DIY Epoxy Frame based CNC MILL

John it looks a decent bit of kit that you are making can't wait to see the complete chasis.

David

Hi All

The base casting is poured!

The components.... Mold, turned parts and laser cut parts to be cast in.
Note the bearing support rails and column support pads on the main base plate. these sit directly on the surface plate and are aligned by holes in plate.

Mold ready for casting.

Electric and control supply ducts fitted.
Also note M20 long nuts with 30mm collar from the local hardware (Bunnings) V grooved in the lathe for extra grip. (attached to the bottom bracingf plates)
The 2.5mm laser cut steel box will contain the Y axis drive motor. Rough ground with an angle grinder for good adhesion.

Foot mount (ignore small hole in collar) the washer was bored and fitted to a step in the collar.

Jacks that press motor void box down made from scrap fastenings, the nuts were grooved 12mm for location. limited travel was needed, they worked well. Small long nuts are M6 from The hardware store, gooved in the lathe The white material is painters gap filler to stop epoxy entering the hole, easy to remove after.

Wooden strip: A 10mm sq timber strip was glued to the back of the bearing support rails. This is to allow through drilling for the tapped holes that will later be required to mount the rails. See previous discussion of this process.
I also mounted a jack at the end of each bearing support rail to assure it did not lift during the pour, It will be cast in and lost.

The casting was done in layers, about 25mm at a time. tamping between each layer, the process took about 3.5 hours, my wife assisted throughout tamping I mixed in batches of half a bucket and tamped at the end, without assistance I would not have been able to place the epoxy in time, after a short pause to allow the casting to partly set the brace plates were removed, in case a side panel decided to come away they were tightened against the inner rails and the last layer was placed, this layer was done with a wet mix, 20% epoxy. After placing the mix was carefully screeded level. as can be seen the surface is flat but not without imperfections, this will be attended to when the casting is removed from the mold.

Mixing was slow, a cement mixer would have sped this up but the clean up would be heavy going, and you would need more tampers! unlike portland cement when hitting hardened epoxy with a hammer the surface is marked but it hangs on.

For mixing I used super cheap buckets, doubled up, 4 were cracked by the stress, also an electric drill on the slow setting and a hardware store cement mixing screw. The drill struggled. I mixed 5 litres at a time by volume about half a bucket.

I am still amazed how accurate the finished molds are they fit together perfectly.
The next step is to remove from the mold, clean up and undercoat the casting.

I cant wait to assemble the base, columns and cross member, then take some measurements.

Regards
John

Edited By John McNamara on 28/01/2019 12:25:14

You are making excellent progress John and I’m very impressed with the finished results. I’m quite surprised at how yellow your “Epoxycrete” looks in some pictures but then appears a nice shade of grey when set. Can you explain this, what type of sand are you using?

Hi Vic

Straight from the mold there are there are dead sharp edges and some small holes, the mix is stiff, also other blemishes caused by the self adhesive mold covering "Contact" I use as release. These are dealt with with car body filler, "Bog" in Australia , "Bondo" in the US I am not sure in the UK. A quick coat and then sanding with a half sheet sander for the flat work. I also sand all edges to a small radius. The grey colour is just an etch primer coat.

If you study the images you will also see that the mating edges of all the castings are 2.5mm mild steel sheet placed within the mold before casting. Also the motor mounts and mounting points for ball screws are all cast in 5mm steel plate. Removable 5mm steel panels are also cast in position for a good fit, I wrap them in self adhesive plastic, fit the mounting screws with inserts and screw them into the mold. when the casting is set they can be unscrewed (the inserts are now cast in) and the protective plastic removed. The motor mounting point on the column is done this way as is the motor covering plate and ball screw bearing mounting plate that forms the U channel in the centre of the base casting. This will be more obvious when the casting is turned over. remember it is currently upside down.

The sand used was washed silica as used by the building trade. one part sand to 3 parts granite aggregate with a maximum screen of 8mm. and a minimum of dust. These materials are both easy to obtain.

As can be seen there is a lot of 12mm, 8mm, and 6mm threaded rod used easy to obtain and low cost. It would be possible to simply throw in concrete reinforcement bar. This I did not think was a good idea at all. Firstly it would not allow precision placement within the mold. I wanted to balance all the forces, any temperature effects will be balanced across the section. Secondly the mold itself is partly supported while pouring by the reinforcement.

When I did research on Mineral casting I found hundreds of academic papers on the subject and hundreds of thousands (yes) of forum postings on the subject. Some say no reinforcement! others use it. Some argue about the mix and pouring method with great passion. Not many of them actually build anything.....

No doubt this machine could be improved by using a high tech approach, by how much? probably not as much as you would think (The ultimate strength is governed by the epoxy itself), that's why we make thick sections. Even at this early stage it has taught me a lot and I know it will continue to do so.

