accuracy

[lgalla]

May sound like a stupid question,But I am not a machinist.How do you accurately mount 2Z rails and 4 blocks to the Y which has 2 rails and 4 blocks also.Could be a nightmare for a non machinist.Any tips or tricks?Is a cheap drill press good enough?One tip I read was to down load 1 to 1 drawings say from THK,print actual size and use as a template.Perhaps using a machine shop is the answer.
Larry

[monkeyturnip]

build in adjustment,

ive attached a pdf of JGRO's (i hope he doesnt mind) drawings in the download section. its a iso' of his design, he has a over size hole for what its for, ie if the rod is 8mm and the bearing is 12mmod then a holde of 16mm was drilled, a block with a 16mm hole attached and then on the edges a hole drilled for bolts to go through, these bolts give you 4 way adjustment.

do this at every end of rods, and bars for guide rails and you can adjust very accuratley. its a smiple approach but works, it takes out the need to be highly accurate in the construction of the machine.

[Zumba]

For high accuracy, machining is the only way to go. THK rails are accurate in profile and dimensions, but are not straight out of the box. Like any long bar, they're fairly elastic and will only be as straight as the surface to which they are mounted.

Mounting surfaces for woodworking machines can be machined entirely with a mill. If it's a big machine, it'll require an even bigger mill. Metalworking machines should be milled and then ground with a precision surface grinder/way grinder that's larger and more accurate than the machine being produced.

For mounting THK and similar linear guide rails, the best way to do it is to machine a slot. The flat bottom and the inner shoulder are the critical reference faces. The outer shoulder is chamfered to allow the installation of tapered wedges. Drill and tap to suit. Basically the same for the bearing blocks.

See page 17 (19 if you follow the PDF page numbers)
http://www.nookindustries.com/pdf/NookProfileRail.pdf

This is how the big boys do it for VMCs, turning centers, bridge mills, etc that have linear guides. Very expensive if you're building a big machine, since you need access to a giant machine.

For a DIY type project, best bet is probably good old fashioned brute force trial and error. Shims and set screws are your friend.

[BobWarfield]

Larry, there is more than one way to skin the cat.

First choice is definitely to machine a pocket for the rail. The pocket would be flat to your tolerances and have a back edge that is perpendicular and flat to your tolerances.

A second rail can live without the back edge, because the first rail can be used to align it properly by placing a rigid fixed length tie between trucks on both rails. This is an approach used commercially.

Next possibility, since you're an epoxy granite guy, is to align the rails by hand, and then pour the E/G to lock them in place. This is not unlike what JerryFlyGuy has been trying to do with a big 5-axis gantry project he's working on. Madvac also did a gantry this way. In the case of the latter two gentlemen, they didn't so much use E/G as a potting substance. Moglice is the "pro" choice, but one suspects epoxy will work just as well in this application.

Hand alignment is a function of being able to measure when you're in the right spot and being able to adjust finely enough to get there. Machinists do this all the time, BTW, when they tram the head of a mill or tram a vise on the table. They "bump" things into position with an indicator, then they crank down the screws to lock it in place.

You can turn to geometry to understand the requirements at hand. Each axis has to travel in its plane in linear fashion, and it has to be orthogonal (at right angles) to the other two axes.

I would carefully review the following and think on the approaches being used:

- Go to my blog page: http://www.cnccookbook.com/CCBlogJune2006.htm

and search for "Rail Mounting Precision for CNC Plasma and Router Tables". Some good references including links to some of these others:

- JerryFlyGuy's long thread about truing up his gantry. Lots of good ideas there.

- Madvac's account of how he achieved accuracy on his gantry:

http://oneoceankayaks.com/madvac/mad....htm#Indexanch

- L.M. Watts approach:

http://www.lmwatts.com/straightedge.html

- My one pager on an epoxy granite gantry table:

http://www.cnccookbook.com/CCEGGantry.html

- Bamberg's thesis "Principles of Rapid Machine Design." Google Bamberg plus the title to track it down. He gives a good description of how to set up axes accurately and economically.

It can be done, and you can build an accurate machine without necessarily having a mill around, but you'll have to be clever to accomplish the task.

You're also going to need some precision measuring instruments. A dial test indicator, and various attachments will be essential. A cylindrical square will be very helpful when trying to measure whether one axis is orthogonal to another.

