[altaic]

Sandi, I understand; I just wanted to be clear.

I actually observed exactly what you described, where the bottom was very compacted and hard, but the top was a little soupy by comparison. I think adding more big particles would do the trick, but pressure I'm certain would help as well.

The next test I will do will be another one with water or perhaps a light oil, and I will add large particles until they saturate the mix and are visible on the surface. I also did a fracture test on the E/G plate I made yesterday, and I noticed that almost all of the large alumina particles broke in half down the clean break (as opposed to the break moving around the alumina). I interpret this as the epoxy is not the weak link (no pun intended), and the mix may benefit from some stronger large particles.

[BillTodd]

I believe that the lower parts of the EQ in the mold will compact together quite well due to the mass of the material above it, but as you get closer to the top of the mold, the EQ will not compact together as well. That is why I would devise some way of applying pressure to the top part of the mold so it can also compact optimally. This would only be required while vibro-compacting the EQ...

Just a thought:
What about applying the vibration from the top (i.e. like a wacker plate, or soil compactor) while the bottom of the mold is supported on something compliant (e.g rubber or foam)?

[altaic]

Just a thought:
What about applying the vibration from the top (i.e. like a wacker plate, or soil compactor) while the bottom of the mold is supported on something compliant (e.g rubber or foam)?

This may be a very good idea-- I'm thinking vibrate the top and bottom 180° out of phase, so that each vibration cycle basically squishes it. Care has to be taken to ensure the lid doesn't slant due to anomalies in the vibrations or imbalanced weight, so the lid should be guided by lid thickness and fit, or a rail or something.

Edit: Now that I think about it, putting a large mass on the top and vibrating the bottom should be equivalent to vibrating both 180° out of phase. The mass's inertia presses the mix down when the bottom vibrator moves up, etc.

[brunog]

I noticed that almost all of the large alumina particles broke in half down the clean break (as opposed to the break moving around the alumina). I interpret this as the epoxy is not the weak link (no pun intended), and the mix may benefit from some stronger large particles.

Will,
did the quartz particles break as the alumina did?

Bruno

[altaic]

I've been hunting for low viscosity, low shrinkage epoxies, and I was wondering if those with more knowledge would give thoughts about some I've found. I'm leaning towards 117LV-239 since it's viscosity is the lowest and it has fewer MSDS warnings. 117LV-229 is the next lowest viscosity and doesn't require high temp curing (120-180°F).

* PRO-SET Epoxy 117LV Resin w/ 239 Hardener - 117LV MSDS - 239 MSDS (high temp cure required)
* PRO-SET Epoxy 117LV Resin w/ 226 Hardener - 117LV MSDS - 226 MSDS (high temp cure not required)
* PRO-SET Epoxy 117LV Resin w/ 237 Hardener - 117LV MSDS - 237 MSDS (high temp cure required)
* PRO-SET Epoxy 117LV Resin w/ 229 Hardener - 117LV MSDS - 229 MSDS (high temp cure not required)

P.S. PRO-SET is one of the only companies I've found in the past few days with easily accessible US distributors, a variety of epoxies and (gasp) very complete technical data sheets. I started to get worried after looking at so many different epoxies with little to no technical data available or no contactable distributors. Talk about frustrating...

[altaic]

Will,
did the quartz particles break as the alumina did?

Bruno

Good question. I'll have to take another look... The quartz particles are much smaller, so I couldn't really tell from my initial quick examination.

Edit: Okay, it's really hard to tell, but I'd say that yes, at least some of the quartz did break. Basically I look for shiny bits the size of my quarts particles, then I look at the same spot on the other half of the break and look for a shiny bit that matches up. Sometimes I can match the surface of the tiny particle to verify a break, but it's really slow. The large particles are much easier to locate and verify.

[ckelloug]

Will,

From the B.W. Staynes paper from the Proceedings of Polymers in Concrete, 1975, it's normal for E/G to fracture through the aggregate. The tensile strength of epoxy far exceeds most aggregate. The compressive strength and compressive modulus of epoxy are far lower than aggregate however which is why you need it. If you're trying to make E/G strong, the parameter to look for is fracture toughness of the aggregate and epoxy since E/G fails though brittle fracture.

