Sciencemadness Discussion Board

Question for a machinist

gnitseretni - 4-3-2012 at 15:37

I have a question for a machinist.

I need a 9 square inch 1/4" thick aluminum plate that's as flat as possible. McMaster Carr sells a 12 square inch 1/4" thick aluminum plate that's flat to within 0.002". Thing is I need it to be 9 square inch. The only tool I have available that I can use to cut the plate is a jigsaw. Is a jigsaw fine to cut the plate or will it mess up my flatness?

bahamuth - 4-3-2012 at 15:42

If you are very gentle and careful it shouldn't damage the flatness.

But why not get someone to cut it on a laser, or a waterjet. Every good machineshop has those things nowdays..

gnitseretni - 4-3-2012 at 16:07

I don't know any local shop that uses laser, waterjet but not laser. The waterjet shop has a $50 minimum. I'd like to save me those 50 buckaroos if I can cut it myself without messing up the flatness. I think I'll do it myself and just go real sloooooooow! ;)

bahamuth - 4-3-2012 at 16:37

Remember to get a blade for the saw that works good for Al, and get e spare or two. In my experience worn tools are the usual cause for failure when doing precision work..

m1tanker78 - 4-3-2012 at 16:47

1/4" Al plate should easily survive a jigsaw, distortion-free. As suggested, use a quality new blade. Just to be sure, you can also spray some cutting fluid on it periodically as you move along. I have some blue stuff that's GREAT for cutting and drilling but the label is history so I don't know the name. Don't forget to bevel the cuts or at least file down the burrs to keep it flat as possible.

Tank

bbartlog - 4-3-2012 at 17:23

Only real risk IMO is if the material has internal stresses already (from rolling or whatever methods were used to manufacture it), in which case there's a chance that cutting it will allow it to warp slightly. Otherwise, as long as the blade is sharp and you have a good flat support while cutting, it shouldn't cause a problem. Of course if you need the edges to be flat you will also need to do some careful deburring.

cyanureeves - 4-3-2012 at 17:31

tap magic oil works great and you can easily cut 1/4 aluminum with a hacksaw.you can cut 4" into the material on all four sides leaving just one inch of meat in the middle which you can cut with the blade alone. oops! no you cant,i see you will start with a 12" square.the jigsaw should eat it like butter.noisy butter!

[Edited on 5-3-2012 by cyanureeves]

watson.fawkes - 4-3-2012 at 17:37

Quote: Originally posted by gnitseretni  
I need a 9 square inch 1/4" thick aluminum plate that's as flat as possible. McMaster Carr sells a 12 square inch 1/4" thick aluminum plate that's flat to within 0.002". Thing is I need it to be 9 square inch. The only tool I have available that I can use to cut the plate is a jigsaw. Is a jigsaw fine to cut the plate or will it mess up my flatness?
"As flat as possible" is a dangerous phrase to give to a machinist. With scraping, you can easily get to 0.0002", ten times the accuracy you specify. Regardless of your expected tolerance, do you plan to actually measure the flatness?

If you use hand tools to cut the plate, you won't be adding any distortion that wasn't already there. 1/4" aluminum will cut quickly with a hacksaw, although you'll have to be tricky about cutting from each side. You can cut aluminum in that dimension with a table saw, for that matter.

If you know what alloy you have, you can look up a heat treatment schedule for stress relief. For certain common alloys, you shouldn't need more than a bake in your home oven.

As always, if we knew the application, it would be easier to assist.

Morgan - 4-3-2012 at 18:35

Maybe a finer tooth blade would be better so that it doesn't grab as easily.

gnitseretni - 4-3-2012 at 19:58

The plate will be used as a bed on a 3D printer. I need a bed with a flatter surface than the one I have now to achieve better accuracy and a better surface finish on my printed parts.

@watson.fawkes: I'm not familiar with scraping. Care to elaborate? Is this something I would do myself or is it done by a machine?
I did a quick search myself. Are you talking about laying the work on a surface plate that transfers a dye onto the work piece to reveal high spots, which are then scraped off, and then simply repeat the process until the work piece is flat?

watson.fawkes - 5-3-2012 at 07:19

Quote: Originally posted by gnitseretni  
The plate will be used as a bed on a 3D printer. I need a bed with a flatter surface than the one I have now to achieve better accuracy and a better surface finish on my printed parts.
[...]
Are you talking about laying the work on a surface plate that transfers a dye onto the work piece to reveal high spots, which are then scraped off, and then simply repeat the process until the work piece is flat?
Yes, that's the scraping I'm talking about.

What's the planarity of your current bed? Is just warped, or does it change under temperature? You can estimate planarity with a surface plate and a dial test indicator mounted on a surface gage.

