Aluminum Flat Warpage

[bassman2914]

I've been looking at SpeedyMetals.com for finding the price for the aluminum for building my cnc router. I'm going to be using a lot of aluminum and am aiming for a very precise machine with hopefully better than .001" positional accuracy over a 15" span. To do this I need to accurately mount my linear rails meaning I need to mill pockets in the aluminum it is mounted to or machine flat a piece of aluminum and mount the flat end right next to the rails which will straighten the aluminum and the rails(if thick enough). My problem is my last email to this company said I shouldn't do that since it will make the aluminum warp making me lose my straightness tolerances. I've been looking at the 6061-t6511 extruded "flat" aluminum sheets they have since it is economical with wide ranges in sizes. Now my question is would I have a problem with warping if I have to drill and tap holes(could be any of the sides) or if I have to machine off .005-.01" on a .5x3x24" piece of stock on the .5x24" side? looking online i have read that it's not worth worrying unless it's really thin and it may warp in the direction you'd expect to warp(the way a piece of cardboard would bend). I've read about so many problems with warping metal i don't know if this is true. If anyone has experience with extruded aluminum(closer to 6061-t6511 the better) or better yet if you've bought from speedymetals. I have extruded aluminum pieces from home depot and drilling holes and cutting metal off one side doesn't seem to do anything(that's with really thin stuff too).

[jsheerin]

Not really answering your question, but you might look into using mic-6 tooling plate.

[keebler303]

MIC-6 would be the better, more expensive option.

I have worked with 6061T6 a lot in 1/4" thickness. Up to around 2" width, you don't really notice the warp. I have worked with up to 8" width and it comes from the factory with a good 1/16" or more warp. Most of my stuff has been profile cutting but if you take some off one side it will move around a bit. I think as long as you maintain sufficient thicness, it will not be a large problem. Just keep an eye on it as you move along.

Matt

[wildwestpat]

IMHO you have the tolerance on the thickness but I do not see any reference to flatness in the Speedy Metals specs on extruded plate. In general extrusion and warping go together and the warping when you cut bits from a bigger sheet get worse. Drilling the odd hole is normally OK but cutting long slots is asking for movement. Annealing helps relieve the internal stresses due to the extrusion or rolling process. IMOH rolled plate would be better suited to your application.

Try and use ground stock and resist cutting long slots OR get the bed surface machined after you have finished shaping it. Although 1/2" thick sounds solid it will bend so you need to work out by how much as 0.001 as a total tolerance has to be distributed between a number of sources of inaccuracy.

Regardds - Pat

[RICHARD ZASTROW]

I vote with jsheerin. Tooling plate is cast and doesn't move around (hence tooling plate) like rolled (wrought) bars and sheets. The con part: it doesn't weld that well, but aluminum welding is a PITA anyway.LOL

Dick Z

[bassman2914]

So it sounds like it may be possible. I don't care too much about how straight it is when it comes as long as it's good enough to where I can make it straight. To give an idea of the parts the parts I'm having to mill will be around .5x4x24 and .375(.5 if needed)x3x24 and I will be milling the .5x24" and .375x 24" sides meaning if the side i was surfacing was the top then the piece would be .5(or .375)inches wide, 24inches long and 4(or 3) inches thick. Now i could see how milling the large face would cause warping since it's .5 inches thick but milling the side means it will be essentially 4"thick which seems like it wouldn't bend. I also had and idea of maybe surfacing very lightly one side, flipping it and doing the other side(while measuring flatness each time) until it stops changing or until one side is flat. This way it won't warp so much that i'm just cutting the ends or the center on each successive pass. I could also just surface a couple inches at a time and flip this way there wouldn't be any warping since the parts the amount that changed is very small(i think). It would take a while and require a finishing pass but you got to do what you got to do. Does anyone have an idea of how deep i'd have to go before i'd get past the stressed outer layer? Again, the best option is stress relieved aluminum so it's not too expensive but will be stay put.(if anyone knows where to get some) Would 1/4 holes for bolts every 4" on the side(the .5x24 side) cause significant warping(or similar ratios of bolts about half the thickness of the material) if they were about 5/8" deep(they're only disrupting one side instead of equally on both).

[jsheerin]

Speedy metals sells cast aluminum tooling plate. It's under 'plate'. I've bought some from a seller on ebay before as well - if you search for 'mic 6' you can find his listings.

[wildwestpat]

Speedy metals sells cast aluminum tooling plate. It's under 'plate'. I've bought some from a seller on ebay before as well - if you search for 'mic 6' you can find his listings.

Go back to the requirement and your measurement technique. The measurement flatness over a 24" x 3" area is difficult and either involves some extra work or relies on existing hardware which or may not have errors.

Positional errors are more to do with lead screw and is used linear scale errors not bed flatness. If you are going to accept 0.001" positional error then you do not need to worry too much about the bed supporting the rails provided the orthogonal axis (Z) is at a true 90 degrees i.e. trams correctly. IMOH the bed deviations show up as an angular deviation and you would need some significant angular movement to get any impact on your 0.001" tolerance over the travel of 7.5" from the centre of the bed.

