[Mark86]

Hey Guys,

Just looking for some input on the design of a lathe bed. Bed is 1000 x 75 and will be filled with an epoxy granite mix. The bed will be ground flat once the epoxy has had a few weeks to stabilise.

If i weld the bed am i likely to get twist in it down the track due to stress in the steel? or will the volume of EG prevent this from happening?

Would I be better off without the 8 plates inside the bed and just having a solid mass of EG maybe using threaded rod attached to the top and sides so there is something for the epoxy to bond to?

Hopefully i have attached the images correctly..

Thanks for any input.

Mark

[harryn]

Hi Mark, Interesting project.

EG is a mixture of aggregate and epoxy, and I am reasonably sure that welding is going to damage the EG locally, as that is far too hot for epoxy. At best, it will debond and smoke a lot. At worst, I suppose it might light up with flames - not sure.

If you are just having one surface ground flat, I would be tempted to do that first, then lay the plate flat side down on a really flat surface (like a poured, self leveling epoxy flat plate). Use the bottom of the surface plate as part of a mold to pour in your EG. Over the distance of 1 meter, a poured, self leveling epoxy plate is very competitively flat with any other method, just not very scratch resistant.

If you need two surfaces ground flat and square to each other, that is more complicated, but perhaps it is useful to consider doing one surface this way, then just having the second surface to grind flat and square.

The bond of the steel to the EG is pretty strong, but you could enhance it with just any sort of threaded rod. The plates are interesting, but I am not sure that they are helping you as much as creating the potential for voids during a pour. (corner effects)

I would avoid thermal processing of your steel plate at all costs, and even just thread the rods into tapped holes / epoxied into place. BTW - what steel are you planning to use for the bed? That is pretty important so it does not twist just from being surfaced.

[RICHARD ZASTROW]

Stress relieve the weldment after welding and before finish machining. Many machine tool builders stress relieve after welding, rough machine, stress relieve again before finish machine then finish machine.

Dick Z

[BobWarfield]

Consider flipping that design upside down so you don't have to worry about stress relieving. Something like this concept:

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

With this approach you're aligning the linear rails while there isn't much structure and then pouring the E/G to "pot" them in place. If you do it right, the result will be very stable versus vibrations due to the E/G and the rails will be very well aligned.

Cheers,

BW

[rowbare]

I agree with what harryn says. Do your metal work, stress relieve and then add the E/G. The only thing you might leave for after adding the E/G is the final grinding.

Have you seen this thread?

http://www.cnczone.com/forums/showthread.php?t=42308

bob

[Mark86]

Thanks for the input.

The plan was to weld the "frame" together, let that settle for a couple of weeks.
Next I would fill it with EG then let that settle for a couple of weeks then I would get the top of the bed ground or milled flat.

I do like the idea of upside-downing it and using the epoxy surface as the surface to mount the ways to...

Would a surface flat to a thousandth of an inch be possible by casting a layer of epoxy on top of the EG?

Cheers.

[harryn]

The surface plates are quite flat - I think yes to your question on tolerance.

The limitation to keep in mind is that self leveling epoxy is designed with wetting agents to really level out well, with no fillers. The final surface is not hard like EG or steel - just very flat.

Any internet search engine will find a lot on these.

[BobWarfield]

Note that the hardness of the E/G in this type of design is less critical, because the inserts provide most of the support for the linear rails. A lathe would be a particularly good application because nothing sits on the E/G but the rails. Hence the job of the E/G is then most to be stiff rather than flat. You could build a slant bed design pretty easily too, if you so desired.

It's worth taking some time to carefully think through how to get the headstock and spindle properly aligned with the bed rails and similarly dampened with E/G.

Cheers,

BW

[Mark86]

So...

Interested in the flipped bed idea and using an epoxy surface plate to mount the ways of this lathe to, I went out and got some epoxy and a slab of granite just to see how flat and suitable the epoxy bed would be.

Here's some pics and the result. In the photos the surface looks like glass however the epoxy hasn't cured yet... Hope it stays this flat.

The granite is just a 500mm square paver thats been cut in half.



I poured 60cc's initially to make sure the epoxy wasn't going to leak out the sides.



Then poured another 300cc's making a total of 360cc's which works out to be about 3mm's of epoxy over the surface.



Next, I ran over the surface with a gas stove lighter to get rid of all the bubbles. This worked a treat and resulted in the glass like surface shown here. Like I mentioned above the epoxy hasn't cured yet so I'm hoping the end result is as flat as this.



I'll post an update once its cured.

Now.. How do I test for flatness? guess I'll be pulling out the straight edge and feeler guages?

[BobWarfield]

Your flatness will be decent, but not awesome. However, you don't need surface plate flatness for this application. You should level your embedded mounting points, bolt down the rails, and pot them in place with the epoxy so they're partially submerged. They'll be well supported that way and flatness will not be the issue.

Cheers,

BW

[harryn]

Hi, just FYI, here is a link to a firm that sells self leveling epoxy and some info on it. It is not "just epoxy", but also includes wetting agents to help it flow out better.

http://www.precisionepoxy.com

One method to obtain an indication of level and flatness is to put some soapy (not sudsy) water on it. Soapy water has very low surface tension, and will show high / low spots very fast.

Another method is to put a sheet of glass on the surface. Float glass is quite flat (not perfect, but pretty decent) and imperfections will be visually obvious. Some people combine the water and glass methods to force a sort of image into the glass / surface plate interface. There are precision instruments made which are based on this technique.

[altaic]

The problem with pouring a layer of self-leveling epoxy on E/G is that one loses the benefit of having a durable and hard granite-like surface. Unfortunately, as one might guess, E/G does not self-level due to the suspended particles. Vibrating the mold can help with flatness, given a properly set up vibrator, but will also introduce waves in the surface whose configuration will be dependent on the shape&size of the mold cavity, the shape&size of the particles, and the vibrational frequency (the vibration will cause the particles to form modes in the cavity).

To get perfectly flat E/G without having a layer of pure epoxy, first pour a low-viscosity, low-temperature-cure, pure (no additives which can cause shrinking, and no particulate which can cause localized distortion whatsoever) epoxy surface plate, perhaps on top of a slab of granite as you have done. I always make the epoxy at least twice the thickness of measured meniscus. Speaking of, since edges of this surface plate will not be flat due to the meniscus of the epoxy on the edges of your surface plate mold, the surface plate should be larger than the final E/G surface you want to make.

Then, build an upside-down mold on the epoxy surface plate and treat the surface plate and mold with a mold release that can have consistent thickness. Pour in E/G and compact.

The choice of epoxy in the E/G is important: since the surface plate is low temp, your E/G will likely have to be low temp as well, but low temp=less thermal distortion which is good. I've found Pro-Set 117LV (low viscosity) resin with their 226 or 229 (high temp cure not required) hardeners to be quite good, however I still haven't gotten around to sending Cameron some samples for proper testing. Too many projects! :drowning: