[louieatienza]

understood, 0 flex means the whole surface must release at once so that could be a problem. You could cast in a few plastic wedges that stick proud from the side, give them a whack to get the delamination started? suran wrap is probably pretty consistent in thickness, stretch that over the plate, epoxy will weight it down flat, plate would fall off under its own weight, that is my educated experience guess.

jh

I wonder if threading some bolts in strategic locations, flush with the mold would help... They would be tightened down like manual "ejector" pins to release the mold.

[79TigerPilot]

Just have to cast in a place to feed shop air to the interface. Lots of square inches at 150 psi should do the release trick. If that is still not enough a hand hydraulic pump would.

Craig

[CaptainVee]

I wonder if air pressure would work.
Again there should be zero flex (theoretically) so the surface can not peel away, so the whole surface on the plate has to release all at once.

Once released there should be microscopic spaces between the plate and the casting, allowing air to go everywhere but after casting those spaced are probably filled with epoxy.

Hydraulics might work but may require a lot of pressure as it again has to peel a very rigid piece.

I guess some kind of foil would be my choice to start with.

Thanks for the brain exercise guys!

[fernando123]

How strong is epoxy granite, and where do I find a good quality unexpansive source?

Enviado de meu SM-G530H usando Tapatalk

[rocketflier]

I wonder if air pressure would work.
Again there should be zero flex (theoretically) so the surface can not peel away, so the whole surface on the plate has to release all at once.

Once released there should be microscopic spaces between the plate and the casting, allowing air to go everywhere but after casting those spaced are probably filled with epoxy.

Hydraulics might work but may require a lot of pressure as it again has to peel a very rigid piece.

I guess some kind of foil would be my choice to start with.

Thanks for the brain exercise guys!

I have seen that in fiberglass molding videos, which I can't seem to find. They simply had a hole drilled in the mold and on face side, they put a bit of clay in the hole.

But mold prep is key here. Even large boat hulls are able to be separated simply with wedges. Watch the draft angles, ultra smooth and sealed mold face with multiple layers (6-10) of wax and PVA.

Where the mold pieces come together, you could slightly bevel the edge of the mating board for wedge placement/starting.

[ger21]

Even large boat hulls are able to be separated simply with wedges.

The difference is that boat hulls can flex a considerable amount.

[LeeWay]

Design wedge keys into the mold. When ready, just tap on the wedges.

[CaptainVee]

I used some epoxy slurry for undercasting the feet of my gantry a few years back, with a slab of granite as a table.

I had a very hard time getting it to release, even though the release wax had not given me any trouble with releasing epoxy before.

It did release in the end, but that gantry was made of steel so I could whack it with a mallet. That made a dinging sound so there was some vibration (=flex) there.

If you hit a piece of EG with a mallet you should hear "B". The sound of the mallet vibrating

I have done quite a bit of casting through the years, in various materials. Thinking back it were always the non-flexing things that gave me trouble.

[petruscad]

Casting on an already surface plate would surely simplify the process. However I have a couple general considerations which mean I would never actually do that; here they are:

1. If I had a granite plate precise enough and large enough to cast this mill's base on, I would want to keep it and not risk destroying it with epoxy
2. Mixing and casting epoxy granite is a very messy process. No matter how careful you are at the end there are tiny spots of epoxy everywhere. I was never able to cast without making at least some sort of mess.
3. Which means, back to point 1, that I would probably screw up the surface plate, which is what I would not want to do! And it does not matter if the whole plate is covered in plastic wrap, it will surely be torn off and epoxy will reach the surface. One dried spec of epoxy and the surface is screwed.
4. Granite surface plates have tiny holes/depressions in them, even when new (I believe they are caused by the cutting of the granite block and natural imperfections of the stone). The epoxy granite will tend to flow down is those depressions reducing the precision of the casting surface.

[LeeWay]

Put a piece of glass on top of the surface plate.

[the4thseal]

the depressions are from the scraping prosses. They are usually talkied about terms of points per inch and area of bearing This is not a lack of accuracy, but a means of defining qualifying or accuracy. Since a "plane" or refrence is defined interms of points that are coplainer. Knowing the points and area give you needed information for transfering these peramiters to another surface. Depending on the nature of the reference the actual composition of the stone the crystiline structure is taken advangae of in terms of durability. This is not the case with a metal surface plate. Just FYI.

[gary_808]

the depressions are from the scraping prosses. They are usually talkied about terms of points per inch and area of bearing This is not a lack of accuracy, but a means of defining qualifying or accuracy. Since a "plane" or refrence is defined interms of points that are coplainer. Knowing the points and area give you needed information for transfering these peramiters to another surface. Depending on the nature of the reference the actual composition of the stone the crystiline structure is taken advangae of in terms of durability. This is not the case with a metal surface plate. Just FYI.

But you don't scrape granite...

Fine adjustment is done with lapping.

[the4thseal]

actually you do scrape granite. lapping is a way of flattening a surface, however when they come to calibrate and certify the scrape the high points. if you do a search for scraping surface plate you will run into the scraping rabbit whole. lapping gets you only so far. sometimes that is far enough, having said that if your surface plate has been lapped the structure of the material will not erode at the same rate. for example a rose quarts surface is sought after for that very reason, the quarts crystals are harder than the matrix, giving longer life.