I have tried to avoid any hard to obtain materials, and really hope a few people try this methodology for their own projects.

Regards
John

John,

You said:

The sand used was washed silica as used by the building trade. one part sand to 3 parts granite aggregate with a maximum screen of 8mm. and a minimum of dust. These materials are both easy to obtain.

I'm off to mix some cement for a small foundation for a wall, this morning, and I wonder

(a) whether you needed to exclude moisture by using kiln dried sand?

(b) why exclude dust? I have used 'granolithic' as a screed, and that is mostly aggregate, with some cement. The aggregate is described as '6mm to dust' and the dust contributes to the strength, possibly because, like a larger mixed aggregate, the different grain sizes of the sand and aggregate lock well together, and provide strength. Would that not work in the same way with the epoxy mix?

(c) Experience tells me that if I lay a screed on a floor, then work it to a fine flat finish using a float, I need to take care not to over-work the screed, because the more I press it, the greater the tendency for the larger particles to work their way to the top as the smaller particles are forced downwards into the spaces between the larger particles. That can spoil the finish, as well as (I suspect) producing a kind of layering effect, with smaller particles tending to be more concentrated further down the mix. I note that you tamped in layers. Did that help the produce a more uniform distribution of particles? Might be rather difficult to tell, of course, but I shall spend the rest of the day pondering... Perhaps someone has done tests and reported results in one of the many papers on the subject?

Great progress, though. And I agree with your thinking re: concrete reinforcement bars ('rebar' in the UK) and 'inexactly placed' reinforcement.

Marcus

Hi Marcus

Drying
I should of mentioned I spread the sand out thinly about 20mm thick on a sheet of thick black building plastic in the sun to dry it out. I raked it with a plastic leaf rack a few times over a couple of hours to dry it. I did the same for the aggregate. once it was dry it is easy to lift the corners of the plastic towards the center creating an easy pile to scoop up. Maybe not kiln dry but dry enough.

Screeding epoxy is not fun it sticks to everything and you have to remove every drip before the epoxy sets
The mix is rather dry compared to concrete, I have not been able to get a good finish unless i make the last thin coat wetter and sand only. I know a couple of contractors that do pebble mix over concrete, they get a good flat finish using carefully graded pebble sand and a wet mix. Body filler is my best solution so far. An extra step but a nice result.

Regards
John

Hi All

The base casting has been quite a project it is heavy. over 200 KG

The underside of the casting carefully leveled before stripping the mold sides, note the inserts; 2 x 6 counter-bored holes for attaching the columns and the 4 M20 tapped holes for the adjustable feet. the rectangular hole in the center provides ventilation for the y axis drive motor.

Here it is finally undercoated and on the little engine crane I use to manhandle heavy objects.
I needed the crane to manipulate it while I ground all the sharp edges, apply filler where needed and finally give it a coat of dark grey undercoat. There is no shortage of M12 mounting points for the eye bolts.

And now comes the main event!

At last the 4 main castings have been test assembled. And I could not be more pleased, As you can see below the face of the cross-member casting that carries the X axis is square to the column mounting faces, without adjustment it was within about 1.5mm certainly less than 2mm. Using the alignment jacks it was simple to set it dead true against the Moore and Wright 600mm test square. This was only a rough test I will need a straight edge to extend the square. the columns are not bolted down they are just sitting there by their own weight. positioned by the 12x200mm threaded rods without nuts. There is no risk of falling the cross-member is bolted. Had I not used laser cut molds I don,t think I could have achieved the level of accuracy obtained.

The six alignment jacks are M10 x 30 Hex grub screws not seen but marked by zinc bolts inserted in the M10 threaded holes 2 above the columns one out of frame and 3 on the face and one marked by an Allen key.

Also I set up a linear rail on the right it is not bolted those holes have not yet been drilled and tapped, As you can see the cams securely hold it in position, once the machine is properly assembled and aligned the rails will be attached.


The Y Axis drive motor is fitted within the base. The motor plate is removable after removing the access panels. Both panels are 5mm mild steel.they also carry the Y axis limit switches and the Y axis ball screw bearing blocks. These panels were cast in position when the mold was poured with the CSK screws in position holding the to be cast in inserts. This guaranteed perfect alignment between the panel and the base, there is no movement possible when it is tightened down. This is important as the ball screw bearings have to be accurately located.

The X axis Drive is fitted behind the cross-member and through the column casting. Note the removable 5mm mild steel motor mounting plate. The M12 long nuts holding the cross-member are for convenience and will be replaced with nuts that fit within the counter-bore's.

The next step is to bolt the columns in position and prepare to attach the rails. Also there are many already made laser cut accessories paint and attach.

Regards
John

H all

Re the jack screws. These are only used to position and align the cross-member while it is being set up.
Once this stage is complete the small gap left will be filled with metal filled epoxy bearing material bonding the components together.

