Casting Machine Bases in Composite Epoxy

When you’re building a machine that needs to be accurate, you need to give it a nice solid base. A good base can lend strength to the machine to ensure its motions are accurate, as well as aid in damping vibrations that would impede performance. The problem is, it can be difficult to find a material that is both stiff and strong, and also a good damper of vibrations. Steel? Very stiff, very strong, terrible damper. Rubber? Great damper, strength leaves something to be desired. [Adam Bender] wanted to something strong that also damped vibrations, so developed a composite epoxy machine base.

[Adam] first takes us through the theory, referring to a graph of common materials showing loss coefficient plotted against stiffness. Once the theory is understood, [Adam] sets out to create a composite material with the best of both worlds – combining an aluminium base for stiffness and strength, with epoxy composite as a damper. It’s here where [Adam] begins experimenting, mixing the epoxy with sand, gravel, iron oxide and dyes, trying to find a mixture that casts easily with a good surface finish and minimum porosity.

With a mixture chosen, it’s then a matter of assembling the final mould, coating with release agent, and pouring in the mixture. The final result is impressive and a testament to [Adam]’s experimental process.

We’ve seen similar builds before — like this precision CNC built with epoxy granite — but detail in the documentation here is phenomenal.

60 thoughts on “Casting Machine Bases in Composite Epoxy

  1. It’s a nice idea – but what was the resulting damping and stiffness? Is it any better than other materials, objectively?

    I’m pretty sure that “Grey Iron” is the standard for machine bases, which is very stiff, and has a much higher damping than other metals (factor of 100 better than steel). This material is notably missing from the chart.

    Still, casting epoxy is a lot easier than casting grey iron for the home user, so it’s a nice idea, but I’d like to see if it’s truly better as claimed.

    1. Agreed. I was a little disappointed to see that the “experimentation” seemed focused on just getting the material to look nice without voids, with apparently no measurements taken. Is “grey iron” significantly different from cast iron? Because cast iron IS in the chart, with about 100 x the stiffness of epoxy and around 1/10 the loss factor. Concrete and stone fall between the two.

        1. Thanks. I guess I could’ve just looked it up. Wikipedia says, “Gray iron, or grey cast iron, is a type of cast iron that has a graphitic microstructure. It is named after the gray color of the fracture it forms, which is due to the presence of graphite. It is the most common cast iron and the most widely used cast material based on weight.”

          Being that it’s the most common type of cast iron, I would guess that it falls somewhere in the oval labeled “cast iron” on the chart.

          Going on with my guesses, and based on why _I_ would be motivated to use a material like epoxy/stone composite, I’d say the main reason for choosing it was that it’s considerably easier than cast iron to work with in a home shop.

          The chart in the article is very interesting, though: according to it, wood or concrete might be a better choice. But everybody “knows” that a machine made out of wood or concrete is crap, so instead you use a material not everybody is familiar with…

          1. I’m not sure about that. I think I saw an article on HAD many years ago about old-school concrete lathes that were actually quite good – but they’re a right bastard to move in any way.

          2. That may be so, but anytime you see an article about machines made out of concrete – and I remember this specifically about that article – there are experts chiming in from all points of the compass in the comments about how poor a choice of materials this was.

          3. There’s a guy on YouTube who makes all of his shop tools out of wood. Two or three handsaw models for axample. The takeaway is that you CAN build incredibly good machines this way (better than many manufactured machines) but that they do take a lot of patience to dampen vibrations, and need to be kept on clean dry shops.
            But hey, the industrial revolution was bootstrapped on wood machines to build the steel machines.

          4. At first I thought you were talking about Matthias Wandel, but I never saw him build a hand saw out of wood! Yes, he is a genius at engineering with wood, but so far I haven’t seen him build anything out of wood that’s made to work metal.

