Printable View
Is aluminium/steel box section straight enough to mount a supported linear rail on opposite sides and expect it to run smoothly without binding.
Here's an example to clarify. I'm thinking of putting a piece of 100x50 box section between those two horizontal rails.
Attachment 2837
Depends on the spec... and how it was manufactured.. Extruded may well be, welded steel box section maybe not, might need a pass with a mill..
'Pass with a mill'
Skim it with a fly cutter on both sides? What about the internal stresses in the material, is that going to be an issue?
[Shame they sold the surface grinder at school :(]
surface grinder would be better still...
I've not had much experience of large 'beam' of that material... JohnS might be able to give a more informed opinion
Assuming the rail is straight and if it's only a few tenths here or there you could use a thin layer of epoxy.
Extrusions are very accurate, either profile Ali or RHS Ali section. Bit more expensive than steel, but alot of steel is ERW (electric resistance welded) with a seam down one side. I remember reading a post where they wished they'd skimmed the steel section before mounting the rails. Factor in the cost of getting two faces ground and the Ali profile is your best bet.
By the way, in your sketch showing the box section assembly wrapped around the gantry this could probably do with a slight re-arrangement. As drawn the top and bottom plates fit on top of the sides. This means that the sides have to be very accurately cut to get a nice just-right preload on the bearings.
If you make the sides longer, and have the top and bottom plates inset, then the play in the bolt mounting holes (~1mm in total up and down) which are now in the sides and not the top will allow you to put it together better, and set the system up with just the right load. Otherwise you risk either squashing the bearings or pulling them apart.
Thanks for that. I reckon I'll go for 80x80x3mm aluminium box section then. There's someone selling it near me for £20 per meter. That does make the centre distance of the Y-rails only 134mm - is that OK?.
Excellent point about the bearing preload - that's probably saved me a lot of hassle! I'll change the drawing now.
The machine I got from ebay has, from memory, 90mm Rhs, a 9mm wall and its welded (erw) and that has been milled for the rail mounting surface (I guess thats why they used 9mm to allow for the removed steel) this has only been done on one face so internal stress wouldnt have been a problem.
Ali is also lighter, dosnt rust (not like steel anyway) and dosnt need painting :smile:
Are you using the rails back to back on the y axis to counter the uneven load distribution? (ie 25% reduction at the open side) if so have you looked at doing the same with the zaxis? if you mounted them on a c channel (or box with a slot cut out, then you would have space for the Drive screw as well)
Just a thought and is some thing I was looking at a while ago for an Ali box and supported round rail machine. Not enough hours in the day..........
True, but it's not as strong and more easily transmits vibrations since it's less dense...Quote:
Originally Posted by Ross77
Exactly.Quote:
Originally Posted by Ross77
I looked at it, but didn't think of using C-channel which would make it quite simple.Quote:
Originally Posted by Ross77
The place near me has some 3"x1"x0.25" aluminium angle. There's also U, 2" x 3" x 2" ... only 3/16" thick though. Time for some more drawing / calculation!
Personally for moving gantry with steppers I would go for Ali. Fixed gantry then heavy duty steel box...You pays your money you takes your chance.....Quote:
True, but it's not as strong and more easily transmits vibrations since it's less dense...
You can design for extra strength and reduce the vibrations in the first place and it dosnt become such an issue. (I'm sure someone will disagree with that.:naughty:)
This is what I was looking at, the idea was that the two horizontal box sections top and bottom with the c channel and vertical box in the middle would provide a pseudo I beam but with improved torsion due the central box, the c channel would house the Ball screw and there fore keep the z axis in tight to the gantry
The idea for the z axis was similar in that the spindle and hence cutting forces were in close to the gantry and the back to back rails to give an even load distribution. I figured that a wider z carriage was acceptable as the increase in rails needed is marginal compared to the bearings.
Anyway thats as far as i got and I know there are many improvements but hope it of some use or sparks another idea
Wow' thanks for the drawings! That looks like a good way of going about things - much less expensive aluminium plate which is a bonus.
I've been trying to get the rails to fit on 3"x1" angle mounted too the plate...but it's just generally not working. I'll post more when I've got a drawing done.
I'd also been playing around with a few ideas recently. Some of them are very similar to yours Ross!
First off, tried to get Z axis as close to gantry as possible (using profile rail, which I'm fortunate to have) and came up with the scheme below (LH image).
Also, playing with gantry sides idea using RHS, with an offcut from the gantry as a spacer to get the loads well into the RHS and avoid just locally distorting the wall (centre and RH image). Lower joint down to the X bearing is always an area for many ideas - thought using these bars each side were a bit neater than the usual double back to back plate arrangement.
