[ckelloug]

Hi PoppaBear10,

Glad you're joining us. The good news is that the aggregate formula is pretty much decided. The bad news is that we're still working on the epoxy.

For aggregate, we've come up with this this is probably fairly close to optimum: see <A href=http://www.cnczone.com/forums/showpost.php?p=371354&postcount=2413> post 2413</A> for where this was originally suggested.


1 part #6 Agsco Brown Aluminum Oxide
1 part #4 Agsco Quartz
1 part #2 Agsco Quartz
1 part #2/0 Agsco Quartz
1 part 3M G800 Zeeospheres
1 parts 3M G200 Zeeospheres

(www.agsco.com for the aggregates listed and somebody like the Cary Company in Chicago for Zeeospheres)

For epoxy, so far Reichhold 37-127 and 37-606 have been tried by several folks. (Available from US Composites as Thin epoxy with the slow hardener). I've also tried Hexion 813 which is similar to Reichhold 37-127 with their amidoamine hardener( a no-go ) and IPDA which looks promising:

For Hexion 813 plus IPDA at 4.51:1 I'm getting a flexural modulus of about 440,000psi and a flexural strength of about 14,000psi which is better than any of the previous samples but by a small margin in some cases.

Consensus from folks like Tony on the thread as I have been able to determine is that we'd be better off with some like DER331 Resin from Dow which cross references to Reichhold 37-140 using Jeffamine D-230 and or D-400 hardeners. Tony's theory is that under vacuum, something as thick as DER331 will have no trouble wetting the grains. (There's a good chance that the the high viscosity epoxy on uscomposites.com is actually reichhold 37-140 but I haven't checked).

I'm a bit dubious about the thick stuff personally because making massive scale parts under vacuum is just not in the DIY agenda. There are also some warnings in B.W. Staynes' paper from 1975 about the thick epoxy making for a mixture that is not very workable.

I'm fairly certain the the Jeffamines are the right hardener to go with for DIY E/G due to low toxicity and excellent published performance numbers. I've had good luck with IPDA in a few experiments with resin alone but IPDA will eat the flesh off your hands and the paint off the floor if not handled with the utmost of care.

Chris aka crr has been ordering up some of the resins and hardeners for the DER331/Jeffamine experiment. I'm going to order some of the Huntsman equivalent to DER331 (Araldite GY6010) and the the Jeffamine samples tomorrow to get on with it. I've gotten sidetracked a lot lately losing several months due to work travel earlier in the year and then a couple of other projects and making a totally vacuum tight sample mold.

Other lessons:

Degassing at less than 29.5 inches of vacuum doesn't really do anything. I watched my epoxy just sit there with no effect until the vacuum chamber gage got above 29.5 inches. SO, going back to the discussion in <A href=http://www.cnczone.com/forums/showpost.php?p=301798&postcount=1373> post 1373</A> More than 29.5 inches is how much vacuum it takes to get to the center of a tootsie pop.

Eson,

I've been using silicone release Agent from http://www.mann-release.com/ At $6.00US a spray can, for Mann Release 200, I can't even begin to imagine a reason to try to improvise something. The dimensional change from the release agent was so small as to not disrupt the transparency of a glass-clear sample I cast.I wouldn't worry about the effect of a silicone release agent like this on dimensions. You'd be more likely to have a problem with an improvised one. I do feel however that a Poly Vinyl Alcohol or wax based release agent will mess up part dimensions because it's possible to build up a pretty thick later of PVA or wax.

[rowbare]

1 part #6 Agsco Brown Aluminum Oxide
1 part #4 Agsco Quartz
1 part #2 Agsco Quartz
1 part #2/0 Agsco Quartz
1 part 3M G800 Zeeospheres
1 parts 3M G200 Zeeospheres

Cameron,

I am guessing this mix is by volume? Is that correct?

bob

[ckelloug]

Bob,

These are by volume. Some of the components have different densities. The easiest way to measure a given volume is to multiply specific weight of the component by intrinsic density to get volume.