Are you sure it's the bed that's the problem with the printer? What kind of printer is it?

gnitseretni - 5-3-2012 at 16:06

I found a website of a guy that does alot of scraping. I contacted him to see if he'd be interested in scraping my current bed. I mentioned that I can get a plate that's flat to within 0.002" but that I wanted a plate that's flatter. He said he didn't think he could do better.

Did some more research and I think I'll look for a local shop that has a surface grinder and have them grind my current bed. Wikipedia says that most surface grinders (depending on usage) should get the surface flat to within .0001". But even if it's .0005", it's good enough for me ;)

The printer is a Botmill Glider 3.0.

condennnsa - 5-3-2012 at 16:49

If you want flat, use a piece of float glass.

All window glass is float glass.
Due to the production process it comes out extremely flat and parallel .
I use 3 pieces of float glass on top each other (each 4mm thick) glued together with double sided sticky tape,
these 3 on top of a granite 50*50cm flooring slab, as a poor man's surface plate. it is extremely accurate.



[Edited on 6-3-2012 by condennnsa]

Wizzard - 5-3-2012 at 16:52

Granite makes and excellent CNC surface, it's very flat and hard!! Glass is my second choice, though.

phlogiston - 6-3-2012 at 03:43

Although getting a flat plate is probably the best way, I have always wandered why you could not correct for inperfections like this in software on CNC machines. The nozzle height could be adjusted for the height of the plate.

bahamuth - 6-3-2012 at 03:58

Quote: Originally posted by phlogiston  
Although getting a flat plate is probably the best way, I have always wandered why you could not correct for inperfections like this in software on CNC machines. The nozzle height could be adjusted for the height of the plate.


If you think a little about it it would not do any good...

Imagine the printer bed is "convex" on the middle with an increase i height of 0.2 mm, would it do any good to move the nozzle up in that area? You would get a convex product...


IMO the glass-plate idea seems like the best option, cheap aso.

watson.fawkes - 6-3-2012 at 06:06

Quote: Originally posted by gnitseretni  
Did some more research and I think I'll look for a local shop that has a surface grinder and have them grind my current bed. Wikipedia says that most surface grinders (depending on usage) should get the surface flat to within .0001". But even if it's .0005", it's good enough for me ;)

The printer is a Botmill Glider 3.0.
The biggest problem you have is that you have a machine that's has plastic connectors as part of its basic structure. This limits its inherent accuracy merely through its flexibility. It's not visible flexibility, but it's a fools game to try to strive for ten-thousandths on a machine like that. I'm guessing that the stock plate you were looking at ordering will be just fine without further processing.

If you want more accuracy, I'd worry about the parallelism of ways on which the axis rests, and the parallelism of the mounting from those ways to the output platform. Unfortunately, there's no built-in way to adjust those ways, although you could use shims. But before even that, you need a way to measure it, which is the traditional surface plate + dial test indicator + surface gage combination. So unless you want to spend on that (or have a friend at a machine shop that will lend you access), there's just so sense in bolting a very flat plate onto an axis that doesn't travel in parallel and'/or has significant skew or parallelism issues. The bed is only a single step in the metrology chain, and without measuring it, there's no particularly good reason to believe it's the worst problem.

I doubt glass will work for you. While it comes out of the factory flat, there are any number of ways that it can deform. In the design of your printer, the bed is bolted to a movable frame. If you use glass, you've still got to mount it. You could drill the glass, I suppose, but if you don't, you'll have to make a sub-bed for it, and that will be just as good for that machine.

Lastly, surface grinding isn't magic. The biggest problem is that the workpiece generally deforms under the mounting force. With ferromagnetic pieces, the typical problem is that a magnetic chuck will pull a warped piece flat. After grinding, it's flat while still mounted, then springs back after demounting from the chuck. The solution, no surprise, is to scrape. You make one side perfectly flat; that side goes face down on the chuck. The piece then doesn't deform on the chuck and it comes out parallel without any spring-back. Grinding aluminum plate has all these problems, plus difficulty in chucking it. Thin plate like 1/4" is likely to deform if simply put down in a clamp. You can laminate to a ferromagnetic plate with superglue and then dissolve it out, but you then have to ensure that the plate is flat and parallel and the mounted side of the Al is flat. This all seems like overkill for your application.

gnitseretni - 6-3-2012 at 10:08

I understand why you question the rigidity of the machine. I don't like the plastic parts either, nor do I like the threaded rods. Getting a new bed is only one of the modifications I plan on making. I plan to turn my printer into something more like the one on this page http://www.hydraraptor.blogspot.com/ - And instead of rods for the y-axis, I plan on using shafts that have a support rail.. the one halfway down this page http://www.mcmaster.com/#linear-shafts/=gjplb4 - and I'll get larger diameter rods for the x and z-axis. But those mods are for later, new bed first, which will be an improvement already.