Option of TIG-ing a closer plate over the sloping part of the RHS, but probably not worth the effort.
p.s. Looked at some of the steel box and RHS in the stores, and the ERW stuff would need grinding. The sides tend to cave in slightly relative to the corners.
I like the idea of rhs with the spacer, nice and neat, you could probably get away with just gluing a closer stip over the angle to keep dirt and chips out.
Where would the ball screws go on the first one? with those rails could you wrap the carridge around the beam and mount the srews and motors on the back or would the leverage be all wrong? not pull ing from the centroid etc...
I guess there are no new designs its all been done before and when you start leaning out or designing close to the wire you end up at the same point any way. I'm sure I ve seen elements of my design else where and thought if i just change that a bit and move that over there it will work how i want it too. :smile:Quote:
Some of them are very similar to yours Ross!
Well, I tried 'borrowing' your design Ross but couldn't get it narrow enough for my requirements. Great ideas though.
This is what I ended up with:
Attachment 2867
Here's my current design. Just trying to work out how to fix the 4 blocks to the aluminium plate that slides on the Y axis. There's not much room and it's made more difficult because the bolts thread into the bearing blocks...I don't really want to start drilling those out.
Attachment 2859
Attachment 2860
One plus of having the Z rails this way round is that that aluminium angle will stiffen up the 20mm aluminium plate significantly. Maybe I can replace the 600x160x20mm plate with 15mm to save a few ££'s?
Right, hopefully this will be the final final design!:
Attachment 2861
I've rendered it with the Y axis off the blocks on the Z as otherwise you wouldn't see much.
Have I done anything silly...does it look OK?
The mass of all that is 25.9Kg, that's not including the stepper motor, screws, bearings and spindle.:thumbdown:
Some good development and ideas, but where were you thinking of putting the Z axis leadscrew? Looks like there's no space now.
I wonder if you just turned the Z axis rails through 90deg and mounted the support rail onto the router plate, and mounted the blocks back onto the main Y axis carriage, you would be better off. Stiffness wise it must be similar, and you are no further away in terms of spacing from the gantry than the above dwg (judging by the dwg above). You could even include the angle pieces in there as well if you want.
I forgot to put a hole in the blocks to fit it, it fits now:Quote:
Originally Posted by routercnc
Attachment 2866
Are you describing what I drew in the first post on this thread, or something different? I rotated the rails to even up the load on the bearings. Do you think this isn't an issue? If so I'll quite likely go back to the previous version as it's easier.Quote:
Originally Posted by routercnc
If I understand the dwg in post#17, then the router plate mounting footprint onto the Y axis carraige is now very narrow. I assume the blocks with the hole in for the ballscrew are the mounting feature into the bearings and into the Y axis carraige.
If this is the case I personally would go back to you original drawing. Ross can make this work, but your geometry requirement is obviously a bit different. Don't think the bearing load would be a problem for the original dwg, especially compared to the compromise on mounting in post#17. If you keep the angle piece you can always add these to the router plate if required. Any further views on that Ross?
The supported rails do have bulk, more so than profile, so it looks like you've done all you can.
[edit: Wonder if you should allow for (i.e. drill the holes for) a second pair of bearing blocks at the lower Z support. Because of your long Z axis reach the loads will be high at the bottom bearing, trying to open it up on the post#1 dwg. Another one next to it would help. Bit like some heavy doors have 3 hinges, with 2 close together. Never tried 3 in a row - don't think there would be a tolerance problem unless someone knows different. ]
Narrow mounting, yes I had just been thinking that. I'll wait and see what Ross thinks. I could separate them a bit more if I make the axis wider but I don't really want to sacrifice the Y travel. At the moment the travels are 1700x740x400 (XYZ).
Adding the L section to the previous design is perfectly possible so I may well do that. Do you think if I did I could use 15mm aluminium plate instead of 20mm for the bit they mount on?
I just based mine on ease of construction and cheap Ali box. Trying to keep it simple to build is also worth concidering. I also based the bearing strength from this thread I posted earlier in the yearQuote:
Ross can make this work, but your geometry requirement is obviously a bit different. Don't think the bearing load would be a problem for the original dwg
http://www.mycncuk.com/forums/showth...supported-rail
Dont forget that I was aiming to get the most out of the rails and therefore designing for the smallest possible size, again to save cost and weight. But if your 2 rails and blocks can still handle the load at 25% capacity then its not worth worrying about and save your y travel
It dose look a bit narrow to me tho. dont forget that making it wider will reduce the torque arm caused by the long z axis. (fun this design stuff eh....)