So for example if something weighs 4g/cm^3 and you want 10 cm^3 then 10cm^3*4g/cm^3=40g.

Alternately, you can tabulate the specific weight (1/density).
something with a density of 4g/cm^3 has a specific weight of .25cm^3/g. thus 40g*.25cm^3/g=10cm^3

Regards all,
Cameron

[thespindoc]

Hope you are all having a great holiday! I am fascinated with the epoxy-granite thread and am working on developing a new machine to market. My business has grown from a home business to a business with a large building and CNC tools. I still am basically a DIYC guy and would like to really move forward with a composite base for my desktop machine and hopefully can contribute to the thread. Do you have a handle on equipment such as mixers, vacuum chambers and ovens that could be purchased for a startup?

Thanks
Keith

[thespindoc]

I think I found sources for all the ingredients and some high end equipment, thanks!

Here is a question. Someone mentioned some problems using measurements by volume. There is the specific gravity and the loose pack bulk density which I assume will vary according the the grit size. Someone else mentioned how to calculate the weight needed to occupy a certain volume but which density do we need to use since there may be quite a range e.g.

SPECIFIC GRAVITY 3.95 g/cm3
OR
LOOSE PACK BULK DENSITY 1.69 - 2.07 g/cm3

Thanks,
Keith

[ckelloug]

The posted formula I developed is for specific gravity (which I've generally called intrinsic density) of the components. The theoretical equations on which that formula is based all work in volume fractions and consider a volume to be a mass divided by a mass/volume aka specific gravity of the material.

Still having dealt with only pure epoxy and pure aggregate in laboratory quantities, I don't have a handle on the production machinery for the final product. I know that some of the producers in germany have huge CNC vibrating tables which are talked about somewhere in the thread index and that vacuum beyond removing dissovled air in the epoxy before mixing with aggregate is uncommon at commercial manufacturers. I have a surplus vacuum chamber which is about an 18" cube and a surplus Grieve 48x36x36 curing oven as well as a lab balance from ohaus.com's surplus and closeout section.

I expect to have strength numbers for the aggregate formula with Hexion 813 and Araldite GY6010 using Jeffamine Hardeners and IPDA in the next few months. Between making apparatus and keeping up with ordering materials, I've made slow progress lately.

[thespindoc]

Thanks Cameron!

Keith

[thespindoc]

I found a great source for vacuum and mixing equipment with a range of economy to industrial. http://www.abbess.com/vac/homepage-vac.html

Keith

[ckelloug]

Keith,

I don't have a handle on what you are trying to make but the problem I run into thinking about it is that a large machining center probably amounts to making batches that involve several tons of material.

Mixing and handling tons of material is very different than making a 1/2 pound for samples which is different again from making 50 pounds of material for small parts. I am going to concentrate on home shop type machines initially when I get to the point of having a product as the amount of capital required to make huge machines is beyond my means at present.

I'd assume that a walk in oven, huge vibrating tables, and huge vacuum chambers would be required to scale the process that I'm using to produce laboratory batches.

One mechanism I've considered is using steel vacuum tight molds with bin vibrators bolted on like those available from FMC syntron. Haven't tried it however.

This would be a departure from what the commercial vendor folks have said on this thread. It's also true that I don't think Thomas Zeitz and friends go to that much trouble and they have excellent results. I'm paranoid about process control and making sure all parts have the same physical properties so I may be overengineering this whole thing.

see
http://www.fmctechnologies.com/Mater...Vibrators.aspx

[thespindoc]

Cameron,

I am just looking at making 2 bases for the x and y axes that will have steel walls all around and 2 steel central walls where I will run the slides. They will just be 1.5 inches thick and in the range of 12"x24" and 4"x8". I probably don't need to go with the epoxy-granite but it will certainly be helpful at eliminating vibration so I am going to go for it. I will just pour the mixture into the frame with no need to remove from a mold...

Keith

[ckelloug]

Finally, with Christmas comes some time without distractions.