[gary_808]

I only build granite based machines from small radius grinders for grinding carbide die parts, to 30+ tonne grinders, calibrate and adjust our granite inspection tables,

And not once have we scraped any of them. Machined and lapped.

Ps I can scrape, i'm pretty good at it as well if I'm honest,

25 foot base V and flat, scraped to with in 7.5 microns straightness, and 3 microns twist.
With points per square well within rolls royce spec.. (highest spec we build them to)

Most machines are just ground and mottled these days, and we are they boys that have to sort these out on site.

[the4thseal]

that is fantastic, i am sure you are right about the state of the art. the lab that i do ME R&D work at has our references scraped, not a whole of scraping by the way. A great of optical test equipment and good bs over lunch.. it also might be because of the locaion of the surfaces. out of curiosity how do you lap surfaces that are used for spotting in the field that wear unevenly, vs lay out ? curious, always on the look out for more info.

[gary_808]

If I'm taking your question correctly,
We don't use any granite surface plates etc out in the field for blue rubs etc, just to heavy bulky and clumsy,
Have a selection of cast iron fish backs and surface plates,
Which are scraped/checked to the master torpedo or table in the shop after every job.

[rocketflier]

If I'm taking your question correctly,
We don't use any granite surface plates etc out in the field for blue rubs etc, just to heavy bulky and clumsy,
Have a selection of cast iron fish backs and surface plates,
Which are scraped/checked to the master torpedo or table in the shop after every job.

At the risk of further derailing petruscad's awesome build thread, what are your thoughts on the mold method(s) given the skill level and tooling availability of many (most?) of us are likely quite a bit less than your own?

1) Using a surface reference covered by thin plastic.
2) Using steel inserts (either machined or threaded inserts) which stand proud of the granite and can be milled, scraped or filed after the pour.
3) Bare epoxy granite and any grinding/lapping needed.
4) Inserts or bare epoxy and post filling with epoxy grout (after rails are aligned).
5) other?

[petruscad]

With the mounting surfaces flat it is time to install the linear rails. I located the bolts embedded inside the concrete with a long pendulum with a tiny neodymium magnet attached to a string. The magnet hovers over the surface at maybe 0.5mm and gets strongly attracted the the bolt. The two end holes were drilled and tapped first so that the rail could be secured in place and the remaining holes located. The drilling was done with a hand drill which was fixed in a guide allowing it to be moved only vertically. This picture shows the setup for locating and drilling a hole.

Attachment 290302

First I drilled through the ~5mm of epoxy granite to get to the embedded steel bolts. I used a 7mm diamond coated hollow drill bit for this, like the one below.

Attachment 290304

After that the steel bolt was drilled and tapped for 1/4-20 bolts. With both rails mounted the next step was to make them straight and parallel to each other. For that I suspended the large ceramic straight edge on three supports, installed one carriage on each rail and a 0.0001" dial indicator on each, like so:

Attachment 290310

After a lot of fussing around and trying different ways of straightening them I found what I think is the simplest method:

1. Bolt one of the rails and set the straight edge so the indicator shows zero at the end of the rails. Don't touch the straight edge anymore!!!
2. Loosen all bolts except one bolt at one of the ends.
3. Start tightening the bolts from that end towards the other end bending the rail by hand while tightening every bolt to get the indicator to read zero.
4. Without touching the straight edge repeat step 3 for the other rail.

By following this process both rails can be adjusted and tightened in about 30min with an error of 0.0002". It's surprising how well the bending by hand works: since the rails is fixed at one end and floating at the other you have a lot of leverage to bend it by the 5-10 thousands that are needed.

[Hezz]

Absolutely fantastic thread. Thanks for posting up your work Petruscad. I'm looking forward to your further work progress. The information in this thread will help me to refine my DIY CNC designs. Although I can see the challenges of getting the EG surface really flat and stable for the mounting of the linear rails there should be one really good benefit of enhanced vibration reduction by having the rails mounted so intimately with the EG.

Paul H.

[petruscad]

Since the last post I made the saddle for the mill. I basically followed the same procedure as for the base: created a mold out of particle boards, an internal steel skeleton, anchors for linear rails mounting, etc. Here's the mold ready for casting.

Attachment 292712

The saddle is cast upside down so what you see here is the bottom. The three blue tape covered areas are large steel plates for mounting the saddle to the the X-axis linear rail carriages and the one in the middle is for the X ball screw mounting block. You can't see it but the two side plates are already drilled and tapped and aligned to fit the carriage blocks exactly. The X-ballscrew mounting plate will be drilled and tapped later. Inside the mold you can also see two smaller plates. They will stick out on the other side, in the middle of the saddle, and the Y-axis ballscrew will be mounted on them. Here's a side shot showing the vibrator underneath and the way the mold is suspended on foam so vibrations don't get damped away.

Attachment 292714

And here is the end result after casting! You can see the Y-screw mounting pads much better now.

Attachment 292718

Just for fun I placed the saddle on the base and placed the Y rails and ball screw in their places to get an idea of how it will look... not bad!

Attachment 292720

Now the tedious process of grinding the saddle flat and parallel on both sides...