I specified alumina for the largest components because it has a fracture toughness (KIc)of 4 megapascal square root meters whereas quartz has a value of only 1. Larger particles usually have more flaws so they are almost always weaker than smaller particles of the same material. From my knowledge of fracture physics backed up from the graph of strength vs. particle size in <u>Advances in Structural Primers and Resins</u> you want as small of particles as you can get away with with as high a fracture toughness as possible.

The calculations I did long ago indicate that if you make the stuff rigid enough to hold machine tool .0001in deflections that the strength isn't really material. My last recollection of this is that most beams that won't deflect under cutting loads have a safety factor of 10 or so for strength.

I hadn't managed to find Proset epoxies myself and I did a lot of googling. I very much like all of the datasheets you presented that I looked at. The only disadvantage to the room temperature hardening variety is that it will make it up to quite a high temperature before cure is mostly complete leading to the possibility of more shrinkage than the slower ones. I can't quantify the effect to say whether it is a real problem however. The epoxy you were interested in will do just fine as far as modulus and strength go. It should work rather well I predict.

You are correct about needing more larger particles. Since you are hard at work actually testing, I'll present you some better model results for using the agsco aggregates you have. The original formula with equal parts was one I posted for Walter who was very clever but not even remotely into measuring stuff. The original mixture was an 85% mixture designed for the sake of simplicity. A mixture with about 89% density using the minimum segregation criteria is suggested by the model as:

0.18744 Agsco #6 Brown Aluminum Oxide
0.38420 Agsco #4 Quartz
0.12950 Agsco #2 Quartz
0.11514 Agsco #2/0 Quartz
0.11352 3M G800 Zeeospheres
0.07020 3M G200 Zeeospheres

These percentages are in intrinsic volume of particles (not bulk volume from like that from measuring the aggregate in a beaker). So, multiply each fraction by it's density to get the mass required when doing the calculation. Also note, I left more significant figures than reasonable to show that it adds up to 1.0.

The model is supposedly good to 3% density so it's in the realm of reason that this formula is good for 88% packing density. With the epoxy you propose and this aggregate formula, theres a reasonable chance of duplicating Accures/Zanite performance numbers.

Best of Luck, Will.

Regards all,

Cameron

[romihs]

Talking about Walter, does anyone know what happened to him?
He simply disappeared.

Regards

Sandi

[ckelloug]

Sandi,
Nobody's entirely sure. I believe I was the last one here to correspond with him. Before he disappeared, he sent me a chip of the sample from akvacnc and then he fell silent.

I got one e-mail more from him saying he was okay after sending him a message that I was developing an extreme concern about his wellbeing and had started checking the obituaries for him. I won't go into detail about exactly what he said since it would be making public what should stay private other than to say he told me he had lost interest in E/G and this community and that it wasn't personal.

I secretly hope Walter's lurking here somewhere as I think he would enjoy seeing his work being built upon. Has anybody else has corresponded with him?

[altaic]

Cameron, thanks for the excellent information. It's really got me thinking and I have many new questions, however I will try to exercise some restraint until I understand things better (and have time for much reading). Perhaps just a couple things...

I understand the varying particle sizes and random crystalline orientation are the main reasons why E/G is good at damping vibration. The epoxy holds the particles together and allows them to absorb the mechanical vibrations. I imagine two extremes, one where the epoxy is very flexible and allows the particles to move fairly easily without transmitting much vibration to the next particle, and another where the epoxy is very rigid, such that any particle movement is transmitted unhindered to the other particles. Tending toward the flexible one seems best for vibration damping (in the scope of many epoxies that already have excellent strength and modulus figures). So, I look at values such as Tensile Elongation % and I think that the ability to stretch a lot before breaking might be an indicator for good vibration damping and toughness. I mean we're only talking about a few percent, but if it's 2% vs. 6%, that may make one a lot better than the other.

Then, things like Izod values seem useless unless I'm making something that needs to dissipate a lot of energy when breaking, e.g. a one-time use shock absorber or crumple zone. OTOH, perhaps it's also an indicator of toughness, i.e. since it takes more energy to break, it will be more resistant to fracture and perhaps more readily convert mechanical energy into heat before that point?