I have thought of glass as well, btw. If I wanted to use it, I could secure it to my current bed with bulldog clips, like others have done.

Rosco Bodine - 7-3-2012 at 06:07

There are methods of electroless plating of low melting bright mirror finish bismuth tin solder alloy coatings which could be useful for this sort of thing.
Some of these coatings are fusible in place by heat treating and basically leave a surface mirror which is close to optically flat ......which should be flat enough...
on the flitter flatness index of course :D

Silver/tin 4/96 solder leaves a mirror and integrates / alloys with many substrates ......solid copper would work best instead of aluminum if you decided to go with a mirror plating scheme

The same techniques used for making telescope mirrors could be applied to what you are doing.

[Edited on 7-3-2012 by Rosco Bodine]

gnitseretni - 7-3-2012 at 09:34

Sounds interesting. I'll look into it.

watson.fawkes - 7-3-2012 at 13:00

Quote: Originally posted by Rosco Bodine  
There are methods of electroless plating of low melting bright mirror finish bismuth tin solder alloy coatings which could be useful for this sort of thing. [...]
The same techniques used for making telescope mirrors could be applied to what you are doing.
Optically reflective is a kind of flatness that's called surface finish. This is a rather different notion of flatness than needed here. Surface finish has to do with flatness across a small patch of surface, say, 1 sq. mm. The kind of flatness needed for machine fabrication is flatness across the whole of the platter. The example of local flatness for a focusing mirror is a perfect illustration of the difference.

In particular, using an adhesive material (like molten metal that solidifies) will do nothing to improve total flatness for this application. There are a few high-tech materials, like special filled epoxies, use to restore flatness in machine tool reconditioning, but they require a reference flat to work. And they give you one flat part at the end, not an entirely-parallel metrology chain.

Rosco Bodine - 7-3-2012 at 16:25

A pool of mercury is pretty flat. You could try that. You can get flatness and surface finish at the same time on something as small as a 3X3 inch piece.
Seriously, a 3X3 1/4" inch aluminum plate could be hand lapped to probably a thousandth or less with a little elbow grease simply using successive grits of emery paper secured on a piece of float glass and then wet lapping with polishing grits and with rouge. An orbital motion buffer with double sticky mounting tape could be used as a motor drive to make the job go faster.

gnitseretni - 7-3-2012 at 16:42

A pool of mercury? I think I'll pass up on that :P

Btw, it's 9 square inch, not 3 square inch.

Neil - 7-3-2012 at 17:18

Speaking from my experience; watson.fawkes has the dead right of it. Lapping to perfect flatness is very difficult, without a lot of practice, no mater how smooth you think you've lapped something you'll find you are off when you start scraping.

As soon as you add a motor to the issue it gets worse...

Personally I'd look into the run out of the mechanism before worrying about the plate.

Rosco Bodine - 7-3-2012 at 18:42

Ummm yeah I was thinking of lapping 9 square inches not 81 square inches

9X9 should be called 9 inches square, and that much area would require machining and could be expensive.

Sheet lucite or cycolac might be a better and cheaper way to go.
What material are you printing, for which such accuracy is needed
for the base underneath? Unless you have vacuum holding down
whatever you are printing, and that material is exceedingly thin,
then that material will be laying across irregularites in the surface
rather than conforming perfectly to the bed surface on which it lays.

gnitseretni - 7-3-2012 at 19:42

Crap, I meant 9 inches square (9X9). Sorry, everybody! :(

Rosco Bodine - 7-3-2012 at 21:14

I mentioned lexan sheet which may be flat enough as is, but since it is
flexible, you could use shims or adjustment screws bearing upon the underside of the lexan for curvature correcting any bow, if it was secured to a rigid backing plate by its corners. A mounting frame with a "stadia" wire used as a straightedge
could be used as a Glint indicator where you could see daylight reflection under the wire as you visually check the surface flatness crossways in something like an asterisk pattern, gauging the surface with the taught wire.

gnitseretni - 8-3-2012 at 03:54

Lexan wouldn't work. The bed is heated from underneath with resistors, to about 70C. The bed is heated so the printed part doesn't warp. It needs to cool off slowly.

dann2 - 8-3-2012 at 06:24


A book, Amateur Telescope making my Albert G. Ingalls, describes making optical flats from glass and metal. These thing can be made to millionths of an inch using elbow grease and time.
Will post some photo of pages if you want.

gnitseretni - 8-3-2012 at 16:11

Nah, I think I'll just get that plate from McMaster that I mentioned in my first post.
Thanks though.