I will use my own formula for the metal filled epoxy grout.
link to power point file: **LINK**

Stefan Gotteswinter did this in an excellent video on you tube showing the method he used on his milling machine to improve its accuracy.
**LINK**

This machine will be set up in the same way.

I have seen a number of posts on the method, it is used by many commercial machine builders to assemble components.

Regards
John

You’re making very good progress John and the results are impressive. I can’t wait to see it all in action when it’s finished. Thanks for your marvellous documentation of the project and sharing of materials and links etc.

Hi All

Assembly of some of the laser cut parts, also a bit of machining.

The two rollers support the swarf curtain for the Y axis. Also note the 30 x 30 x 3 mm Aluminium angles bolted to the cast in place inserts in the base. These together with the top of the Linear ways form the support for this curtain. See the start of this thread for a 3D view.

The rollers are 40mm light steel tube with machined ends and 32 x 12 x 10 mm sealed ball bearings.

The laser cut 5mm mild steel brackets are bolted to cast in position inserts, no welding.

Laser cut control system 2.5mm mild steel box bolted to existing cast in position inserts.

Laser cut control system 2.5mm mild steel box bolted to existing cast in position inserts. (Not finished)
Note how it wraps around the cross-member casting.

More to come.........

Regards
John

Hi John, nice work!

This is exactly the size machine I would like to build.

I am interested in some of your thoughts, could one keep the steel cast bolted to the expoy granite as part of it? Is there an advantage or disadvantage either way?

Also I am wondering about the metal rails cast into the epoxy, could one just have the rails milled and ground by a machine shop after? or did you have a reasons for doing it this way?

Will you be using preloaded linear bearing blocks or standard?

Thanks

Ian

Hi Ian Baxter

Glad you like it,

I have been AWOL from the workshop for a couple of weeks. I feel guilty for not posting the next update I hope to rectify that over the weekend.

I chose to make a laser cut steel molds because they will be shared with a mate. Many large industrial machines, particularly when made from weldments are filled with epoxy/aggregate or Portland cement to damp out vibration.

The engineering lab at MIT led by Professor Alexander Slocum have done a lot of work on machine design**LINK** There are a number of free resources available for download Check out the FUNdaMentals series.

Also this Theseus paper by Eberhard Bamberg is well worth reading.
**LINK**

Yes a machine shop could mill the cast in rail supports. The rails themselves are **LINK** medium preload type 25. The machine base is large and heavy over 200 kg a large CNC or manual mill would be needed, Not many jobbing shops have a grinder with a big enough table, I did get a couple of quotes big milling machines charge out at several hundred dollars an hour.

I chose to use a surface plate to make the machine easier for a DIY builder to construct the machine. Surface plates come up regularly at machinery auctions at reasonable cost, the biggest issue is moving them. I just hired a small utility truck, they had a fork lift at the auction site (I made sure before bidding) I Unloaded it with an engine crane. The granite slab is 1000 by 630 x 150mm granite roughly 250kg.

Note a high speed 20,000 RPM spindle is shown in the illustrations, I have one so I will use it first, Good for aluminium not much good for steel even with small cutters as a number of videos on the internet show. I plan to replace it with a lower speed spindle for steel later. The machine is designed to be stiff enough for steel.

Regards
John

Thanks John.

Did you vibrate your mold? I see your from Australia, I am from NZ. Where did your get your granite aggregate?

I am also a mechanical engineer, I have no problem designing things, its the details about molding expoy and knowing the right way to do certain things I have not done before.

Regards

Ian

Hi John,

I am also wondering what motors you are using, steppers? Any encoders?

Thanks

Ian

Hi Ian

I do not have a vibrating table although I have seen a few you tube videos of one being used. The molds were placed on the granite surface plate to align the bearing mounts in the case of the base and main cross member I did not want those to be vibrated. Note there are holes in the mold, the bearing mounts sit directly on the granite only separated by a thin film to protect the granite from any epoxy leakage that would glue the whole setup together and damage the surface plate.

The granite aggregate came from a Victorian quarry. on the Mornington peninsula. The type of rock a quarry mines is location specific I just made a couple of phone calls and I found one that mined granite.
I have also used coarse silica sand and fine mixed gravel from a garden supply with epoxy It worked OK,

Here is a Google search that should give a good idea of the process and many PDF documents. Note how by searching for images it shows if a document will open, I find this useful.
**LINK**

I am using DC servo motors with integral 1000 count encoders. All the motors are located on laser cut steel plates to make it easy to swap I left plenty of room for the motors.

Regards
John

Edited By John McNamara on 01/03/2019 14:52:20

Hi John,

Thanks for the info. I have formulated a few ideas on my approach. I saw the video on you tube where epoxy composite was molded around a simple cast iron frame made from square. Then the cast iron machined for rails afterwards. I think this approach will suit me best, I will need to approach a few companies to see about machining.

