[oxford]

What kind of questions should I be asking when I talk to the rep?

What kind of buzz word should I be researching to make this a meaningful conversation?

Thanks

[RomanLini]

...
Bigger aggregate particles won't increase modulus. An aggregate distribution with larger maximum size is easier to get a high packing density for high modulus but if you just use a bunch of large particles without the smaller sizes, you will get a decrease in modulus. Larger particles also lower the strength as they are more likely to contain a flaw of the Griffith critical size. RotarySMP Mark is right that 1/5 of feature size is a good rule of thumb for max aggregate size.
...

I'd really like to hear a "laymans language" description of WHY smaller particles are more rigid beacuse just throwing some intellect and common sense at it comes up with the opposite result.

Here's my take on it;
The stone is much heavier than the epoxy and larger stones will sink to the bottom and be in direct contact with each other, there will essentially be zero epoxy between the stones.

Stones have basically zero compression (compared to epoxy) so larger stones will act like non-compressible beams in a lattice structure.

Larger stones have higher density per surface area and will sink harder, sitting on each other with more force per "joint" area, displacing epoxy better and causing much less epoxy in the "joint" where the stones join.

So now you have a lattice work of large non-compressible beams in full contact with each other, the only flexion is going to come from the joint pulling apart ("stretch") on the opposite side of the beam (the tension side).

If this is encased in a permanent metal former that stretch will be very difficult as the stones will be locked in full contact with the outer former and with each other.

Now as a comparison, if you have small particles like "grit" they won't fully displace the epoxy from between each particle or at least not with as much force as larger heavier particles. So under flexion of a beam you have a higher number of the epoxy joints which can cause stretch under flexion, and there is more epoxy in the joint so each joint itself will be more flexible. That's a double whammy.

So a self supporting beam (like a kitchen countertop) would be better with smaller particles as it will be a little more flexible and a LOT stronger (resist breaking).

But if made with larger particles it would be much more rigid, and much greater risk of breaking (if not supported).

As for the ideal size, in the 2" void like some of the lathe bed side beams as shown in this thread, I would have used marble sized stones, which is probably very close to the ideal 1/5th of the beam.

[johnohara]

ckelloug,
Thank you for your offer -- very generous. As I get better at
this I may take you up on it.

RomanLini,
The aggregate in this casting runs 0.25 - 0.50mm in size. But
it does show what would happen if all the aggregate were
approximately uniform in size.

My first thought after demolding this piece was "now I understand
why you need you smaller sizes to fill the voids."



1. Uniform size + epoxy. 90:10 aggregate to epoxy.



2. Bottom of the piece (smooth side).



3. Closeup of bottom.



4. Loose masonry sand to see what goes into the voids.



5. Simple tapping of sides with fingers fills the voids.

~John

[RotarySMP]

Here's my take on it;
The stone is much heavier than the epoxy and larger stones will sink to the bottom and be in direct contact with each other, there will essentially be zero epoxy between the stones.

Stones of uniform size will give roughly point contact between each other with huge voids. Thing of a pyramid of marbles.

You need a range of filler sizes from roughly 1/5 the thinnest feature down to fine grit. There are a number of recipes on this thread ranging between theoretical "perfect mixes", and practical home bake mixes.

You could throw in a few big rocks to fill out thick features, and reduce the epoxy needed, but normally such a feature would be designed out, or filled with a foam core.

[johnohara]

(All captions will appear above pics from now on)

1. Fine white sand (probably silica) -- 100g



2. Course play sand -- 200g



3. 0.25 - 0.50mm aggregate -- 200g



4. Dry mix (no epoxy) all. Vibrating the dry mix causes the
larger pieces to rise and the finer sand to fall.



5. Dry mix + epoxy resin (40g without catalyst)
Mix as long as you want to cover everything.



6. Dry mix (500g) + epoxy resin (40g) + catalyst (20g)



7. Side view after manual vibration and tamping.



Will demold after 24 hours. Cure will take 7 days at room temperature.

89/11 mixture.

Aggregate: 500 grams
Epoxy Resin: 40 grams
Epoxy Catalyst: 20 grams

~John

[lgalla]

John good work.
What is the name of your larger aggregate?eg,pea gravel,river rock etc.
Roman larger aggregates espically crushed may be fractured having cracks.Previous tests showed under stress,hammer blows,the aggregate broke before the epoxy.This is a simple reason not to use large aggregates or rocks.
John try to put a thin layer of epoxy on your mold first.It may reduce the pot holes
Larry

[lgalla]

John,I just read your last post in detail.You must never mix the aggregate with the resin withought the cataylist

[ckelloug]

Oxford,

I'd say you want a low viscosity epoxy with a flexural strength of at least 10,000psi and a flexural modulus of at least 400,000psi.

Romanlini,

It has been shown empirically in the book <u>Advances in Resins and Structural Primers</u> by A.J. Kinloch that the stiffness of an epoxy particle composite is dependent only on the ratio of aggregate to epoxy by volume.

Assume a sample of an optimal mixture where the large stones segregate and fall to the bottom. The top half now has a reduced grading span (difference between the largest and smallest sizes) while the bottom has an increase in large particles. It can be shown from packing models that reduced grading span and excessive large particles both lower the packing density. It was shown in the Kinloch Book that modulus is proportional to packing density. Since both the top half and the bottom half lost packing density when the mixture segregated, the segregated mixture has a lower modulus than an unsegregated one.

It's the ratios between the stone sizes and percentages in the mixture and the shapes of the stones that affects the modulus, not the absolute sizes of the stones. RotarySMP was right on target with his mention of pyramids of marbles.

[johnohara]

lgalla,

Got the impression from this video that mixing resin first before catalyst was okay.

[ame="http://www.youtube.com/watch?v=vrgbCrhkJTI"]YouTube- scott bader epoxy granite[/ame]

I think they're casting "cultured marble." But the catalyst is added after to extend
the pot life.

~John

[johnohara]

.

[ckelloug]

If you're using a true catalyst such as a tertiary amine for homeopolymerization, I can see adding the hardener after mixing as being possibly reasonable.

If you're using a primary or secondary amine hardener that is a reactant in the hardening process, it seems to me that it won't get distributed well enough to make the epoxy harden correctly.

Personally, I'm still dubious about mixing with aggregate first and then hardener, even if you are using a true catalytic hardener but that's just my opinion

[johnohara]

ckelloug,

I'm using US Composites 635 resin with 556 hardener. Their site does not give MSDS info so I don't know if it's a tertiary amine hardener or not.

I bought it for the pot life which is supposed to be 30-40 mins and because it was mentioned in this thread and the price was right. The West System epoxy looks good but for now is too $$.

My next sample will exclude the 0.50mm aggregate and I'll use mixed resin/catalyst.

~John

[lgalla]

"I think they're casting "cultured marble." But the catalyst is added after to extend
the pot life."
That is correct.Cultured marble,solid surfacing etc.are very different from E/G.The resin is polyester with MEKP catylist,methyl ethyl ketone peroxide.The mixes are very thin and can be 50/50.I think the catylist is added at the end as the cat may vaporise under vacuum.
Epoxy is difficult to get a proper mix.In the video you see them scraping the mixer.This is another no,no for epoxy.After mixing the epoxy and adding the fillers and pouring it is recomended not to scrape down the mixer and sides.There is always under mixed epoxy that collects on the mixer and mix vessel sides
Larry

[johnohara]

White sand -- 200g
Play sand -- 400g
Epoxy -- 45g (30:15) (US Composites 635:556)

Mixed resin and catalyst then added aggregate. Swimming in air but I'll deal with that later.

~John

[RomanLini]

...
It has been shown empirically in the book <u>Advances in Resins and Structural Primers</u> by A.J. Kinloch that the stiffness of an epoxy particle composite is dependent only on the ratio of aggregate to epoxy by volume.
...