So relax and get building as JSHEERIN has suggested start a new build thread with the proposed build in sketch form and details of the rails bearings / trucks. Above all define the envelope that the tool must encompass in terms of axis travels type of material to be cut and the accuracy required. Do not forget accuracy has two main factors - absolute accuracy and repeatability - so be clear as to which you are seeking to achieve. I have often seen requirements for very tight tolerances when what was required was a fit with mating parts made on the same machine (or process in the case of moulded parts). All machine design is a balancing act and trade off between the precision of parts and their contribution to the overall accuracy of the parts produced. So do not just home in on the rail supports - consider the tolerances imposed by the sliding of the trucks / bearings on the rail as seen from the other axis as all the errors are contributors to the precision of the manufactured part.

Regards - Pat

[bassman2914]

Well, I'm not sure if people of clear on what I'm saying so I'll post some pics. In the end, if i have to i'll use mic-6 but that's only if i can't get what i need from 6061-t6511. As you can see from the pics it shows two pieces one of which(the left) is warped(highly exaggerated). To the piece on the left there will also be a linear rail attached but this is just for concept. The piece of the right is the piece I've been talking about that needs milling. Imagine there were bolts going from the piece on the left into threaded holes on the piece on the right about every 4 inches. When they're bolted together the rigidity of the piece on the right will keep the piece on the left straight to hopefully within .001" meaning the rail will also be straight. The piece on the right is .375(.5 if it's 6061 and .375 if it's mic-6)x3x24(24 if mic-6 and 30 if 6061). The side that would need to be milled is the side that will be touching the piece on the left. Is sounds like you guys have been thinking I've been talking about facing the top side of the piece on the right but I'm talking about facing the small side. Now that we're all sure of what side I'm talking about we can get back to the question. **If I mill that side, since I will be removing little metal relative to the total mass will it cause significant flex in the same orientation as the piece on the left?? I see how facing the top might cause a lot of warping because of the volume of removed metal but the side has very little and it will have to bend a 3 inch thick piece. If you guys still think it's going to flex too much then where do you suggest I get some mic-6 or similar stress relieved metal?? Thank You for all the help guys. Especially WildWestPat since you've posted on all my threads. My questions just don't seems to every be previously asked questions(unless they're deeply buried in a long thread). I have $600 and am ready to build so i got to get this question resolved.

[bobgerman]

Forget the mic-6 and think shims or scraping. (During final assembly of course)

First off, the work your describing will make little difference on the straightnes of your parts relative to the starting condition of the stock. Aluminum will move more then your desired tolerance merely by temperature differences. You can't "chase" flatness.

If you want rigidity and stablility use steel...or better yet cast iron. Even though steel is three times heavier then aluminum it is also three times as stiff and one third the cost.

Unless you have access to highly accurate machinery and inspection/measurement tools to do your work you won't even be able to hold those tolerances anyway. Even the best of machines are still hand scraped to achieve high accuracy. Shimming will get you there. One thou (.001") in 15 inches is considered quite accurate.

Look for the book "Machine Tool Reconditioning and Applications of Hand Scraping" by Edward F. Connelly (1955)....out of print and difficult to locate but well worth the read before starting your project....You'll see what I mean.

[bassman2914]

First off I don't think I have any way of machining steel to any tolerances. I feel like scraping the metal would be ridiculously slow and couldn't be done if the metal isn't already straight to within a few thou. I stated in there somewhere that the pic was extremely exaggerated and the actual warping of the rails is only warped .008" over 24" and I've calculated based on the rigidity of the piece i'm using to straighten the rail compared to the warped aluminum and the rails, it should be straight to within .0005" when it's screwed together. I don't feel like expansion from temperature will be a problem since the metal will expand uniformly and won't induce warping. even if it does I can't imagine it'd be much since this machine will be in my basement and the temperature will be fairly uniform throughout the year. In the end I'd just like to get the parts as close as i can to start with so i don't regret it later. If i got it within .001" straight and then it warped .005" on a hot day that's be .006" total. If it started at .01" plus .005" that's .015 total. I'd much rather have .006 than .015 any day. again this is a hobby machine and I'm likely wasting money to get more accuracy than possibly usable and if you all agree please say so. this is all just for the frame so a $100 difference in a $1000 dollar project isn't exactly a huge increase in cost.

[bassman2914]

I asked a close friend for his opinion on this matter and he recommended pretty much what bobgerman recommended. I will be using shims since they're pretty cheap and he brought to my attention that apparently cast aluminum tends to crumble when it is tapped so that is a major flaw in that idea. After a little more research it looks like thermal expansion can be a problem but i won't bother worrying about what I can't do anything about(since I can't machine steel). If anyone knows where I could get something that's about 24" long that is extremely straight that I could use for adjusting my machine using dial indicators to determine it's level of straightness that'd be fantastic. I could use some mic-6 for that but only if I can get a very thin strip so the cost is low.