These are obviously very complicated subjects, so I guess I just want to know if the Pro-Set epoxies I listed are all just about the same and I'm over-thinking them, or if it warrants more thought.

[ckelloug]

Will,

In my scan of the proset datasheets, I didn't see much difference except for a small decrease in modulus as the setting times got shorter. They look better than the stuff I've been working with performance wise. The bad thing about the room temperature stuff I do see now that I have thought more about it is that the gel time is pretty short and it may not leave enough time to adequately mix and place the material. Some of the slower ones achieved full modulus after two weeks and I know that this is the strategy used by at least one british manufacturer (roach's employer).

I have very little idea how to model the epoxy granite's vibration damping aspect. My intuition tells me that because the modulus difference between the epoxy and the mineral components is so great, that it will always be a pretty good damper for a rigid material since elastic waves sustain a loss each time they bounce. I believe that this is the same reason that meehanite cast iron is a good damper compared to steel: it's an iron with graphitic inclusions. What really happens on a microscopic scale is probably reserved for epicly large coupled blobs of masses and dampers in a computer model.

I agree that Izod numbers are of minimal usefulness here. Once you have a beam with 6 inch square cross section or something like it, chances of breaking it are vanishingly small and far less than breaking a concrete part of the same size. As far as I am concerned, Izod numbers were never very meaningful but they were all that they had before theory of brittle fractures was invented. I prefer fracture toughness as a criteria because it is a true measure of the energy required to propagate a crack in the material which is an intrinsic property in a way Izod is not. If you have a KIC value for a material then you know the material with a higher value will always require more energy to fracture it than a material with a lower value. Applying fracture mechanics came from materials science which is a much newer field than the remainder of mechanical engineering so human inertia and the expense of the tests have made it less used than it seems like it should be. See http://en.wikipedia.org/wiki/Fracture_toughness

Feel free to muse, ask questions here etc. There are lots of people that will jump in and answer or brainstorm. I've been working on this for a long time but I certainly don't have all the answers and none of my answers are based on E/G measurements at this point, only measurements and calculations on the constituitive components. The two major answers I am certain of from my research are that you need to hold 88% packing density on the aggregate and find epoxy with a flexural modulus better than 400,000 to make material that will be as stiff as zanite or accures.

[Bada Bing]

We had some bases made at ITW, a company I worked for manufactured a photo plotter that was very precise. The material ZANITE is very stout and held the vibration to a minimum.

[altaic]

I've made progress in the past couple days getting set up to cast E/G.

I managed to pick up a nice new Welch vacuum pump off ebay for a fraction of the cost because it had a superficial crack in the motor's cooling fan guard. What a find! I did a lot of research before deciding what kind of pump to buy, though, so I'll briefly outline my findings.

First I looked at Gast pumps, since it was recommended here to pick one up for cheap. Gast, however, makes a large variety of pumps, as do other major manufacturers such as Welch, Edwards, Alcatel, Busch, Leybold, etc. Here are the various types:

• diaphram pumps: quiet, oil-free operation, but requires significant upkeep and is slow and weak (not enough vacuum to be useful to us)
• rocking piston pumps: loud (60+ dB), oil-free operation, and some are capable of about 29" Hg of vacuum (negative pressure as measured from relative atmosphere), or 25 torr (as measured from perfect vacuum, which is how most manufacturers I've found do it)
• piston pumps: about the same as rocking piston pumps, but tend to pull less ultimate vacuum, perhaps 28" Hg, or 50 torr
• direct-drive rotary vane pumps: fairly quiet (~50 dB) and fast, need oil (otherwise maintenance isn't supposed to be a problem, and as far as I understand the pump is designed to run for long periods at high vacuum), and have an ultimate vacuum of around .001 torr (29.9999" Hg) or better, however they are very expensive unless you buy second hand
• belt-driven high-vacuum pumps: even higher vacuum levels and better speed, but most of the ones on ebay were pretty decrepit looking

I went with a rotary vane pump because I'd intended to do some high-vacuum testing and I happened to come across an awesome deal. A rocking piston pump or a good second hand belt-driven high-vacuum pump would probably be my second choice. However, there are so many decent looking rotary vane pumps on ebay it's really probably cheaper and better just to get one of those. Of course, YMMV, and I'm sure a lot of those pumps require some attention. Speaking of, rebuild kits for most of those pumps are between $100 and $300 from what I saw. It might be worthwhile to locate and price one before making that highly reliable ebay purchase.