But did he do tests with the mixture cast within a steel structure?

I've worked with epoxy for a lot of years, starting in industry in the '80's and more recently quite a lot over the last 11 years. Finer particles don't have the density required to squash the epoxy from between the particles. The weight of the particle is the cube of its diameter while the surface area is the square of diameter, and epoxy has a lot of "cling" to particles especially so with the high strength epoxies that grip the surfaces well. So as particles become much smaller they don't have the mass to overcome the surface tension and the particles won't "touch", they are all separated from each other by an epoxy skin.

Check in your own tests, the small particles sitting on the bottom of the mold won't displace the epoxy undreneath them, so you still get a smooth epoxy surface when demolded. To get around this you use less and less epoxy in the mix, to try and "force" the deal (as A. J. Kinloch discovered) and get the particles in a greater contact with each other with less epoxy between. But then you get the problems with difficult mixing and air voids and difficult potting. Even with the vibrator table you will probably get air voids. And vacuum degassing is probably not practical for most lathe builds. You can see in John's post above picture 7 there is nowhere near enough epoxy, the top 40% of the product is full of air.

In a supported iron structure the rigidity comes from a very low amount of epoxy between the particle joints, and using larger stones will enable an easier mix, easier pouring and they will lock themselves into the iron structure very effectively with their much higher mass/surface ratio.

If you are interested I did some tests about 5 years back of rigid composites. I can't disclose all the data because of commercial reasons but one thing I can tell you is that for the finer particles you will get much better results from garnet sandblasting particles, which are not too expensive. It is heavier than sand (hence better settling and less epoxy in the joints), and the complex fractured shape of the hard garnet particles is extremely good for epoxy adhesion.

[johnohara]

1. No vibration when casting. Shows air pockets. The epoxy surface is smooth as glass.

2. Closeup shows smoothness of the surface.

Getting there. I admit it's not perfect but I'm learning how to work with the materials and more importantly how the various ratios "feel."

[ckelloug]

Johnohara,

I just wrote to U.S. Composites and asked for the datasheet information on your epoxy. We'll see what I get. I called them last time and got the info on the thin system.

I think Larry is right that you need to mix the hardener and epoxy before adding to the aggregate. As the epoxy hardens and thickens , the mobility of the hardener decreases and if they are not thoroughly mixed, you will get regions with an incomplete cure and other regions with too much hardener. A catalytic hardener is probably less subject to this but in the end, I don't think you'll get a complete cure adding epoxy to aggregate before mixing. Pot life of the mixed epoxy aggregate will be better than that of the pure epoxy because the specific heat of the aggregate will keep it from getting so warm.

It almost looks like you don't have enough epoxy for the aggregate mixture you have although it may be an artifact of the way the hardener was added.

RomanLini,

Your results sound quite interesting and I as well as others I am sure would like to hear more. I haven't seen any tests of filled structures.

I've also thought more about it and the Hashin Shtrikman equation really doesn't apply to your case as I argued before because it assumes homogenous material. Thus I can't legitimately use the packing model and the Hashin Shtrikman equation to predict the modulus of the filling of your structure. I am certain it is quite valid for the case of homogenous mixed aggregate where there is no reinforcing structure.

I assume there is a small layer of epoxy coating every particle and given this assumption, you still end up compressing the epoxy when a part is deflected.

The method you are describing will produce highly anisotropic reinforcement. My understanding is this: if x is the long axis of the beam, y the short axis and z up then you will have extra stones at the top of the beam. These will be in close contact in the x direction and if you deflect the beam in the downward z direction, they will be loaded in compression and in contact while the tension section of the beam will be in tension. If you load in the upward z direction, these stones are in tension and the higher epoxy percentage with your large stones guarantees at minimum that there will be regions of increased deflection locally. I think it's probably true that there will be increased deflection globally and that the overall modulus will be lower than a beam with maximum packing density.

