[bhurts]

Thanks for the reply's and suggestions. I have several gallons of polyurethane from rencast that were given to me. It is a part a/b mixture and gets quite hot while setting. I have in the past experimented with making eg with it and it is very strong and rigid. It will stop a bullet from a 7mm without shattering. It will fracture a bit but stays together well in that situation. The reason I'm looking for a alternative is I had a bad experience with it the last time I used it. I mixed to large a batch and it got so hot it was curing to quickly. I abandoned my scoop and used my hands to fill the cavity before it got to rigid to use. It busted through the rubber gloves I was wearing and took a week to come off my hands. It was pretty bad. My thinking with the tightbond was if all else fails it cleans up with water. I guess I'll have to take my chances with the urethane again.

Thanks
Ben

Sent from my HTC6435LVW using Tapatalk

[ihavenofish]

Titebond is far too thick. You'd never be able to mix it with sand. If you tried, it would start to get even thicker than when you started. Also, it doesn't stick to aluminum very well.

You can get 3 quarts of epoxy from US Composites for about $45. Mix it with a bag of sand, and your all set.

titebond would set up before youd finished mixing. its also rather expensive - though so is epoxy I guess.

[phil c]

Depending on how serious your machine is, you could just vibrate/tamp small gravel or even 2B road gravel(it's been screened to get maximum packing). Fill the pieces and plug with a couple inches of filler mixed with epoxy.

Just having the mass of stone solidly packed in the frame will damp a lot lf vibration. Most important is using a heavy duty vibrator while tamping to get all the particles solidly packed.

You can also add an epoxy filled mix the same way, in small portions. it doesn't all have to be poured at once, like concrete.

[DaBit]

Why do so difficult? With throrough mixing of the EC mixture epoxy quantity can be brought down to 3% by weight. The resulting mix is very dry and has to be rammed in (with a stick, nothing fancy), but once cured it is still a solid mass and that will improve stiffness. Which you need to prevent the straight aluminium sections bending under their own weight.

So a single kilogram of epoxy, cost $20 or so, will give you 33kg/72lbs of epoxy concrete. In volume that's about 15 liters/4 gallons. In my book that means 'not expensive enough to risk my expensive aluminium sections by messing with polyurethane or wood glue'.

If this is still too expensive, then what's wrong with regular cement-based concrete? You can get that at any home improvement store for little money. It doesn't adhere to the aluminium that well, but why should it?

[bhurts]

I think you are thinking of polyurethane in the wrong way. I am not talking about the stuff from home depot or lowes that's used for wood finishing. It is a two part resin just like the epoxy you are using and is used to cast high quality molded parts. I have about 6 gallons of the stuff. 3 of a and 3 of b. I also have several different tints that can be added to color the parts. I have made eg in the past with it and it was very high quality and very dense.

Ben

Sent from my HTC6435LVW using Tapatalk

[DaBit]

The problem with polyurethane is that it is often quite viscous/thick in mixed state and that it sets fairly quickly.
That doesn't render it unuseable, but you do need more of it and work a lot faster.

[rowbare]

If this is still too expensive, then what's wrong with regular cement-based concrete? You can get that at any home improvement store for little money. It doesn't adhere to the aluminium that well, but why should it?

Cement based concrete is alkaline and reacts with aluminum so you shouldn't just pour cement in the extrusions without some kind of barrier. Also concrete shrinks as it dries unless you put in some additives.

bob

[DaBit]

That 'shrinkage problem' is blown up too much if you ask me, but my opinions often differ from common wisdom.
Maybe if you are making a CMM or similar ultra-high-accuracy device you might have problems with it, But as a filling for the average aluminium framed router? When mixed with the correct amount of water shrinkage is 0,05% or so after 28 days and although it continues to shrink it does so extremely slowly. Changes in temperature have more effect. There are also many mixtures with even lower shrinkage, and inside an aluminium extrusion the excess water doesn't evaporate.which lowers shrinkage too. I poured steel pipes with concrete a few times. Years later there is not a loose concrete core in them, but they are still solidly filled.

epoxy concrete shrinks too, by the way. That was how I got here in this topic.

Good point about the alkaline though. I don't think it is a real problem; who cares if the inside of the extrusion becomes corroded slightly. But honestly I don't know exactly.

