new build opinions required

[JAZZCNC]

Eddy is quite correct and most of the larger machines I've built have been welded and bolted. It just makes good sense when you think about.!!
Also I've said this many many times.!! . . The more adjustabilty you build into a DIY machine the easier it will be to set it up. Bolting makes for another area where you can tweak a machine into alignment etc.

I also mostly Agree with Eddy's take on it except that full epoxy is better for high tolerances than Shimming because it's not really, it's just easier(Quicker). I'd actually go has far to say it's potentially less accurate because your soley relying on the epoxy with no room for error or improvement where has with shimming you can tweak out any error it just takes a lot of time. Personaly I'd do a mixture of both on a large or long machine that requires higher accuracy.!!

[Jonathan]

I'd actually go has far to say it's potentially less accurate because your soley relying on the epoxy with no room for error or improvement where has with shimming you can tweak out any error it just takes a lot of time.

The problem with shimming is you can't compensate for rail twist with it. I discussed it at length in the recent build log. I didn't include the readings from the shimmed rail in that post, however if you want I can dig out the graphs obtained when the rail was shimmed and you'll see the error was still far worse than the final result obtained using self-levelling epoxy. I'll quote myself, from the build log:

At this stage it is interesting to note that the bearings slide smoothly on the rails, so although compensating for the height error with shims or a machined profile is sufficient for the bearings to run smoothly, this does not by any means assure accuracy since neither method compensates for the angular error.

It all depends on how accurate you want the machine to be and its intended use. The error at the tool introduced due to rail twist will be exceptionally small, so you'd need a very strong machine to notice it. The additional wear on the bearings caused by the rail twist will depend on the stiffness and accuracy of the structure they're mounted to, so it's hard to judge, but in a DIY CNC situation is unlikely to be too significant. I'd still choose epoxy though as it's so much quicker and easier than messing about with shims and enables one to get very good accuracy with less measuring equipment.

[JAZZCNC]

Shim steel is more expensive that I thought, do people just cut up food/drink cans or is there somewhere to buy an assortment cheap ?

Often I use tin foil has it's thin cheap and won't rust.!! If I need more than 2 layers I'm upset.!!

The problem with shimming is you can't compensate for rail twist with it. I discussed it at length in the recent build log. I didn't include the readings from the shimmed rail in that post, however if you want I can dig out the graphs obtained when the rail was shimmed and you'll see the error was still far worse than the final result obtained using self-levelling epoxy. I'll quote myself, from the build log:

Why not.? Shimming is all about where you place the shims.! I agree it's not easy in comparison to epoxy but it can be done, I've done it many times.

Your experiences with epoxy are on a short machine but on a longer machine then things are slightly more exagerated and just relying on one method that is fixed or not very flexible is not so clever IMO. That's why I say for a Long machine(requiring higher precision), longer than 4ft I use both methods.!

Why make life hard for the sake of some plates and bolts, which if Epoxy hits the mark first time you don't need to touch, if it doesn't then fine it can still be tweaked if needed. Adjustabilty is Key and the more of it you have without compromising the machines strength the better at DIY level.!!