is different than these rails im just trying to point out that you need to measure in the components final intended setting.
In this video you show that the edges of the surface are higher than the center. How do you know this isnt intended to be that way? you can think measuring your parts to be high precision independently from each other is better but that does not mean they will be good at all when assembled and tightened. n real world applications everything moves and flexes depending on the forces that act upon it. This is important.
What I am saying is the same as others have said in these comments. We understand what you are doing perfectly fine. I do not think you understand the concept we (acme GP and Chris Ranyer) are trying to tell you.
@@bobc.2810 if the thickness of the rail (top of the carriage to the bottom of the rail) on one end is 35.090mm than in the middle is 35.000mm and than at the other end is 35.070mm. and you bolt it to a flat surface. no matter how hard you will bolt it down. it still will be higher at the both ends, and lower in the middle.
side walls in my base are with in 1 micron. so lets say the rail will confirm to the curvature of my 1micron-off surface.
so the exaggerated hypothetical end result is:
35.091mm, 34.999mm 35.071mm.
total variation of 92microns
thank lord i got a replacement rails which are below 1 micron.
so if the rail with carriage is 50mm.
With both worst-case combined errors, will be:
50.001mm , 49.999mm , 50.001mm
total variation of 2 microns
i can check and correct if i can before i put it in. right?
You are literally working on a surface plate! Why no simply blue it up and reference the bottom surface of the rail of the surface plate? You would get it as flat as your surface plate is! You could even reference it directly off the mounting point!
The rails will be pulled straight when mounted to a flat surface. Meanwhile, the critical dimension is the top of the carriage relative to the bottom of the rail.
Без перевода понял что даже если сделать все плоскостя- то все равно точности не будет из за кривых рельс .Я в шоке .Они же денег стоят и все равно кривые.
The reason your rails are not measuring out perfect is simply this. The exact part is the ball bearing contact point - NOT TOP OF RAIL. Measure rails like this. Install them and tighten down to specific even torque - tightness. Install bearing block, next place measure tool on top of bearing block and measure the height of bearing block every 2 or 3 inches - 50 to 75mm. Now if there is error, shim up contact point of rail or grind rail bottom - not top. I am betting your "bearing block" top measurement will be vary good - much better then top of rail itself - bearings DO NOT RIDE ON TOP OF RAIL - say this ten times to yourself so it sinks in to your brain.
-
Also - important don't start with used junk rails. If you buy used rail first measure fit of bearing block. If there is too much movement return it - only buy rails from sellers that offer return. Bolt on a long bar that is strong to the top of each bearing. The bar should be as long as the rail. Place bearing on one end of rail with bar sticking out parallel to rail. Clamp rail down and with clamps on each side of bearing block soit cannot slide. do not clamp bearing block itself. Now clamp dial indicator on rail so it can measure movement at end of bar.
-
Next gently move bar end up - down and back and forth. Be aware of bar stiffness so you are not creating flex in bar. Move bar gently. How much movement is there in bearing fit to rails? If much then find better rail and bearing blocks or buy slightly larger balls for bearing block. Best bearing blocks have what is called pre-load. Do enough research to know about rails and bearing blocks so you know ahead of time what is the right ones and wrong ones to buy. Large cnc mills and lathes use these same rails and bearing blocks but they use type with preload so there is nil or almost no play. Then the design mill bed to take advantage of rail and block design so it places load properly.
-
Now I know this is lot of text but this is important info if you are building this machine. Here is very important detail for you. With this machine design there is really no need to attach granite side plates like is done in this design unless you wish to expand usable bed area. Simple as that?? yes simple as that - read it again and let this sink in to your brain. Place rails directly on top of plate along outer top edge. Want larger base and usable area for milling? buy larger granite plate. Next make sides and top out of granite for low twist and stiffness.
-
Another good hint for you? Don't use larger granite surface plate - use single thickness or maybe double thickness granite countertop glued and screwed together. Not as stiff?? Yes it will be stiff enough and lighter. What are you doing? Milling with 20 hp spindle at top speed and making airplane wings that are huge? No your not so be more reasonable and use counter top granite. Odds are you have counter top company local in your area? Just go to them and ask them to cut you any size base you need and maybe they can even drill it for you?? they have carbide and diamond drills so they can drill it easy and they might even like your idea very much and give you good deal and help you a lot. If you build it with counter top granite make a video and share it.