[ckelloug]

I have spoken with lazlo on the homeshopmachinist board who has looked into this and says that Harlow is using accures castings.

[ckelloug]

In general, I have to agree with Andrew that the E/G that's been made here on the thread is way less than the 2e6 psi modulus that I've been using in sample calculations. It's also less than the 4.5 e5 psi incorrect modulus from accures as well as the 4.5e6psi correct modulus.

My formula for modulus based on walter's jig estimated 170psi as the flexural modulus of Walter's sample which is somewhat discouraging and means that it is likely only applicable as a vibration dampener in a steel structure due to what will likely be large deflections.

Given the US Composites 635 epoxy we have and aggregate designs that aren't made of a single type of round sand, the numbers ought to be better than that. I believe that the fact that they aren't, combined with Walter's descriptions of gumminess, indicate that there is either an error in the epoxy mix ratio or some other problem with the epoxy curing correctly. It is also possible that there is some issue with the reactive dilutants in the epoxy.

Based on the data currently available, I would think that it would be wiser to design a router with metal load carrying members filled with E/G (or rubber bagged portland cement) as I said in my first post on this thread and Andrew has said all along. I say this because casting a bad beam or table could be expensive and the reasons for the poor performance in the epoxy tests on this thread are not yet understood.

I realize from Andrew's comments that people might be unsure of my position so I will answer that my idea is to build gantry routers out of whatever works best be it E/G or not. Unlike some in the thread readership, I was interested to explore the idea of E/G gantry router top beams and also, what is ultimately possible for the material itself.

In general, assuming that the epoxy setting issues are solved and 2000ksi modulus E/G is achieved, I see unreinforced E/G as a possibility for 24 inch and smaller router beams (for .0001 deflection with beam weight carrying 100 lb live load) though I still need to do a shear deflection computation. Between 24 and 36 inches, a hollow gantry router beam that weighs around 150 lbs and has a 4x12 inch cross section with 1.5 inch thick sides and 2 inch thick bottom and top seems possible (with .0003 or so static deflection with beam weight and 100 lb live load) although perhaps not desirable. In pieces like gantry beams longer that 36 inches, unreinforced E/G likely qualifies as a complete waste of time.

<B>
In all cases, the risk will be a lot lower to make the load carrying parts steel and use E/G to add mass and vibration damping.
</B>

I'll continue to post what I determine about the material properties achievable and I'll post formulas as I learn more in case anybody is still interested. For all of you out there building routers, sorry if I've led you astray. I still think it will be possible to build most beam like parts under 24 inches in length from E/G once the issues with epoxy setting are resolved. I seem to be posting materials science on a router forum which some people seem to frown upon so I'll make the offer once again to relegate myself to another thread.

Best of luck to all those out there building routers!

[walter]

accures castings, now that is incredible. He had it all along!

[harryn]

CKelloug,
The E/G may work well for you. I think as filler or as a table you are pretty much there. For a gantry, where stiffness is the premium quantity, there are fun things to mess with: http://stores.ebay.com/Honeycomb-and-More has aluminum honey comb for sale.

I wonder what a multi layer sandwich of aluminum sheet, epoxy, honeycomb,epoxy,aluminum sheet, assembled on a flat plate, repeated to a beam 6x6 inches would do. All I know about this is what I read in Nat. Geographic about Burt Rutan and scaled composites, but it is available to the small shop through Ebay.
Keep up the good work.
roger

Sorry to take this thread OT a bit.

It is even more available than EBay actually. Locally here in N CA, I found a shop called http://www.pacificpanels.com

They carry bulk honeycomb, unexpanded commonly in 3/8 and 3/4 cell size. You just tell them what you want in terms of thickness and sheet size, and they will make it for you. You can get it rough cut or precision cut.

They can fill edges, add reinforcements, etc.

As a budgetary number, a sheet of 3/8 or 3/4 in cell size x 1 in thick costs approx $ 8 / ft 2. Example, a 4 x 8 ft sheet, 1 in thick, is 32 ft2 x $ 8 / ft2 = US $ 256 if my math is working tonight.

A bit pricey, but not terrible, and in line with what I have seen on line at other honeycomb panel shops around the country. There is a shop in CO, and probably others around. (just search "honeycomb panel your area)

Deflection - For my project - worst case loading (using their calculations)

7 ft beam x 4 ft wide x 3/4 thick, supported only at 7 ft ends
- no load = 0.024 in deflection
- 100 lbs evenly spread = 0.045 in deflection
- 100 lbs - center loaded = 0.073 in deflection

So, I have a bit to go.