[ihavenofish]

Good point about the alkaline though. I don't think it is a real problem; who cares if the inside of the extrusion becomes corroded slightly. But honestly I don't know exactly.

they are anodized, which makes them inert so it shouldn't be an issue.

shrinkage on concrete im not sure though. the extrusion is somewhat high tolerance, and the concrete may cause it to warp or cave in as it shrinks. yes, it may only be 3 or 4 thou over several weeks, but if this is having linear rails attached, you might be in for a nasty surprise.

youd have the same worry with epoxy of course, especially with the aluminium expanding and shrinking from the heat during curing.

believe it or not, ab foam is really good at dampening vibrations in an aluminium structure. not on the same level as granite of course, but a vast improvement over nothing.

[aguskov]

Wow, what a wonderful thread, so many interesting things are being discussed.
Anyway, I have not visited this forum for a long loooooong time pretty much since I build a little home brewed cnc machine and ran into a problem of backlash compensation. Being a total noob, I was super lucky to get an internship in an engineering company which developed direct drive precision rotary unit for custom application where I stayed and worked for the last years. Our units are used in a variety of applications: dispensing, electronic assembly, machining to name a few. I also have seen several machine utilizing polymer concrete frames, those were used for grinding.

First of all, I think P/G a wonderful material. But, because of the unique properties, it has very specific range of application. When talking about machining equipment, there is a number of factors required from the base material, vibration absorption, chemical resistance, thermal stability, compliance and toughness are of primary importance. P/G shines @ the first four surpassing steel and cast iron (probably similar performance in compliance); however, its toughness is a different story. Probably nothing new here but P/G is excellent in compression, OK in its shear strength and terrible in tension. It's also highly brittle and badly performs at perpendicular loads. Every design I've seen (a dozen of so) utilizes the same principle: P/G is used for the massive base, cast iron/ steel/aluminum for overhangs, consoles and other elements where the loads are primarily of torsional, tensional and bending nature. There is, of cause, an option to use pretension to overcome this issue but the level of engineering required for such endeavor is probably far beyond an average DIYer. Event professional machine builders try to avoid pretension since it greatly complicates the process.
From what I have seen, I am pretty sure that even with average performance composites what we can mix in the garage, a decent machine is possible, IF designed correctly. P/G would make an excellent base, but I would not build an entire machine out of it. Well... I would if the structure is very carefully designed, load optimized, most likely mono-poured and pre tensioned. But again, can we do it in our home shops?

Continuing the theme of reinforced P/G. There are reasons why the big guys don't do it. These are primarily the differences in thermal expansion rate and vibration absorption coefficient. There is also a problem of internal shrinkage and lamination which can be overcome by... Yes, pretensioning, but it's expensive and complicated. The bottom line is that it is very hard to design a monolithic structure (which would behave like a single block) from a number of materials with vastly different properties which would undergo thermal and dynamic stress and would not destroy itself.
On shrinkage, suppose I frame my base with rebar and fill it with P/G. The later would eventually shrink away from the rebar and it would become loose, even a little hurts a lot. That's lamination in a nutshell. I hear there are 0 shrinkage epoxies now, but are they truly 0 or just round up from 0.0-something.

So design is the big factor. A good place to start is determining you loads in both direction and magnitude. One company we worked with was making a small machine for cutting table legs. They had problems with deflection and began studying the kinematics by building a sensor jig which could determine loads affecting the spindle at different speeds and materials. Now that's a great place to start.

Hope it's not news to anybody, machine building is a very very complex area. If I was building a cnc machine, i would not strive for the result, I would try to enjoy the process, designing experiments, maybe building my own testing equipment or buying some if affordable, reading physics books and studying FEA. And who knows, maybe in several years you actually say to yourself: " Hey, I am actually building a machine comparable to some on the market. How awesome is that?"

Cheers, Andrey

[DaBit]

however, its toughness is a different story. Probably nothing new here but P/G is excellent in compression, OK in its shear strength and terrible in tension. It's also highly brittle and badly performs at perpendicular loads.

One could say it behaves like stone

P/G would make an excellent base, but I would not build an entire machine out of it. Well... I would if the structure is very carefully designed, load optimized, most likely mono-poured and pre tensioned.

In reality it is not that bad. I agree that epoxy concrete is not the most suitable material for the moving members by itself, but it does allow a degree of freedom of form which may come in very handy and it is quite easy to build large cross section members with it. It also doesn't resonate so easily.