~

Anyhow, I also ordered samples of the Pro-Set epoxies I listed (except for the 339 hardener since it wasn't immediately available), which should be here before the weekend. I'm also waiting on some other vacuum stuff like fittings, a gauge, and some vacuum bagging supplies.

My plan once everything arrives is to first make a 5cm x 5cm x 20cm bar mold. I have some good plaster that I think will do very nicely for surface finish and strength, although I'm not certain it will hold up to vibrating the mix. Might as well try, since plaster would be nice for any sort of big mold, where buying a huge block of aluminum or doing tons of welding wouldn't be so nice.

After I make the mold, I will do a pour without vacuum so I can observe the flow and wetting of the new resin. After that, I'll do a pour with the following (preliminary) steps:

• coat plaster mold with silicone mold release
• coat lid similarly; it will be an aluminum shouldered/stepped lid with one or more holes in the center for evacuation of air and excess epoxy
• let the release set up
• fill mold up with largest particles (#6 alumina, in my case) such that it's full but the lid is able to tightly seat on the mold, and then empty the mold into a tray and weigh it; this will serve to give an approximate scale for the recipe (sanity checking is good)
• measure and mix epoxy for 15% (some will go to waste the first time to ensure a complete wet-out)
• pour into mold, put the lid on, stick in a vacuum bag, and degas under around 29" Hg vacuum (make sure no epoxy gets sucked into the pump!)
• remove vacuum bag, take lid off
• weigh each component of the aggregate and write it down so the epoxy content of the cured product can be calculated
• spread in largest particles first, then next largest, and so on, mixing a bit and leveling it out roughly (not sure about this, maybe mixing it all up first is better, OTOH the vibration step should ensure even dispersal
• cut a piece of fiberglass breather/bleeder ply (pre-coated with a release agent) to the shape of the lid and put it on top of the E/G mix (note that the lid is supposed to have been made to account for the ply's thickness)
• replace lid
• put breather strips around the periphery of the lid so the vacuum bag doesn't get sucked into the gap between the mold and the lid, and so the lid's movement wouldn't be able to damage the bag
• put a breather strip at the top to ensure vacuum gets to all the holes in the lid
• stick in vacuum bag
• fire vacuum pump up briefly and fix any leaks that can be found
• fire it up again and degas, keeping an eye on the vacuum tube to make sure no epoxy gets pulled into the pump
• it might require a couple cycles of degassing, depending on how much air is in the epoxy and how much room it can froth before getting to the pump
• with vacuum pump still running, attach vibrating contraption to lid and vibrate it until the lid's shoulders sits snugly on the mold
• maintain vacuum until the epoxy finishes the first cure cycle, which should be something like 15 hours for Pro-Set infusion epoxies (BTW a vacuum transducer actuating a relay for the pump power can be nice for not running the pump all day if there's a tiny leak somewhere)
• pinch off vacuum line and move the bag and everything inside it (including the vacuum) into an oven which is set as per epoxy technical data (120°F - 180°F, IIRC, depending on desired properties)
• cure for either 8 or 16 hours, as per epoxy technical data sheet

That wasn't supposed to be so long, but apparently I needed to make some notes to myself.

Oh, also (if anyone's still reading), I'm curious if there might exist a substance that can be dissolved into epoxy to temporarily reduce viscosity, and then be removed via rapid evaporation in ultra high vacuum once wetting out is complete. Such a thing would be really nice for this process, but I don't know the first thing about epoxy chemistry. Perhaps I can get my girlfriend interested in E/G (she's a chemist)... Okay, I am in danger of rambling here if I haven't already (not much sleep last night), so I'm going to cut myself off. Sorry for the lengthy post. Hopefully it's comprehensible.