I think you theory hold true if you have square bricks with high contact area as your reinforcement. I'm not so sure I believe it works well for real aggregate with point contacts.

If you have more info, I'd enjoy learning more about the situation.

Regards all,
Cameron

[johnohara]

The 93:7 ratio seems to need the following to be optimal:

1. A good pre-mix of dry material. The recipe found in the summary index for this thread is reasonable. The mix I used in this sample could use some form of mineral powder 7 Mohs or above.

2. Gentle vibration to settle the aggregate and epoxy. Concrete vibration tools will work but they seem violent unless the cast is large. I'm thinking more along the lines of multi-speed dental vibrators and hand-held orbital sanders.

3. Vacuum to de-gas the air from the epoxy. Mixing aggregate and epoxy under a vacuum (as shown in the scott bader video) would work well. Maybe a diy vacuum mixer or some sort of vacuum bagging technique that subjects the entire mold to vacuum. The air HAS to go.

4. Pressure to properly shape the material. Manufactured quartz surfaces are exposed to 100 tons of pressure in the Breton Process but they use polyester resin. The 93:7 ratio creates moist material similar to cast architectural stone (http://www.haddonstone.com) which employs vibratory tamping.

My original cast -- the brown one -- was 15-16% epoxy. It seemed to be okay except for the microscopic bubbles. Good vibration and vacuum would eliminate that.

My questions are changing -- that's good.

[RomanLini]

...
The method you are describing will produce highly anisotropic reinforcement. My understanding is this: if x is the long axis of the beam, y the short axis and z up then you will have extra stones at the top of the beam. These will be in close contact in the x direction and if you deflect the beam in the downward z direction, they will be loaded in compression and in contact while the tension section of the beam will be in tension. If you load in the upward z direction, these stones are in tension and the higher epoxy percentage with your large stones guarantees at minimum that there will be regions of increased deflection locally. I think it's probably true that there will be increased deflection globally and that the overall modulus will be lower than a beam with maximum packing density.

That's exactly how I see it. In the case of the lathe bed where the bottom of the bed is the sprue it will be weakest in downward deflection. But in sideways deflection the beam is stronger as it has steel on the tension side. If it was a box tube filled from the small end that problem would be solved, as the tension side of a tube crushed in which would support the non compressible lattice. But anyway, for any aggregate mix in a lathe bed that open bottom sprue is the weak point.

For really high rigidity I would use larger stones (ie marble sized) and physically compress them into the steel frame. Pressure from somethign as simple as a wood chock cut to the shape of the bed with a couple of hundred pounds on it. The larger size stones would guarantee point to point contact of the non compressive (larger) stones, and the pressure would be extremely effective at removing the epoxy between the non compressible parts including the steel.

I never said above that smaller particles should not be used. But I would use a smaller quantity basically just to fill much of the voids between larger stones. Not use the small particles as the primary lattice structure.

You could also hold the steel frame at a slightly elevated temp (say 50'C) for the entire curing period, this promotes a stronger epoxy cure and will also provide additional compression of the structure onto the stones later when it cools.

I assume there is a small layer of epoxy coating every particle and given this assumption, you still end up compressing the epoxy when a part is deflected.
...

I think this is the crux. Compressing the mix, with larger particles will give the least epoxy in the joints.

Johnohara- If you are testing mix ratios it will depend a lot on the particle sizes, smaller particles will soak up more epoxy around them and require vibration, or even better vacuum or physical compression to allow you to use the low ratio of epoxy you are testing. You can't just mix it and let it sit when it has such a low epoxy amount.

Maybe you could try a test with a wood former on top of your mold and put some weights on it...

[lgalla]

John I asked previously what is your aggregate called.Also I think you have a typo on the size.At.25-0.50mm would be sand.the decimal should be one place over???I am attaching some aggregate pictures mainly 1/4"minus and different pea gravels.Your gravel is first.
Hopefully you can see you aggregate has many different sizes and shapes.The pics I posted have mainly constant size and round shapes.This is better for packing density.
Larry