Epoxy/granite machine concept

[Gizmot]

Hi all,

first of all I would like to present myself. I'm Mathieu and a soon to be graduated mechanical engineering student in Montreal. As part of my last semester project, I tought about making (or at least designing) an epoxy/granite CNC machine. First of all, let say that I have been lurking in the background for about 3 months now and that I went through the whole 200+ pages of the other thread.

Ok,... now I'm planning to create some kind of report including all the details concerning choice of linear motion, motors, software, mixture, process, concepts, drawings, blahblahblah

First part as been to do some kind of technological review of everything available. Some choices have already been made but some are in the concept zone (ex: air bearings). Been doing that for the last month almost full time.

Next part would be getting a good mixture. I checked a couple books by François de Larrard and ckelloug seems to be doing a great job. Gonna leave this part to him

I have done a couple of sketch on how to get some crazy accuraty without blowing the bank. This topic will be about showing and discussing idea on how to create that big *****.

to be continued soon...

[Gizmot]

Well well...

One of the first home made CNC I have seen here was the Madvac design. I was quite impress with the result. What did interest me particulary was that Moglice stuff. I found that quite funny because people tought I was crazy at one time in NZ win I used epoxy to create some accurate cast surfaces. Obviously, you can see that I'm definitly planning to do it again.

After that I ended up seing that: http://www.elektronikforumet.com/for...er=asc&start=0

What is funny is that its looking very closely to what I had in mind.

ok enough stupid stuff about myself

---------------------------------------------------
some question I have been asking myself

1. what is the usual tolerance between correct bolts size and bolting holes in linear rails? trying to figure what kind of possible adjustment.

2. guys that played with some epoxy/granite, how easy would it be to join together (or cast in 2+ times) different sections?

3. How the bloody hell do you plan to vibrate a medium-large machine base?

----------------------------------------------------------------------
Now some concept ideas (no pictures yet, I'm sleepy)

To create the base (for a machine looking like the swedish guy- link above)

1-create a large accurate bottom reference surface in the casting box with epoxy. This would create a neat accurate surface to bolt you rails into.
2-Do the same thing but with a countertop granite plate. You can even ask the guys to precisely drills some holes in the bottom for the bolts (fill the gaps with clay when casting).
3-Do exactly like the swedish guy, then fill the gap under the rails in a later stage. This would allow to create a molt in the up position.
4-Do like the swedish guys and hope that gravity to going to create a nice flat surface for rails.
5-Create/buy/built some steel plates and cast them in the bottom on the box (upside down casting)

My personnal favorite is currently #2. What I like about it is that the product is final, no need to tweak to make sure rails are level, good rotation angle. Just have to check alignment.

-----------------------------------------------------------------------------
Crazy concepts I'm playing with

1. the use of air bearings. Super hard flat surface? seems a good choice. Only problem is that you need constant air supply, safety system, blahblah. But then, would it be possible to lets say cast some kind of granite rails? it is sure hard to beat to have free rails!
2. linear motors. I have honestly hate super high precision mechanical devices subject to wear and dust such as precision ballscrew. The biggest problem with linear motors (exept maybe for cost and controller) is that metallic tables dont like them. A problem we dont have here...

sorry if this is a mess, I will try to do some sketchs to post

[svenakela]

I moved your thread to the project log forum, this will hopefully be a very interesting log to follow

[Gizmot]

no shiny drawings yet, sorry

I'm still attempting to solve building problems. I'm just wondering, based on some kind of "open source". What do you guys thinks about the concept of building an high tolerance steel mold, and when you are finished with the casting, to ship it to the next guy? That makes much more sense than building a great mold for only 1 use.....

[romihs]

Hello Gizmot,

I'm sorry to see that you have received so few replies so far... you have some very interesting ideas.

I am also currently working on a slant bed lathe design with EQ.
The only 'logical' conclusion I have come to is to build an accurate mold that will produce accurate reference surfaces for the linear rails. The only parts of the mold that must be accurate are the beds for the rails (plan to use ground steel plates), the rest can be less accurate (wood, mdf, steel, etc.).
I would only have small threaded inserts so that there will not be any long steel inserts embedded in the EQ which could cause unseen stress in, and possibly warping of the, machine base.