          5. Search “concrete lathe” on Youtube. So far, one guy made one that I can find. He’s using an offshoot of the design by Pat Delany. over on Open Source Machine Tools The idea goes back to WW1, and honestly should have been the most disruptive technology of the era because it made machine tools literally cheap enough to simply abandon. The ways were simple polished round stock. The bed, headstock, and tailstock were one casting, left to cure for a month, then ways, spindle, etc. were jigged up like 27 different times and secured with molten babbit metal. Instead of months of skilled labor to create a big lathe which then had to be transported and then setup, it took a crew of semi-skilled men eight hours total to finish out a concrete casting that could then be used as is. The catch is that they literally weren’t worth reselling, instead just busted up and used as fill.

            Honestly I think the epoxycrete might be an easier material to work with, but concrete is probably cheaper. Concrete would allow a hobbyist the ability to make a 10in swing lathe with 40in between centers for hundreds of dollars. Have my own planned for after the move in a few months.

      1. It isn’t the epoxy that provides the stiffness nor the hardness of a polymer concrete mix. It is the binder and is part of the reason polymer concrete can have an order of magnitude better damping properties of grey cast iron.

        Just using cast epoxy would be a bad idea for almost any purpose.

    2. I know at least one precision machining outfit used filled epoxy over the previous welded steel structures. They produce limited edition custom machines that would not be economical to cast. The bases were about a 4 foot cube overall. The deal with machine stiffness is that it is also controlled by section properties. A material with 1/4 the bulk stiffness can be equivalent with 4 times the section, for example.

      Ultimately a material that one can obtain and use is much better than any theoretical substitute that is unobtainable.

      1. Meehanite is a trade name for grey cast iron with some guarantees of the properties. Nothing more, nothing less.

        Nowadays ordinary grey cast iron have similar properties even without the name (in general) – metallurgy have improved and licensing a name adds complications and expense.

    3. Epoxy Granite or “Mineral Castings” are by superior to cast iron, and ultra high accuracy machines use it extensively. It is much more expensive than cast iron due to the vibration/vacuum and other processes required to achieve optimal mechanical properties. Epoxy granite is less stiff than cast iron, but you can use a huge amount of it and the damping properties are unmatched. It also has very low thermal expansion.

      Its primary application is in grinding/micromachining type stuff but it shows up in large machines as well.

    4. It is better when done right, that’s the reason polymer concrete/epoxy granite (and a lot of other names for the same stuff) is used for high end machines today. Cast iron is relatively inexpensive but takes a lot of effort to get right (melting, casting, aging etc.) while epoxy granite is relatively expensive but a lot easier to work with. Another advantage are that one can add additional parts, reinforcement etc. into the mold before adding the epoxy mix (as illustrated in the build of the article where an aluminum frame is used) reducing machining and making it possible to get a much better properties than a monolithic iron structure.

    5. Author here,

      True, cast iron has a higher Young’s modulus. However, by simply making the epoxy granite walls significantly thicker, a rigid frame can be constructed. Deflection is driven by the moment of inertia, which in itself is a third power of the thickness of the member that is being deflected. So by doubling the wall thickness, in effect you are increasing the stiffness of the frame by a factor of 8.

      As well, cast iron is a pain to work with. I do not have a forge, or the ability to cast the iron, never mind the scraping that would be required to bring the surfaces to a true flatness after forging/casting.

      Epoxy granite on the other hand is incredibly easy to work with (as seen from my article). Knowing and understanding the reduced Young’s Modulus allows the design of the frame to easily compensate for this.

      Thanks for reading!

  2. Aluminum shell with rubber inside works great… Don’t believe it? Go to any drag strip and look for ATI pulley dampeners.

    Let’s try not to solve the same problem more than once.

    1. It’s a whole different problem. Crankshaft harmonic dampers are designed to absorb torsional vibration, with stiffness being a very minor factor. Machine bases actually need some stiffness.

      Let’s try not to jump to the same solution for every problem.