Images say more than words. Recall my machine under construction with it's EC gantry. EC was chosen for freedom of form and the ability to cast on a surface plate:


I also have a 'play around' video online: New mill; playing around running G-code - YouTube
(sorry for the bad photographs and video)


If I put one ballscrew in servolock, remove the other, and apply 40N to the other side of the gantry to push it out of square then this is the out-of-squareness that results (over a distance of ~750mm):



That is not bad at all. Most gantries rock much more.

'limited stiffness', 'bad in tension and bending', etc. are all very true, but also very well solvable by adjusting the design, and it's freedom of form allows optimizations that well surpass the superiority of for example steel for a hobby guy in his shed with limited tools.

About the resonance: this is not a real test and more playing around, but this is what's happening when I hit the (partially constructed) steel Z-axis column with a hammer:


Resonance is measured using a 5um linear scale (I'm using linear scales in the feedback loop)

Doing the same with EC shows a flat line...

The later would eventually shrink away from the rebar and it would become loose, even a little hurts a lot. That's lamination in a nutshell. I hear there are 0 shrinkage epoxies now, but are they truly 0 or just round up from 0.0-something.

I noticed warping due to epoxy shrinkage, and I used as little as possible and a low shrinkage/low viscosity type.

However, you could use the 'fiber reinforcement' tactic. I threw a lot of wood screws in the mix.

The machine is not up and running yet, but all in all my measurement results are very promising...
At some point you just have to stop thinking and start doing. Epecially in the <0,1mm area there is a difference between the virtual world and the real world.

[gozos]

Very good work!!!

I have a few questions:

How heavy is the gantry?
What is the power of the servo? Model?
Pitch of ball screw and ratio of the belt pulley?

Many thanks!

[DaBit]

Gantry itself weights 80-90kg or so. That is based on the amount of raw materials used; I never weighted the finished gantry. Moving the gantry around is a 2-man job and the few times I had a second pair of hands available I rather used that extra pair for more useful things.

Servo's are old previous century Omron AC types on analog controlled OMNUC-U drives (these are in fact rebranded Yaskawa Sigma I drives). I managed to trade 3 steppers+drives for these with closed wallet.
I am running the drives in torque mode using LinuxCNC and Mesa 6i25+7i77 hardware to control them. I could not use velocity mode; I was unable to tune the velocity loop in the drive well enough; it was either a 'huge' following error or an instable motor. Bypassing the velocity loop in the servo drive and letting LinuxCNC handle all the PID stuff improved things a lot.
I still have to fine-tune those loops BTW, but I have to straighten the screws, mount 1um resolution scales on the Y-axes and complete the Z-column before I can do so. Torque mode tuning is quite tricky.

The Y axis/gantry is powered by 2x 400W motors running a 1:1.5 reduction to 10mm pitch ballscrews.
The X-axis uses a 750W AC servo with 1:1 reduction, 10mm pitch ballscrew.
Z uses a 5mm pitch ballscrew and a 4Nm stepper motor.

All ballscrews are 16mm diameter and bought from Chai/linearmotionbearings2008. The X and Y screws are using a variant of fixed-fixed mounting to increase stiffness by a factor of 4. These 'skinny' screws reduce rotational inertia a lot, improving dynamics. Accelerating/decelerating the rotating mass is what uses power from the axis drive rmotors. Accelerating the linear moving mass uses less power than one would think, and the milling forces use almost no power from the axis drive motors at all (at least not with my HF spindle).

And decent dynamics is what I was looking for. It is very well possible to put 15-24k rpm on a 8-12mm endmill in aluminium and produce a scary stream of chips, provided that you can keep up the feedrate.

All three servo motors provide way, way more power than needed BTW. Since I am running the drives in torque mode I can see the amount of torque required and even with 1G acceleration and 20 meters/minute top speed they don't even get close to their nominal output torque, let alone their three-times-nominal peak rating. 100W motors would have been fine too.

I see you are from Belgium. Here is a build thread in Dutch on the local CNC forum: ? Bekijk onderwerp - Wat was ookalweer het nadeel van graniet?

[lgalla]

also polyurethane is moisture sensitive.Will bubble

[Aaron1975]

First time post, but too interesting a subject to go by without having an input based upon the more recent advances (eg. Binci, Gemci & Kaplan, 2012; Binici, 2013).