If the normal linear math works on this stuff, a 7 ft long beam, 6 in thick would deflect approx 0.009 in.

[romihs]

Cameron,

I've been following everything you have had to say, some of it may be of use, some of it may not, but at least someone is talking about all the options.

Your research is invaluable, and some of the information you have posted is important for making a good E/G or E/Q mix. With the input you have given, I believe that a fully E/Q machine can be built in the home workshop.

We need to nail down the recipe and process.

I for one will keep away from exotic additives as I believe that a suitable result can be achieved without them. Besides, I haven't a clue were to get my hands on the stuff anyway...


I have stopped with my testing for now, as I need better equipment, and I also need to setup a workshop. I'm currently living a small flat with my wife and kid, so it is not appropriate to do this kind of work in a place like this.

But what I have noticed with the few samples that I mixed up is that 8% epoxy, by weight, is too little, while the 12% mix worked out way better.
The 12% epoxy mix is about 20% heavier that the 8% mix when comparing two samples of similar size, implying that the particles packed closer and tighter together in the 12% mix compered to the 8% mix.
I have not seen the problems that Walter has been having with his epoxy. Then again, I have not used extremely fine fillers like carbon black either...

I will continue to test different options in an attempt to get the best results that I can get.
I hope that you will also continue with your work so that we can together work towards a useful solution.

Thanks again!

Best Regards

Sandi

[LeeWay]

Nice work, Guys. Haven't been following as close as some, but skimmed the cream pretty good.

What I am deciphering is that the testing is bearing out that the mix and ratios that are best are nearing that of standard concrete. Swap the Portland for epoxy and you just about have the formula. The exact size of the aggregates aren't as precise as what you guys are testing when you get a truckload, but they pretty much even out when mixed in volume. Standard rocks, gravels and sands do cover a wide range of sizes. Standard concrete is what I am referring to. High strength and special formulations certainly use a better defined and measured mix like what this group is trying to nail down.

I may have missed it, but since there will be three main uses for the E/G for us, will there be three BEST formula's as well? Are we rather looking for a one size fits all solution?

I do need to fill my mill extrusions eventually, so I am in no hurry to see the final formula, but I will stay adhered to this thread until then. I would love to help with some of the testing. I am just plagued by a popular product that is in good demand and apparently I am the only one in the World that makes em.
That isn't a complaint.
Keep up the great stuff, Guys.

[qroger]

Harryn,
It is off topic, but I was wondering where the advantages of the Honeycomb come in. If you had a 16" wide sandwich with 1/8" aluminum sheet on both sides of a 3/4" honeycomb and turned it sidways, would it act like a solid 1" thick aluminum beam, with the honeycomb just helping to fight torsion? If you continued until you got a 16" x 6" x 6' it might be getting pretty interesting.
Thanks
roger

[walter]

Given the US Composites 635 epoxy we have and aggregate designs that aren't made of a single type of round sand, the numbers ought to be better than that. I believe that the fact that they aren't, combined with Walter's descriptions of gumminess, indicate that there is either an error in the epoxy mix ratio or some other problem with the epoxy curing correctly. It is also possible that there is some issue with the reactive dilutants in the epoxy.

Gumminess is probably caused by carbon black- I'm sure the latest batch will have the same problem. It dramatically affects the curing time..

Samples made with straight sand and 15-20% epoxy didn't have any of that- they ring and feel like real granite.
_

[ckelloug]

Walter,

Thanks for the comment on the carbon black gumming up the epoxy. That's good data and I'll work on a fuller understanding of it just as soon as I have somewhere to do lab work. I had the impression you were having general problems with the epoxy, not just with the carbon black. I don't know whether it interferes with the epoxy chemistry or it's just such fine particles that they get in the way of something. To figure this out, I'd personally try with something like silyated silica fume like cabot TS-530 or TS-610 though I doubt you're interested in that experiment right now.

Responding to your comment about accurately controlling the curing temperature, it's really not absolutely necessary. The main difference it makes is that epoxy that is cured at higher temperatures appears to crosslink more and have a much higher hardness and lower creep as it is exposed to higher temperatures.

DAK3333 do you have any comments on carbon black hosing up the epoxy setting?