      1. Implying that motors turning and causing chatter in the platform is not actually a torsional vibration. I would hope I wouldn’t have to point out the error in that. We are talking harmonics here, doesn’t matter the source, all harmonic vibrations behave in the same fashion. Harmonic dampeners are virtually universal. Give the energy something to vibrate, and make sure that something has enough mass, and you’ll be fine.

        1. All Harmonic vibrations DO NOT behave in the same fashion. Harmonic dampeners are not virtually universal.

          ATI dampers operate on the rotating mass where as the mass for cnc machines is in the base. If all harmonic dampeners are virtually universal, then why don’t cnc machines have some sort of ATI dampener at the other end of the spindle, and then you wouldn’t need that giant honking mass of gray iron that they sit on. (by your logic at-least)

          Feel free to prove me wrong, but please use citations.

          1. Citations? You mean like the ones you provided?

            Harmonics are all the same. It’s an oscillating vibration. That’s it.

            CNC Machines, as in industrial grade CNC machines have a heavy pedestal to negate vibrations, because no customer at that level would put up with not having something sturdy to do. It’s cheaper to pour a cast iron part than to mold and machine a space for a rubber ballast. It still does the same exact thing, provides enough mass to absorb the energy being released that causes the oscillations. On a smaller machine, rubber works just fine.

            Just to point it out, because it’s apparent that you don’t really understand what crank dampeners do, they do not just reduce the harmonics of the crank. They reduce the harmonics of the explosions going off in those cylinders. They reduce the thrusting forces and vibrations associated with the pistons changing directions multiple times a second. They reduce the vibrations of the cams and their lobes spinning and oscillating, as well as the pressure of the valves opening and closing.

            Engines are the harshest engineered environments that we have. If plain black synthetic rubber is good for an engine, it’s good for your dorky tabletop CNC device.

            Harmonics are harmonics. The source doesn’t change the method of mitigating them, which is increasing the mass of the device.

          2. internal cylinder combustion engines are so far from the harshest engineering environment it isn’t even funny.
            liquid rocket engines?
            turbines engines running at speeds and temperatures where even the best normal steels out there would behave like wax, it requires steel made from special alloys and grown so that the entire part is one single individual metal crystal.

            what about the large research machines, some of them produce temperatures that make the sun look like an icecube.

            in short to say that internal combustion engines are even close to the harshest engineering environment is akin to saying that an old banged up beetle is the most sophisticated car, it has never been and will never be.

          3. ill leave my reply here because for some reason i cant reply directly to you..

            First of all, thanks for making it blazingly obvious that you are a troll. but let me explain why..

            “Citations? You mean like the ones you provided?”

            I never made any claims that needed citations to back up.. Quite simply i stated that all harmonics are not the same and all dampers are not the same either, thats pretty simple to figure out considering harmonics is a field that people can and do specialize in, you are right in that i don’t know much about it but i have had the pleasure of meeting people who have specialized in the field. To be so flippant as to say that all of it is the same seems very rude and inconsiderate to those people and to be frank, if it was the same then why would there be a need for any specialization in it.

            the rest is just you being dismissive and rude, you make grand claims of things that you don’t seem to know much about.

            Harmonic dampers such as the ATI ones that you have referenced have nothing to do with the valves or any part of the valve-train. Thanks for playing the troll.

            Also just a FYI harmonics isn’t just about increasing the mass, at least i know that. Its the mass, spring rate and dampening force, which once put together (due to being a differential equation) with a Fourier transformation can give you the harmonic frequencies.

  3. Yes, that’s why we’re still using spark gap radio transmitters, arc lamps, and mechanical calculators, because once a problem is solved it’s a complete waste of time to try to solve it in a different and maybe better way.

    1. So you’re saying that multiple pouring expensive composites to find the right mix with other materials is cheaper and easier to manufacture than a well measured sizing of a box filled with common generic synthetic rubber…

      Yeah, well, if your CNC machine vibrates that much, there is something else wrong with it.