I would also review Tormach's certificate of inspection document:

http://www.tormach.com/document_libr...Inspect_V4.pdf

Why? Because it explains how the test their machines for various sources of error. Many of the setups are what you would need to use while adjusting your machine to be square and true.

I hope you didn't find that too exhausting. Once you understand what's going on, all these things say the same thing, they just say it slightly differently to give you a fuller understanding and help you see the possibilities.

Cheers,

BW

PS Both you and Zumba may find a discussion Geof and I recently had about milling a 6 foot workpiece on a 4 foot table to be useful:

http://www.cnczone.com/forums/showthread.php?t=53618

It describes two different methods to do precision machining that is "beyond the capacity" of your current mill. There are limits, of course, but it is interesting.

[lgalla]

Thanks for the help guys.Very useful.The rail mounting is easy with epoxy,but looks like mounting the blocks square and parallel is the nightmare.Pretend I know nothing[easy to do].
What tools are used or required to mark the holes for the blocks?Precision square,calipers etc.Are lines scribed?Do you need blue?Come to think about it,I know so little on the subject I don't know what to ask.
Bob,any more epoxy casting?
Zumba,I have not forgotten about you.The supports are just plastic shaped as an L with a pivot in the corner and a spring.The nut simply hits one end of the L and swings the other end up contacting the screw.It is so simple its one of those"Why didn't I think of that"
Larry

[ger21]

Larry, if you already have flat, square panels to mount them to, I would think that you could do it on a drill press with a fence setup to keep the holes in line. Perhaps after drilling one hole, bolt the rail to it and use the rail holes to guide you ?

[BobWarfield]

Let's think about what we're trying to accomplish (and Larry, do go read all that stuff I gave you links to--it explains this process!).

In essence, you are going to mount another axis to your rails by bolting it to the blocks. I would not expect to be able to achieve the required accuracy by putting perfect holes in the axis that mate up to the blocks. If nothing else, bolts are not precision locators, they are clamping devices (yes, there are exceptions, but they're probably not relevant here).

So, think of the bolts as clamping, and the holes as clearance, not locating. Now your issue, as monkeyturnip said, is you need to build the attachment point so that it is adjustable. That just means there is enough play available that you can assemble the second axis on the first and tap it into proper alignment before tightening down the bolts.

How you measure when it's in proper alignment is in the links I gave. Start out trying to understand that Tormach document. Read the thread from JerryFlyGuy.

Let me give a for instance. All the pieces interact with one another. Get one rail accurate, and you can use it to position it's mate. WIth those two square, your next job is to make sure the axis attached travels perpendicular to the first. How to tell? Well, if it isn't, then as you slide along that second axis a test indicator, there will be motion in the dimension of the first axis. So you need to create a rig to measure that and adjust the 2nd axis until that motion goes away.

Without a detailed design of your machine, nobody can give you precise directions for how to do it, and you can't create the precise design until you understand the principles. So you gotta go read the stuff and become your own expert!

Cheers,

BW

[lgalla]

Thanks Bob;Gerry,Lots of good reading.I was afraid of a drillpress being good for the task as
a machinist once told me"you cannot drill a round hole on a drillpress"Perhaps I was looking
for an excuse to buy a small mill.
I understand now that you would use socket capscrews smaller than the holes to allow final tweaking.I have seen posts where guys drilled their X rail holes with a hand drill.I think I would at least rent a magnetic base drill press for that job and My $500.00 Taiwanese drill should be good enough for the X&Yplates.
I want to build a double Z as Gerry is working on.I will use steel plate for the Y&Z and attempt to epoxy surface plate them.That should give an accurate surface and eliminate
shiming.
Ihave a THK 2005 and a 1605 ballscrews for the two Z.
Now to decide on the rails/blocks from stock.
4 HRW-21
4 HRW-27
4 HSR-25
10 SSR-15,short rails 200mm
I am leaning towards the HRW-27.These are TANKS.
For X the orginal plan was for T-7 screws,but 10ft is better suited to R&P.Now I have a few BK-15 Angular contact bearings.
Once I decide on the parts,I will E-Bay what I don't need.
Thanks again
Larry

[schbrownie]

I really like the technical section in NB's catalog. It's got some great pictures of how to mount profile rails to ensure high accuracy.

http://www.nbcorporation.com/product/slide_guide.html

Click on "Technical Information" and go to page A-7

Hope this helps.

~Jon