I am currently pondering connection design between the y and x Axis frame. I also need to find a source for servo motors.

Regards

Ian

Hi All
15 Images follow

It’s been a while since I posted on the epoxy mill, Inspired by the accuracy of the castings made so far I decided not to use the low cost option Asian bearings and ball screws. Instead I will use Hiwin Taiwanese bearings these can be ordered to a proper engineering specification. I have asked suppliers to quote. This has imposed a delay. While waiting for the Hiwin bearings I have used some Bosch Rexroth bearings I have in stock for testing, these are earmarked for a CNC Router to be built in the future. They have exactly the same dimensions as the Hiwin bearings, so are interchangeable as long as you don’t mix manufacturers, the rail profiles are different. I have used them in the photos below. The router requires 2400 travel and I have two 3 metre lengths and some short lengths acquired as new second hand, hence the need to keep the stock bearings together as a set for use later. It would have been nice to use Bosch Rexroth Linear rails for the CNC Mill however they cost more than double the cost of Hiwin. While waiting for the linear rails and ball screws there is plenty to do.

My original plan was to only use the surface plate for flattening surfaces, I did this for the X and Y rail mounts as described earlier in his post it worked well however it was very slow progress. For the saddle I have used my vertical mill. It is a Shizouka VHRG, being a rather old machine the knee is a little worn. In order to get the best accuracy I had to use the overarm travel as the cross feed leaving the knee locked. The over arm never gets much use over the life of a machine so there is no wear at all. I also set some bolts in the tee slots and machined them using the overarm method. By doing this the only the table longitudinal travel and the over arm travel influences the cut. I am very pleased with the results.

The Z axis saddle casting has been done, This unit is a little different to the frame castings already made, it is more of a hybrid being made from 10mm laser cut steel plates and quite a number of drilled and tapped spacers that have a good fitting shoulders inserted into laser cut holes and glued in with epoxy making a very rigid unit. The spacers were very turned carefully to shoulder length.
This saddle unit fits into a laser cut steel mold that forms the side and the centre channel walls, it also locates several threaded inserts on the top of the unit. It will be cast this week

The saddle unit caries the Z axis, when thinking on the Z axis design I anguished for some time on the best way to set up the linear rails, should the linear bearings move or the rails move? Or to put it another way should I attach the linear bearings to the z axis and put the rails on the saddle or the other way around with the rails on the moving Z axis and the linear bearings on the saddle?
In the end I chose to attach the rails to the Z axis and put the linear bearings in the saddle.

With the saddle and the Z cast and machined at last they can be test fitted for the first time. It is exciting to see the actual components sitting on the bench having seen them on the computer screen for so long, now they have come to life.

Saddle / Z axis assembly

Inserts inserted

Ready for casting

Placed in Mold

Milling Face

Face milled

Bearing blocks test fit OK

Limit Switch housing

Z axis retracted

Z axis extended

limit switch housings fitted

Z axis drive motor plate attached

Set up for milling Milling bolt heads flat for perfect alignment with travel.

Edited By John McNamara on 23/03/2019 11:3

That is a very impressive build John. With all the embedded parts did you have a check list to ensure you did not leave anything out before pouring the epoxy mix. If it was me doing it I would be sure to miss something out and you would hear the scream in Australia from me in Scotland!

Doug

Thanks for the update, looking good. I like those linear rails, shame the Chinese don’t put them on their milling machines.

Hi Douglas
Thanks Vic

It is not the end of the world if you miss an insert, or have to add one later. Epoxy/aggregate behaves the same way as normal concrete, you can drill it with a masonry drill. A new steel insert can then be made and epoxied in place, this is done millions of times a day on building sites.

More difficult is machining epoxy/aggregate. It can be done with carbide tools but they will be blunted very quickly. The saddle photos show a lot of large holes as you can see they have been filled to the brim with material. For the saddle I only used washed sand as the aggregate, if I had used gravel it would not have been able to flow into the small spaces in the mold. epoxy sand alone is not that difficult to rough machine, Automotive body filler machines quite well. BTW body filler shrinks a lot on curing.

I used some automotive body filler to smooth the surface of the holes before machining them. The machining went well enough, I did a couple of roughing cuts then finish machined the surface with a .004" cut. Yes It did cost me 2 points on the triangular tool insert. The 25mm holder has provision for two tips however I have found for this operation one tip gives a better finish, I ran at 1220rpm 9 inches a minute feed, about .007" feed per rev. A bit on the fast side for me however I found slower spindle speeds produced a rougher finish. The mild steel material is a bit gummy.

Regards
John

Edited By John McNamara on 23/03/2019 23:16:38