      Let’s not make this into a problem, when it really isn’t. It’s not like a change in materials is going to reduce the mass required to dampen the specific vibrations, and if anything adding materials will focus the effectiveness to a specific frequency over all others, which doesn’t solve anything to any real degree, as no multi-motor machine is going to have just ONE frequency of vibration. Making your specific dampener less effective than a general rubber filled box.

      1. Good luck getting any kind of precision or repeatability out of your rubber-filled box. As I said before, stiffness is more important than damping (that’s the term, not “dampening” – nothing’s getting wet here) in machine tools.

        1. Ah, come on.
          It’d be fun trying to work out what tool path speed vs. spindle rpm vs. cutter diameter vs. material vs. path curve radius vs. temperature results in the greatest precision. Bet you could develop combinations for creating different surface textures (it’s a feature). The varying surface textures within a piece means it’s a unique artisanal item. You might even invent a new musical instrument.

      2. “It’s not like a change in materials is going to reduce the mass required to dampen the specific vibrations,”

        have you by chance driven over a working suspension bridge before? have you heard of the tacoma narrows bridge? that had lots of mass, still had a harmonic vibration problem. mass != dampening .

    2. Moglice epoxy is still used for truing knee-mill castings, and even rebuilding worn ways… Its been around since the 1970’s … And note, most large castings and frames have dampening cavities filled with soft-lead, pellets, or sand.

      There are some great alloys around these days, but most raw manufacturing was exported overseas in the 1990’s…
      Due to the recycling programs, it is difficult to find raw material in small quantities at a reasonable price (ebay is ridiculous). While the expectation of precision was lost somewhere during this trend.. People now expect equipment will barely function for long, break in some unforeseeable manner, or simply prove uneconomical to operate.

      Young engineers who have never visited China, have no idea what a large factory running German CNC machines looks like. There is nothing that massive in America, as it could never compete in the modern economy.

      1. I mean, I worked in manufacturing in Australia and we had a factory running something like 30+ Japanese CNC machines, metrology in house, etc.

        I am almost certain this exists in the USA on a large scale. Perhaps you mean on an even greater scale though.

      2. I know some of the US still has German CNC machines, or American, even Swiss machines. They don’t compete in the modern economy, they dominate the niches they occupy, commanding higher prices, because their product is more than worth it.

        I can forgive you for the mistake, as these places tend not to open their doors very often. I’ve been lucky enough to work with a number of these places, to the point where they outnumber the cheap crap factories I’ve worked with.

        I can’t comment on the general state of this though, as I’ve done mostly stuff meant for industrial or aerospace applications, or high performing(quality wise) rust belt factories.

        1. Hardinge bought Bridgeport then built a new factory to manufacture the knee mills. They were able to apply modern manufacturing processes to produce about 3/4ths the peak volume of the old factory in 1/3 the space, and said they should be able to ramp up production to equal what Bridgeport produced at their peak. But being a Hardinge product means they’re bleeping expensive.

          Unfortunately the Bridgeport knee mill is still pretty much identical to the design that was cast in iron not long after the “J” head was first introduced. It works but could definitely use some modernization and using FEA to improve the design. It’s likely that the castings could be altered to have better vibration characteristics without increasing mass, or even decreasing it some. The mechanisms in the head could possibly be made less complicated. The system that runs the mechanical power down-feed would make Rube Goldberg and Heath Robinson proud.

          At one time, Bridgeport did make a one piece head and support arm specifically for CNC models. For some reason they quit that and did what all the Bridgeport clone makers do, graft a servo or stepper motor onto the J head and drive the quill via the point where the downfeed stop bolts on. Bridgeport’s CNC J heads generally don’t have all the manual and mechanical feed parts installed, but they still use fully machined heads. The clones most often take “fully stuffed” heads and remove a few parts to add the CNC setup.