Will

[greybeard]

Will, re "viscosity reducing solvent" that could be vac'ed off. It might be worth contacting Nick at www.tomps.com .(no connection, but I have bought my epoxy from him.)
He sells stuff for casting, but my contact with him has proved him to be very knowledgable about real life casting problems, and he just might be able to suggest a suitable material.
John
ps it hardly needs pointing out that this volatile material is going to go through your pump

[altaic]

Thanks John, I'll give them a call. Regarding the solvents going through the pump, it is a problem; I definitely wouldn't do that to my pump. However some pumps I looked at were designed to deal with nasty solvents... Even so, having the pump I do, I was thinking maybe vacuum a 5 gal. pressure pot, then hook that up to the mold, and when it's full of nasty stuff, take it outside and open it. Rinse and repeat... A PITA, but a perfect wet out might be worth it.

[greybeard]

It has just occurred to me that if you had a "salt/ice" mix surrounding the pipe or whatever between the pump and the pressure pot, you could might condense most of the vapour.
Obviously it would take a bit of juggling (sp?)with flow rates and design, but it might be possible .
Conversely, depending on what the material was, break it down with heat.
I'm thinking off the wall here before I retire for the night.
Does it show ?
John

[altaic]

No, those are good ideas, but they may be too tricky for me to pull off.

[altaic]

I've been thinking about vacuum compaction a bit more. I'm guessing we'll degas at a medium-high vacuum (say 10^-3 torr) and I'd like to use the vacuum in conjunction with vibration for compaction, but afterwards I'm wondering what the effect such a vacuum would have if left that way for the entire cure cycle, both on the cured piece and the curing time. Conversely, the negative pressure causes a heavy clamping action via the lid, and I'm wondering if that might create stresses.

Cameron, I read that you have the ability to do flexural strength testing of 7.5" x .375" x .5" samples, and that you prep the pieces using a diamond saw. Would you like some samples, both large (to be cut down to size) and small (to be tested without cutting)?

[ticica]

I have been thinking about using self leveling (table top) epoxy to get a flat surface for linear rails. I would like to have the inserts sit in the E/G mix held by a fixture above the surface. However, I don't have a precision surface that would actually come in contact with the epoxy to keep it straight. It would rather be suspended above the EG and just hold the inserts in position. Now, this would allow the epoxy to climb (or drop?) as a result of capillary effect. Does anyone know whether this would be a problem? How does one model this scenario to get the curvature of the liquid against the steel to see if it leaves enough clearance for the mounting holes? I imagine this could also be a problem around edges of DIY surface plates.

[altaic]

Hello again all,
I'm so eager to make some proper samples, it's driving me nuts. Some of my equipment was delayed and won't be arriving until Monday, I'm sorry to say. I suppose it'll give me some time to tidy up.

Anyhow, I've been finding more bits of this thread that help give me a better picture of how E/G works. I found this post and the following ones about epoxy surface tension and particle surface energy to be interesting. It definitely makes sense, however I'm not sure where I can get that data; the manufacturers I've found don't seem to publish it. Is it reasonable to suppose that low viscosity epoxies would, generally speaking, also have low surface tension?

Also, I may be getting ahead of myself, but I really like the idea of the Inhance Ti particles to create a very hard ultra-low abrasion E/G.

On a related topic, it seems that it'd be very beneficial to be able to apply a different material to the surface than the bulk material. I'm thinking to spray a thin coat of epoxy on the inside of an empty mold (pre-treated with release), pour in surface aggregate of choice, and shake like shake-n-bake. Perhaps do this a couple times to get thickness up if desired, and then pour in bulk aggregate/epoxy. I'm pretty sure the bulk material will disturb the surface layers, so I'm thinking to let the surface layers set up a bit, and then add the bulk material. I'm not sure exactly what happens to epoxy if one attempts to bond a fresh mix to one that's somewhat set up, however I'd venture to guess it'd be fine. Alternatively, spray the mold with a temporary adhesive such as 3M Super 77 to get a single layer of surface particles, and then fill with bulk material. Is this off the wall, as John might say?

Cheers,
Will