I just finished casting a part with hexion 813 and Jeffamine T403. I think the ratio I used was 2.84:1 from extrapolating the numbers given in the Jeffamine datasheet for a slightly lower EEW epoxy (I have it written down but that lab- book is across town right now).

T-403 with Araldite GY6010 was the highest flexural modulus given in the Huntsman applications note. I think that T-403 Hexion 813 will come out about 10% lower in modulus with a nice increase in workability due to low viscosity when it comes to actual E/G. I also found in a Hexion app note for Hexion 8123 which is the same as Reichhold 37-127 and looks like Hexion 813 is stiffer and stronger.

As you can see, in the pic the sample, which was vacuum degassed and infused, is crystal clear except for a satiny finish caused by the mold release.

I hope to get ASTM D-790 data for this and a couple other Jeffamine configurations in the next few days.

[brunog]

Cameron,

I am just looking at making 2 bases for the x and y axes that will have steel walls all around and 2 steel central walls where I will run the slides. They will just be 1.5 inches thick and in the range of 12"x24" and 4"x8". I probably don't need to go with the epoxy-granite but it will certainly be helpful at eliminating vibration so I am going to go for it. I will just pour the mixture into the frame with no need to remove from a mold...

Keith

Keith,
E-G will certainly do the trick, in doing your project please keep in mind that adding hardware to anchor E-G with your steel casing/frame to allow a good mechanical bond, however I suggest NOT to use rebar or allthread in one piece connected to both ends or sides as this would tend to increase vibration harmonics.

Best regards

Bruno

[brunog]

Cameron,
unless your thinking of molding parts that exceed 1000 pounds in overall weight, one vibrator will be sufficient.

Best regards

Bruno

[ckelloug]

Bruno,

I agree one vibrator is probably sufficient for small parts but it should be sized correctly. There are some bin vibrators in the link-belt catalog that would shake my building down. Some folks have asked me in private about building huge machines is how this came up. These bin vibrators occasionally pop up on e-bay in various sizes. I found out about them because a small one in need of repair came attached to the surplus vacuum chamber I bought though I have not put it to use yet.

I'm still working on the epoxy formula but I'm getting close thanks to the input from Tony. It looks however like both the jeffamines and IPDA will require a heavy dose of tertiary amine catalyst to avoid the oven treatment and unfortunately, tertiary amine often causes brittleness and can worsen fatigue life from what I've read.

IPDA cured at room temperature comes out brittle like sugar cubes. The brittleness probably goes away after a month as the cure completes to some extent but I really like materials with defined properties when they leave the mold so I won't have to worry about them if and when they ship. I've also noticed that demolding while above the glass transition temp can induce some permanent deformation in the part.

Also, while it's a different project than most HSM's want to do, I think there are some incredible tricks that vacuum infusion can achieve. One should be able to make an "indian sand painting" of materials with different properties in different layers. Ultrahard titanium carbide for a way surface with quartz in the interior comes to mind as an example.


So how are you doing and what are you up to bruno? I haven't talked to you in a long time.

Anybody else working on anything they want to share?

Regards all,

Cameron

[thespindoc]

Cameron,

If I am going to make a frame and fill it with E-G, why do I need the particulate components. Would epoxy alone be enough?

Also, for small components, I assume vibration is not necessary especially if mixing under a vacuum? For my application with frame, I would not think that small voids or poor wetting would even be a problem. Am I missing something?

Thanks
Keith

[ckelloug]

Keith,

There are several reasons for the aggregate in epoxy granite which are not necessarily obvious and not necessarily applicable to all parts.

First: epoxy is between $30 and $70 dollars a gallon.

One you get over the cost hurdle, there are some physical properties of epoxy that need some help. First, Epoxy has compressive yield strength on the order of 10,000 psi. Quartz on the other hand has a compressive strength about an order of magnitude higher, around 100,000 psi. Thus, a pure epoxy part when loaded heavily will deform substantially whereas a composite which is mostly quartz will barely deform at all.