    3. @localroger: You are trying to be sarcastic, but each of these inventions solved their particular problem. If some solution is good enough and cheap, typically there will be no reinvention worth a mention. New inventions we use now solved other, additional problems, that were introduced later or which were not anticipated earlier.

      For instance we don’t use spark gap transmitters any more, but we didn’t reinvent crucial parts of them: the LC tank circuit, and we still use resonant antennas.

      Arc lamps are still used where needed (see: xenon arc lamps)

      Mechanical calculators ruled market supreme decades into the age of electronic computers, because electronic computing didn’t reap the fruits of miniaturisation and automated mass production for quite a long time. Complex mechanical automatons are a lost art today.

  4. What about the heat generated while curing epoxy? Is it reduced (diluted) with the sand?

    I have the experience that even small leftover batches of epoxy tend to get very hot or even melt the mixing canister.

    1. There is a long thread at cnczone about resin casting I read a few years ago and heat was one of the issues discussed. I don’t remember the details but iirc you shouldn’t use to thick castings and do the pouring in low ambient temperature. I think the ballast fractions made a difference too. A good ratio minimized the amount of resin used and therefore kept the temperature down.

      I want to try this with crusher run some day. The quarry has taken care of the fraction ratio for you.

    2. I used a 48 hour cure epoxy, which generated almost zero heat during the cure. The curing of epoxy is an exothermic reaction, however, when you spread this heat generation over 48 hours, the actual heat generated is easily wicked away by the surround ambient air.

      A low/zero heat cure is VERY important, as a resin which generates a lot of heat during curing will cause the mold to twist and distort.

  5. I wonder if the author would comment on why he’s using this as a base but not a precision surface? Most of my research indicates this is a fantastic way to get a dead flat (< .001") surface with non-precision equipment. It looks like he cut out the center and put an aluminum bed in…

    Also totally overkill for his application, but that's part of the fun, isn't it?

    Also, I've done a lot of internet reading on this, for keyword searches, also include "synthetic concrete", "epoxy granite", "synthetic granite", etc. Haven't bit the bullet yet, but looking to do this.

    1. Author here,

      I didn’t use the epoxy granite as a precision surface since I didn’t have a 36×24 granite surface plate to cast this on top of. Although that would have been super cool if I did have one. As well, if I did cast the top surface as just epoxy granite, I would have had to embed many threaded inserted to attached the linear stages.

      Casting in machined aluminum with tapped holes proved to be a simpler solution. The aluminum pieces were precision machined to the exact size and flatness required to the application.

      Maybe for the next one…

      And yes, this was quite overkill :)


    1. I’m just guessing here, I believe it should be a match between the ballast and the metal attached to the casting. A cross section of the cast should be mostly filler material surrounded by very little epoxy.

    2. Author here,

      Epoxy by itself as a thermal expansion coefficient that is quite a bit higher than aluminum. However, the embedded sand has an expansion coefficient that is quite a bit lower than aluminum. Overall, the composite expansion coefficient is most likely still higher than the aluminum skeleton.

      With that being said, this device was designed to sit in a metrology room, where the temperature is closely monitored and maintained at 20C.

      A 48 cure epoxy was chosen for this base, as that resulted in almost zero heat generation during cure, which would minimize any potential error that would come during casting.


  6. Light Machines (later bought by Intelitek) made a line of CNC milling machines with a one piece epoxy granite base and column. Look up the PLM 1000 and PLM 2000. The 1000 has stepper motors and uses a large, separate box for the motor drivers and a proprietary ISA or PCI card for the control computer. There’s DOS and Windows software available for the 1000.

    The 2000 uses servo motors and its control system is self contained in a box on the back of the column. It only needs an RS232C connection to a control computer. Unfortunately the only software made to control that model runs in DOS, and it uses EMS. XMS need not apply for that job. All the G-Code processing is done by the servo controller on the mill. The PC can be as old as an IBM Model 5150 PC, as long as you have it setup with enough EMS memory to hold the G-Code, or use the splitter program to chunk it into linked pieces small enough to fin into available RAM.