The part I am 'battling' with right now is the dimensioning of the slant bed so that I don't have to build a mold that has strange dimensions.

I must start my own thread here too....

Best of luck.

Regards

Sandi

[Gizmot]

Hi Sandi

actually I kind of dropped this thread because of low interest (until I start getting drawings). Most people are actually only interested in shning drawing. I wanted to use this place as a brainstorming center but oh well....

Since then I kind of changed concept to an overhead gantry (2mx3mx.5m). I'm trying to design a low cost laser feedback system to eliminated the need for high accuraty part. I'm trying to use either flat way or kinematic rolling elements as linear motion is concerned.

as far as your problems are concerned, I think the best solution is probably replication technique. What I would do is:

build you mold with the up side down (wood frame). it should be looking something like this

l'''''__''''''l
l__l'' l__l


then use low viscosity epoxy to just fill the low sides. this is going to give you a perfectly flat and level reference surface. To make sure both side are same level, use a passage way or simply add the midle section afterward (wood box). Fill with eq. flit, remove and perfectly rocking base.

oh btw, feel free to squat this place. the more ideas the better

[hesham morsy]

the mold is the machine


Frame will be welded from two identical bended profiles. Profiles are made from 4 mm cold rolled steel. Flat bars will be welded for guides and for screw mounts. Profiles will be filled with concrete. All mounting surfaces will be machined after welding
http://www.diy600.net/

[will gilmore]

i'm following this thread with interest. i dont have any knowledge to add so i'm keeping my mouth shut and hoping you figure something out that i can learn from i'm thinking about building a cnc bench lathe using a granite surface plate as the base and bolting linear rails to that. could you do something similar?

[Gizmot]

Originally Posted by will gilmore i'm following this thread with interest. i dont have any knowledge to add so i'm keeping my mouth shut and hoping you figure something out that i can learn from i'm thinking about building a cnc bench lathe using a granite surface plate as the base and bolting linear rails to that. could you do something similar?

yes I tough about it. It is a good idea since granite is "usually" flat enough to bolt rails. The other good option is the one I posted a bit higher: using self levelling epoxy in a mold to create a base reference surface.

What is going to be interesting is to see if I'm able to build that goddam laser feedback. Could then just use cheapo DC industrial motors, chains, gears, etc

[greybeard]

Hi Gizmot.
Would you care to share an outline of your laser feedback idea ?
I'm still progressing very slowly with my "spin cast" EG idea, discovering all the things that don't work in the process
I'm also keeping to my air bearing approach, which hopefully will make a major reduction in the part count. It's based on carbon block filter medium and a glass layer embedded in the top surface of the castings.
Regards
John

[Gizmot]

At first I wanted to use air bearing but kind of dropped the idea since it would have required some kind of huge compressor.

My laser feedback idea:

Use if possible a commercial high accuracy laser unit for feddback on the X-Y axis. How I see it currently is either take some unit that can connect to a PC or either smash it open and take directly the analog signal going to the display unit. Wire this up to the PC and use it (or modify some code) with EMC2. Then simply use standards DC motors... or maybe steppers...
In that case the linear motion doesnt have to be accurate or backlash free. I would not fell ashame to use a roller chain as a rack. $$$
The whole question is if I can build quite easily that system. Have have some help from some school teachers and technicians on that one so its not too bad.

[greybeard]

Originally Posted by Gizmot At first I wanted to use air bearing but kind of dropped the idea since it would have required some kind of huge compressor.

Not sure why. I'm only moving a 4lb router in Z and Y axis, with a roller feed for my X axis, but with x4, 8 sq in air bearings on both Z and Y, I'm looking at quite a small through put of air, via the 10 micron blocks, along with sufficient pressure to give me the required stiffness.
If that proves to work will only come with the build, put I'm happy with my plan so far.

Originally Posted by Gizmot My laser feedback idea:........ Use if possible a commercial high accuracy laser unit ...........

What kind of laser unit would this be - do you mean something like a laser tape measure with DRO ?

Regards
John