They should do, they are rated for 862N/1372N which is well above the cutting forces. I'm not sure about the torque ratings of the bearings though which is probably more of an issue.Quote:
Originally Posted by Ross77
It's 160mm wide and the centre distance of the rails is 100mm. I don't really understand which torque you're referring to that could make a difference to the Z rails?Quote:
Originally Posted by Ross77
From www.clickmetal.co.uk the 20mm aluminium plate (1220x160x20 total) will cost £88 inc. VAT and postage. I could get a 2 meter length of 6"x0.75" flat from www.aluminiumwarehouse.co.uk for £93 posted....I can't find anywhere nearby to get it :(
Ok i will try to explain, but im not the best. Maybe some else can re-phrase my ramblings......Quote:
They should do, they are rated for 862N/1372N which is well above the cutting forces. I'm not sure about the torque ratings of the bearings though which is probably more of an issue.
Basicly you have a force at the cutter that is then transmited over a distance to the bearings (force x distance = N/mm) if nothing is able to resis this force then it will rotate about the centre point. In this case it will be in the middle of the 4 bearings, so yuo will have 50mm from the centre line to the bearings.
So say you have 10N cutting force over the 350mm (for aguments sake) thats 3500 N/mm, now as all forces have to equal and opposite (one of Newtons laws.....) the bearings will have to resist the 3500N/mm but only over 50mm so the bearing wil see 3500/50=70N
So you see that the actual force on the bearings isnt just the cutting force and in this case is 7x the cutting force, although divide that by 4 and each bearing will onlys see 17.5 N from the cuting force
Its the centre line of the rails that counts, as far as the lever arm is concerned. theres also the fact that a long thin beam will more suseptable to vibration........( but that a whole over can of worms):naughty:Quote:
It's 160mm wide and the centre distance of the rails is 100mm. I don't really understand which torque you're referring to that could make a difference to the Z rails?
The danger is that this calc only deals with cuting force and on the face of it makes the bearings look overkill. In reality there all sorts of other forces that need to be allowed for. but rather than spend the rest of you life trying to calculate them just make the machine as ridgid as can and then cross your fingers :whistling:
Ahh thanks for the description. I was thinking in the wrong plane...now it makes sense!
I'll put the rails right on the edge of the plate since that gains about 10mm.
Yes, I'll make it as it is now. I'm currently waiting on my Mum/Dad to give me permission to buy the bearings. Then I hope to find somewhere better to buy the aluminium plate!Quote:
rather than spend the rest of you life trying to calculate them just make the machine as ridgid as can and then cross your fingers
I'm intending to have this router finished before I start University, so 21-09-10.
Can I be picky and say torque is measured in Nmm, not N/mm. That's stiffness.:tup:
Haha, I thought that just wasn't going to mention it. Strictly speaking isn't it N·m! But not a Joule...Quote:
Originally Posted by routercnc
Duh me bad :redface: only fair since its normally me who's picky :smile: still if thats all I got wrong I'm happy, thought it was going to be ripped apart.......Quote:
Can I be picky and say torque is measured in Nmm, not N/mm. That's stiffness.:tup:
you should have, I don't mind, everyone makes mistakes and you never know everything, I'm by no means perfect and isn't that the point of the forum that mistakes get corrected?:naughty:Quote:
Haha, I thought that just wasn't going to mention it. Strictly speaking isn't it N·m! But not a Joule...
I don't think it has to be Nm tho (dynamics maybe), as manufactures rearrange units to fit tables or make them look better with big numbers, I used mm to keep the units the same as you quoted, as long as you keep then the same through out the calc it doesn't matter as pointed out its force x dist (Nm) and then divide by dist to get a force (N).:tongue:
What are you studying at Uni?
Electronic Engineering at Nottingham - purely because I've always been interested in electronics and that sort of thing. I'm hoping they'll let me use the workshops there since it'll be a bit disappointing not having access to big milling machine/lathes after having had access at school for several years. I was in the workshop every lunchtime/free period in the 6th form.Quote:
Originally Posted by Ross77
Good course... I did Electrical Engineering with Electronics at Imperial many many moons ago now ... you have to get to know the labtechs in the mech eng dept :lol:
On your original question of straightness of steel box section, this guy went with a steel gantry and got the THK rail mounting surfaces ground. Cost him $200.
http://www.woodsolutions.com/diy-cnc.htm
Also of interest is the spindle setup (one for Ross and Steve), and the unusual X axis bearing arrangment.
A quick question if I may.
http://www.valuframe.co.uk/Technical-details.html
The Valuframe profiles have a 0.1mm camber either side of the profile slot. To my mind, mounting a profiled rail to this slot slightly to the right or left of centre may result in the rail ending up rotated with respect to the plane or possibly twisting slightly along its length.
A) Does this in fact happen?
B) Is it a significant effect?
Something to be aware of, I was using 6mm wall Al box section from Aluminiumwarehouse.co.uk. It was pretty straight, but slightly twisted. so that in 1m, on a 50mm box one side was about 0.5mm off the supported round rail. At the moment I have just built the thing, which I will probably regret later, though it is not the worst issue with the machine.