    I was able to obtain a bunch of technical info and software for the Animatics servo controller. I’ve posted it here and there and elsewhere so that people with much more programming wizard ability than I may Do Something with it. As yet, no fancy new software for operating a PLM 2000 has surfaced.

  7. I’ve seen a lot of interesting comments here and a lot of argument too. I can’t remember the last time I’ve seen so much argument on a post.

    This build is really incredible, and it looks like the machine he built is very similar in design to a CMM. I have been mulling over for many years building my own hyper precise desktop micro jig borer for micromechanics prototypes, and it looks like this might work for the main machine frame.

    For the record, there are many US companies using plenty of high precision german, swiss, italian, japanese machines. They aren’t your normal job shops though. Plenty of super high end manufacturing being done in the US still using hyper precise machines- I’m really suprised so many people don’t think so. Aerospace, medical, nuclear parts shops use a lot of super high end stuff.

    1. One of the biggest advantages of epoxy granite is that you can cast against a precision surface plate, or apply an extremely expensive self-leveling epoxy to achieve surfaces that are flat and parallel at the sub-micron level.

      His mates are just plain aluminum bar stock so its essentially a low quality frame inside a block of epoxy.

      1. I was thinking of this, using precision ground steels instead of aluminum for pertinent parts. Wonder if that would be worse in the dampening aspects though. Where can you find the self leveling versions of this stuff he uses? I don’t know a lot about these types of epoxies, but I’d want to use the very best, to hell with cost. I’m curious to see what the cost of the very best out there is.

        To the people I noticed complaining that this guy did overkill- have YOU ever built a machine from scratch? Some people are looking for exactly that- the best possible, cost be damned. Overkill is not something they care about. They want overkill. I am one of those people. There are many who want to do more than even metal milling at home- some want metrological levels of accuracy in their equipment for their own purposes. Not everyone is ok with tolerances some consider “good enough” for most things.

        A jig borer, a true jig borer, is a good example of that. Small ones that can fit on a desktop for instrument makers and watchmakers are very rare, and ridiculously expensive. There is a use for things like this epoxy base if it is the best for small one off machines that need super accuracy and dampening.

        I’ve entertained the idea of 3D printing a machine frame in PLA, and then having a foundry lost PLA cast a grey iron machine base for this jig borer, but that could be far more complicated than going the epoxy route, which would allow the nice aspects of composite frame to be easily done.

        Yeah, maybe people can argue over the need for stuff like this- but there is always someone like me needing things like this for niche uses that are nigh impossible or unaffordable otherwise.

    2. Author here,

      Thank you for the kind words. This is exactly what was being built, and automated vision measurement system. A rock solid frame (high stiffness and damping) was needed to reduce settling time of the camera to an absolute minimum.

      With a 30x magnification camera on the end effector, and movement or play at all is immediately visible, so the epoxy granite + aluminum skeleton provided the answer.

      The idea was inspired by the ultra high precision CNC machines (and CMMs) that are build on a similar principle. Given the amount of work that went into getting the mixture and process right, it seemed worthwhile to post the information for others, so they can skip right to the good stuff on their build!


  8. First, amazing job! Ignore the trolls and haters.
    What sort of finishing work is done post-cast? Is a sealant of some sort used? In the concrete turntable posted a while bac k ( ) the maker included voids filled with BB’s or shot. I recently took apart a HDD and found a small capsule on the arm that appeared to be filled with sand. I assume that capsule was there to damp vibrations. Would a similar approach would work here? Lastly, after all the work that went into creating this amazing base – what goes under the base? Does the epoxy sit directly on a table? Are there cork or rubber feet under the epoxy? Three feet or four? Feet on threaded leveling bolts? Again, thanks for sharing!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s