The potential to use waste PET with Quartz sand for the home machine shop is huge, simply melt in whatever container you intend to fill, then mix or if the container has intricate shapes/webbing, mix then melt (ibid). High quality quartz sand/silica is a basic construction material and available virtually anywhere, while waste PET should hardly pose significant problems to source, so too an oven capable of melting PET (>200C: Binici, 2013).

The compressive strength of Quartz/PET was 32.15MPa while the control, normal cement was measured at 20.37MPa, thus the Quartz/PET was 1.5 times harder to compress than normal concrete (Binici, Gemci & Kaplan, 2012). It doesn't seem to be as tough as normal cement when taking into account flexibility and so forth, but as a reinforcing material in a laminated structure, that should not be too difficult to deal with. I especially like the potential given the ease with which this could be used - make a frame, or purchase a shell and then reinforce it with box section, then fill it with the mixture, heat it, let it cool, then do it again until it is full when cooled (or to the point where the shell can be trimmed to leave a flat surface on the base.

The heat/mix or mix/heat process should prevent voids/entrainment/degassing issues, while careful attention to vertical alignment of laminations could preclude a lot of issues relating to shrinkage. Finally, the fire/explosion hazard is much lower, there is no need for volatile resin monomer, MEKP or accelerants, nor for medium/heavy vacuum. If it works, and only one group of researchers is on this particular trail (or has published anyway), then this would be a much simpler route for the home workshop builder.

AS


Binici, H., Gemci, R. & Kaplan, H., 2012 'Physical and Mechanical Properties of Mortars without Cement' Construction & Building Materials, vol.28(1), pp.357-361 <PDF>

Binici, H., 2013 'Effect of Aggregate Type on Mortars without Cement' Eu. J. of Eng. & Tech, vol.1(1)<PDF]Page not found | Progressive Academic Publishing, UK>

[thegolfer]

Hey All,

I started doing some epoxy granite experiments yesterday and so far I'm going ok. Below is a shot of my first attempt.



I'm pretty happy with the result, except the open face of the block where the epoxy degassed.



Has anyone else had this problem? If so, it is possible to remedy it? I degassed the epoxy mixture on my homemade vibration table. As you can see, the degas essentially made a "foam" on top of the block that cured as such.

Apart from the foam face, I think my attempt was a good first step. The block feels cold like stone, sounds like stone when you tap it and has replicated every flaw from the wooden form I used. The block has a density of approximately 2250kg/m³, which is getting close to the same density as the epoxy granite manufactured by Epucrete in Germany. I would love to be able to measure the tensile/compressive strength and Young's modulus of this stuff!

[ckelloug]

The golfer,

Your block looks pretty nice. I'd suggest vacuum deairing the epoxy before mixing the material up. If possible, a few percent of BYK 525 deairing agent will also help. In about 30 formula tests, I never saw that problem although I always vacuum degassed and tended to use a smaller amount of epoxy and used a pretty thin epoxy.

Regards,
Cameron (ckelloug)

[ckelloug]

Aaron 1975,

That info on using PET as the bonding agent sounds interesting. A mix that can be dry blended and then melted has a lot of advantages in terms of capillary action working for you rather than against you however it all depends on the flow properties of PET which I don't know anything about. Does the literature suggest anything about the coefficient of thermal expansion of the overall mixture?

Regards,
Cameron (ckelloug)

[svenakela]

Thegolfer, what type of epoxy did you use?

[ihavenofish]

Hey All,

I started doing some epoxy granite experiments yesterday and so far I'm going ok. Below is a shot of my first attempt.



I'm pretty happy with the result, except the open face of the block where the epoxy degassed.



Has anyone else had this problem? If so, it is possible to remedy it? I degassed the epoxy mixture on my homemade vibration table. As you can see, the degas essentially made a "foam" on top of the block that cured as such.

Apart from the foam face, I think my attempt was a good first step. The block feels cold like stone, sounds like stone when you tap it and has replicated every flaw from the wooden form I used. The block has a density of approximately 2250kg/m³, which is getting close to the same density as the epoxy granite manufactured by Epucrete in Germany. I would love to be able to measure the tensile/compressive strength and Young's modulus of this stuff!

looks nice. a propane torch will get rid of most of the larger bubbles and foaminess. just run the torch gently over it after the pour.

to get a properly smooth bubble free surface (one you intend to use mechanically) I think vacuum degassing as mentioned is probably the best way.