Thanks for that - definitely something to keep in mind. I'm currently intending to use 80mm, 3mm wall, box section. Maybe that's not the best of ideas.Quote:
Originally Posted by daveshorts
Thinking about it if the box section is twisted then with a rail on either side they will still be the same distance apart so should run smoothly. In an extreme case I suppose it would mean the Z axis wouldn't remain perpendicular to the bed and would get slightly translated at the cutter tip - pretty negligible for 0.5mm though I think?
The number of times I change my mind on this because of all the advica I get/ ideas I have is getting ridiculous!
My impression is that the rails will resist torsion better than the box section anyway. I certainly don't have any issues with binding with supported round rail.
im thinking that it might be a bit of a pain if your using 15mm rails and a little less so with 25mm rails... either way make sure you buy the groved slot nuts and your rail should sit pretty close to central then do your best to centralise before swinging on the cap heads ... the camber is suposed to act as a locking device and will pull out under tension so if your on 25mm rails there shouldnt be any twist at allQuote:
To my mind, mounting a profiled rail to this slot slightly to the right or left of centre may result in the rail ending up rotated with respect to the plane or possibly twisting slightly along its length.
I don't pretend to understand much of this but precision aluminium profile (and box? it is extruded) apparently conforms to DIN 17615:
DIN 17615
Outer Dimensions: Depending on size, 0.2 to 0.4 mm
Straightness Deviation: Max. 1.5mm/2m
Flatness Deviation: Max. 1.5mm/2m
http://www.maytec.org/technische_daten_e.htm
and
http://www.drivesystemsgroup.com/PDF...ue%20(DSG).pdf
also this is an interesting read, or would be if it wasn't so late
http://www.easy-conveyors.fi/materiaali/NTS_GB2002.pdf
Some interesting reading there certainly...
Somewhere it says 'For one manufacturing set-up the variation within one profile is 0.0004 inches'. That's roughly 0.01mm which is negligible. So even if the 80mm dimension is +-0.45mm as it says in your first link, the actual dimension should be stable.
That's where I became totally confused Jonathan
if the 80mm dimension is +-0.45mm as it says in the first link and the actual dimension varies by about 0.01 then where do the straightness and flatness deviations per meter fit in? Also the 1.5mm/2m is not necessarily a gradual change along the full length, it could be bumps or waves or??? I don't know.
Brain is starting to fry, again!
This guy had problems with the concave surface of his ali profiles...
http://www.cnczone.com/forums/showpo...7&postcount=33
... his fix was a mix of long t-nuts and epoxy to flatten the face, but the area he was dealing with was quite small.
This got me thinking, could epoxy be used to to flatten all mounting surfaces on steel box and smooth over the dimension issues with ali profile? The above thread makes for good reading and at some point leads to this build...
http://oneoceankayaks.com/madvac/madvac_index.htm
This epoxy is amazing stuff, the two linked builds make heavy use of it to ensure the joins in the steel frames are true.
So, with enough epoxy, a precision level and a thick sheet of glass do you think it possible to true ali profile? The end result must be better than building with 1mm irregularities?
Two big problems:
1) The cost of ali epoxy, http://uk.farnell.com/araldite/arald...0ml/dp/1759545, need to add in the cost of a dispensing gun too, I think.
2) The fact that even with surface prep the special epoxy is not a guaranteed strong, long lasting join.
Since all I'd be using the epoxy for is to make the filling between two lumps of metal is it necessary to use ali epoxy? And indeed, am I obsessing over trivial detail ?
Not got time to look at those links now - about to go into school to make a few bits I can't do at home :)
Try this site for epoxy:
http://www.ecfibreglasssupplies.co.u...k/default.aspx
Maybe reinforce it with some of there glass/carbon fibre?
I've sortof assembled mine with just the box section as it is...and it seems ok. I'll have it assembled properly later on today, then we'll see!
Hopefully you shouldn't need much epoxy, where you need to fill anything over a few tenths of a mm you could put glass fibre cloth in the space. It will be better to use a thixotropic or fairly viscous variant than the west system stuff which is designed to wet out composites and is a pretty thin liquid. I don't think you need to worry too much about bonding if it's trapped in a clamped interface - using epoxy to actually make the join is another matter and you would need very good surface prep and glue / material selection in that case.Quote:
Originally Posted by ecat
I've a length of 3" square 1/4" wall aluminium that I'm intending making my router out of but I haven't had chance to do any measuring on it. I'm hoping it will be close enough that I can just use a thin layer of epoxy on the interfaces and by initially clamping the rails down very lightly until the epoxy sets ensure that they are not deformed by the box section. It'll be at least a couple of months before I get to that stage so I'm looking forward to reading how you get on :smile: