Options for installing linear guide rails ?

[flanagaj]

Auto-collimator or laser Interferomter are options, but it all depends just how accurate you want to go.

No point going for guides in the accuracy of tenths over their length, when the spindle flexes a couple thou under any kind of load.
Also, unless you get the base heat-treated/stress relieved after welding, it'll gradually flex as the weld induced stresses gradually relief themselves over the next few years, throwing your original accuracy out.

That's a bummer. What do most home builders do, do they not use steel ?

Thanks

[Jonathan]

What do most home builders do, do they not use steel ?

Most attach one rail with a couple of bolts (this is the reference) then assemble the gantry without the ballscrews/drive attached. Push the gantry back and fourth by hand several times so that the second rail is automatically pulled parallel to the first. Put one bolt in the second rail... move gantry back and fourth, add another bolt ... repeat. Then just check there's no tight spots at the end. This is of course not the best way to do it, but it works.
For Y and Z you can do the same. Alternatively as the rails are closer you can attach a DTI to one rail, as Mark suggests, and indicate off the reference surface (there's sometimes an arrow to show which - check out the Hiwin linear rail guide). Look at how much it varies and tap it into place.
Before mounting the first rail you could just put one bolt in, then attach DTI to one bearing block and indicate off the frame to get it parallel to the frame.

Edit: This may interest you if you've not already found it:
http://www.cnczone.com/forums/mechan...nterlines.html

[flanagaj]

Most attach one rail with a couple of bolts (this is the reference) then assemble the gantry without the ballscrews/drive attached. Push the gantry back and fourth by hand several times so that the second rail is automatically pulled parallel to the first. Put one bolt in the second rail... move gantry back and fourth, add another bolt ... repeat. Then just check there's no tight spots at the end. This is of course not the best way to do it, but it works.
For Y and Z you can do the same. Alternatively as the rails are closer you can attach a DTI to one rail, as Mark suggests, and indicate off the reference surface (there's sometimes an arrow to show which - check out the Hiwin linear rail guide). Look at how much it varies and tap it into place.
Before mounting the first rail you could just put one bolt in, then attach DTI to one bearing block and indicate off the frame to get it parallel to the frame.

Edit: This may interest you if you've not already found it:
http://www.cnczone.com/forums/mechan...nterlines.html

Ok. I am happy with that for getting the rails parallel to one another, but what about the alignment of the first rail. For example, how do most people get the rail level across it's full length making sure there are no high or low spots and ensuring the rail is linear rather than like a piece of spaghetti ?

[Jonathan]

Ok. I am happy with that for getting the rails parallel to one another

When you're drilling the holes to tap for the rails start with a spot drill the right size to make sure they are in the right place. Could perhaps machine some steel bar with a point on the end and harden it, like a center punch, but the same size as the hole in the rail. Use it to mark then drill the holes one at a time. With an M6 bolt in the hole in a profile rail you get about 0.5mm clearance (from memory) so if it's more than that off you would have to start machining the bolt heads down... bad.

For example, how do most people get the rail level across it's full length making sure there are no high or low spots and ensuring the rail is linear rather than like a piece of spaghetti ?

The simple answer is, I'm pretty sure, that most people (myself included) don't bother. They rely on the aluminium extrusion or steel being flat enough to start with. Aluminium extrusion is much closer than steel but obviously not perfect.
You could scrape the steel surfaces that the rail will mount on (after the frame is bolted down and has 'settled')...that would be ideal but I bet it takes forever. As you mentioned in the first post a taught wire plus optical magnification (with weights to tension and put them in bucket of water / oil to stop them swinging) can act as a straight edge to align from. This enables you to take out any bends is the rail, which there will be, but I'm thinking removing twist is going to harder to achieve?

[m_c]

Ok. I am happy with that for getting the rails parallel to one another, but what about the alignment of the first rail. For example, how do most people get the rail level across it's full length making sure there are no high or low spots and ensuring the rail is linear rather than like a piece of spaghetti ?

Without having access to (very) expensive measuring equipment, a big straight edge and Mk.1 Eyeball.

A lot of people don't require a great deal of accuracy on such a large machine. For woodwork, accuracy to a few thou is good enough, and a more typical use for that size in metal is a plasma table, where accuracy to 0.25mm would be acceptable.
You need to consider what exactly you want to machine, and what kind of tolerance would be acceptable. It's all very well aiming to build a very accurate machine, but will you need the accuracy, and will it be worth the additional cost?