Epoxy also a high rate of thermal expansion: 1e-3mm/mmC whereas quartz is about 4 orders of magnitude better on the order of 5e-7mm/mmC. This means that if a large part changes temperature by a few degrees or has a temperature gradient across the surface, a pure epoxy part could deform substantially.

Cured epoxy is also combustible and could be charred or damaged by hot chips. It is also quite soft compared to metals meaning any sort of abrasion will remove material. Adding aggregate means that the surface area of epoxy that can be harmed by hot chips is greatly reduced as quartz or alumina will limit the damage to the surface both from abrasion and heat.

Finally, the reason that epoxy granite has excellent damping characteristics is that every particle of aggregate causes the reflection of a vibration traveling through the material and aids in the absorbtion of the vibration. Much as the beam of a flashlight shined on a mirror is not as bright as staring into the flashlight, a vibration that has bounced off of 100 pieces of aggregate losing 1% of its energy each time is reduced to heat much more efficiently than by a material with no discontinuities.

In short, pure epoxy is not hard enough to maintain dimensional stability under load, it expands too much on temperature change, it can be damaged by hot chips or cold ones or stray tools. . ., and it is much less effective as a vibration damper than a multiphase material.

Also, keep in mind that I have studied aggregate from a theoretical standpoint and pure epoxy in the lab. I do not have data yet on the mixing properties of epoxy and aggregate and how much vibration is necessary if any to compact various parts. I can say that I believe much less vibration will be necessary if the epoxy is added by vacuum infusion but this wouldn't be worth it unless you were doing high volume or special parts.

A few small voids won't likely be a problem if the parts have no significant load but griffith fracture theory tells us that a huge number of unwetted particles in a high load situation will behave as fracture initiation points an lower the strength of the material. I worry about these things because I don't want a 100 lb machine spindle held by a solid E/G part to take flight. There are several additives that substantially improve the wetting behavior (3m fluoro surfactants) and the bond between the epoxy and aggregate (siloxanes and titanates) but I as of yet have not tried any of them though I do have samples.

In conclusions, aggregate will make your piece perform substantially better and last longer. (AS an example I tightened a screw against pure epoxy samples today in a sawing jig and it left a huge mark).

In that this is a table section you want to make with steel structural components, you probably don't need to be as cautious as I am about voids. I want to make unreinforced structural parts and that means keeping the flaws as small as possible in hopes of keeping the Griffith stress (where cracks propagate to failure) well above the load for the remaining flaws.

Finally, you should be able to get a nice flat surface by pouring a layer of thin epoxy perhaps fortified with something like fine alumina or silicon carbide or titanium carbide over the surface as well as some titanate like Kenrich Lica 34J to keep the viscosity down. If you use a precision level to level the surface before the pour, you can even get it parallel to the rest of the machine. I haven't tried this yet but I think getting the machine tool grade surface plate formula right will be a big boon to HSM's everywhere.

Best of luck,

Cameron

[brunog]

Cameron,
I've started working in design for a concrete plant manufacturer last September, I have also had the chance to stumble on a few vibrator specs on my own. I go to this thread at least once or twice a week to see how things evolve. My shop is on the design stage and I should start building it in a couple of months. It's actually a family room that used to be a garage that will be modified to a workshop, finally room to play!

Best regards

Bruno

[thespindoc]

Thanks Cameron for the very detailed response, I appreciate it very much!

Keith

[greybeard]

HAPPY NEW YEAR

to one and all

from

The Frizzleton Vole Unicycle Display Team

[ckelloug]

I ran a couple tests the other day on Hexion 813 and Jeffamine T403. I ran the first one at a low hardener level as dictated by the Jeffamine datasheet for mixing with pure epoxy and the second one at stochiometric ratios.

The low hardener test yielded a flexural modulus of 333,000 plus or minus 24,000.

The stochiometric test yielded a flexural modulus of 419,000 plus or minus 25,000.

The behavior was quite different from the predicted values in the Jeffamine datasheet for the results of T403 with pure epoxy.

The crysyl glycidyl ether in the Hexion 813 is having a negative effect on modulus.

What's interesting however is a quick perusal of the datasheet for crysyl glycidyl ether at cvc specialty chemicals. It appears that when crysyl glycidyl ether is added in quantities of between 10 and 15 parts per hundred as opposed to the 20-30 I estimate for Hexion 813 that the flexural properties of the mixture are actually better than those of undiluted epoxy (at least with a TETA cure). It looks like the viscosity is likely to be somewhere between 500 and 1000 cps at this CGE level as opposed to the 300-500 cps viscosity seen in hexion 813. See the following <A href=http://www.emeraldmaterials.com/epm/cvc/micms_doc_admin.display?p_customer=FISCVC&p_name=E RISYS%20GE-10_TECH%208-21-08.PDF>link </A> for the properties of cresyl glycidyl ether.

Using this effect, it should be possible to get our raw epoxy modulus up to about 500,000 psi. Adding some nanoresins nanopox, I expect that the modulus of the epoxy component can be raised to between 600,000 psi and 1,000,000 psi though the likely value is probably about 800,000 psi! This is still not as good as a cure with TETA but the exotherm and low pot life of TETA make it unworkable.

Using the handy graph I produced from the data in Kinloch's book back at <A href=http://www.cnczone.com/forums/showpost.php?p=444339&postcount=3119> post 3119</A> it appears that with quartz as the particle content, that between 600,000 and 800,000 psi Epoxy all perform about the same.

So, from the aggregate design in use and the predicted modulus of the final epoxy with nano particles, I expect the flexural modulus of 92% quartz and 800,000 psi epoxy mix to be about 6,000,000 psi or about 50% of class 20 cast iron or about 20% of the modulus of steel. I believe that this is fairly close to the limit of what's achievable with epoxy granite.

Accures Polymer Castings' material has a modulus of 4,500,000 which means the material I am aiming for with quartz should be 33% stiffer than the accures numbers.

Redoing the post 3119 graph for alumina instead of quartz brings us to the interesting conclusion that an epoxy granite made of 92% alumina and the currently hypothetical 800,000psi epoxy ought to have a modulus of 15,000,000 psi or 75% of class 20 cast iron.

With the 420,000psi epoxy we have now, I'd estimate the modulus of the alumina E/G to be about 10,000,000 psi or about 1/3 that of steel.

Finally, the Reichhold 37-127 37-606 combo, using 92% alumina, ought you yield a flexural modulus of about 6,000,000 which is better than off the shelf mixtures like Accures and Zanite. The only disadvantage to alumina other than a small cost premium is that it has a larger coefficient of thermal expansion than quartz.

Conclusion:

Slow hardening epoxy isn't likely to have a modulus much above 450,000psi. Low viscosity slow hardening epoxy with a viscosity of 500-1000 cps likely has the highest flexural modulus around 500,000psi. To get 800,000 psi out of a slow hardened epoxy, nano-additives are almost certainly necessary.

Ideal Quartz Eg tops out at a flexural modulus of about 6 million using 800,000psi epoxy.

Ideal Alumina Eg tops out at a flexural modulus of about 15 million using 800,000 psi epoxy.

Six million psi alumina E/G should be achievable with almost any epoxy formulation including the 320,000 psi 37-127/37-606 formula.

In all these conclusions about E/G recipes, it should be noted that the modulus numbers are predicated on coupling agents (silanes and titanates) ensuring a near ideal bond between the aggregate and epoxy.

What remains:

1. Tests with aggregate to validate the predictions.
2. Tests with aggregate coupling agents to find dosages and show the necessity or lack thereof.
3. Tests to see if something like DETDA which is a non-hazmat hardener when combined with nano-reinforcements can produce a less toxic/hazardous formula.
4. A few more calculations on quartz and alumina aggregate to make sure the size distributions are spot on.
5. Figuring out if it's possible to work these very thick mixtures.(It should work fine for vacuum infusion).