[harryn]

Harryn,
It is off topic, but I was wondering where the advantages of the Honeycomb come in. If you had a 16" wide sandwich with 1/8" aluminum sheet on both sides of a 3/4" honeycomb and turned it sidways, would it act like a solid 1" thick aluminum beam, with the honeycomb just helping to fight torsion? If you continued until you got a 16" x 6" x 6' it might be getting pretty interesting.
Thanks
roger

Hi - I have not used this material before. From general reading (and a few mechanical engineering classes 25 years ago)
- Think of a large rectangular beam
- Most of the strength that is actually used is within 2 inches of the surface
- This is why rectangular tubing and I beams have relatively low deflections, not too far from a solid beam in many cases
- Hollow tube sections tend to buckle at point loads and defects / scratches
- Filling the inside with structural foam, honeycomb, etc, tends to reduce buckling.
- All of these also tend to dampen vibration
- If the interior material is similar to the skins in structural strength, and there is a lot of bond area to the skin, the shear strength is very high
- Small honeycombs tend to make very strong, high shear strength panels, with quite low weight
- Modern adhesives that are used in the bond lines make these very tough, strong, and stiff.

Al is very common. Fiber filled high tech epoxies are available with even better properties, but the price is nearly unrealistec.

[walter]

I don't know whether it interferes with the epoxy chemistry or it's just such fine particles that they get in the way of something. To figure this out, I'd personally try with something like silyated silica fume like cabot TS-530 or TS-610 though I doubt you're interested in that experiment right now.

I don't know.. I sort of liked my early results.

I might go with a barebone system just to get it off the ground. I'm thinking Quartz only, three sizes, zeeospheres for color.
I need to visit Agsco and see what's in their 'Quartz' department.
_

[ckelloug]

Walter,

I realize you're interested in bare bones at this point and I don't want to distract you so I'll be succinct. Puffed wheat (Malto-Meal Brand of course) would almost work as aggregate except the strength is a bit low. The mix of stones you showed in your last pictures looked great regardless! My cabosil suggestion was just my idea for figuring out if carbon black hinders the mixture setting, not advice to you necessarily.

I can say that you're on the right track using multiple sizes of aggregates. If you're not aiming for a heavily graded aggregate approach like calling agsco, just use several sizes available from your favorite sources. The guys at Huntsman use 3/8 gravel, #100 sand and fly ash which is probably as fine or finer than zeospheres.

As harryn pointed out in post 1150, washing the aggregate in 91% denatured alcohol from the pharmacy and drying it ought to do quite a lot in terms of making the epoxy stick better and may turn out almost as good as silanes.

In short, if the epoxy sets correctly, you should have reasonably good luck with a mixture that uses multiple sizes of aggregate. A higher percentage of larger aggregates will have less chance of having mixing or adhesion difficulties though it will be harder to cast into complex shapes.

In short, getting reasonable strength should not be hard with multiple aggregate sizes and the US Composites 635. Getting the highest possible strength will be harder than reasonable strength but likely unnecessary.

The CNCZ-1 formula I posted was my attempt at the most advanced formula reasonably possible. I thought in those terms because if that formula fails to show serious strength after carbon black and epoxy are optimized then the whole approach may be wrong. If it works, then it is time to see what must be subtracted or changed to make it easier for everyone to do.

In short, Walter, in a stroke of brilliance, you extracted the exact right lesson with your comments about aggregate: aggregate isn't as important as getting the epoxy to stick. Bigger aggregate should have fewer problems.

Also, take a skewer or some small rod and poke it into the the mixture in the mold after it has been poured every few inches to help consolidate the mixture.(rodding) You might also try pouring only an inch or so and then vibrating before pouring the next inch. These are two techniques mentioned by one of the experimentalists in the the 1975 conference proceedings.

I'm anxious to hear how things turn out and I hope I haven't wasted your time with any of my ideas and suggestions.

[walter]

I wasted your time, actually. It was your suggestion to optimize the mix for my app but I wanted 'the best'. Turns out 'the best' takes time and at least 100 samples to test&optimize each component. I can do 3 per week.

I'm just going to put it on a back burner.

Really appreciate your work!
_

[davo727]

Hey Walter, do you want me to make a test block with West system epoxy #105 epoxy and #206 slow hardner and send it to you? Dave

[brunog]

Cameron,
Does shore hardness of cured epoxy have an effect on overall hardness of E/Q ? it just might be an important "ingredient" that's being forgotten in the recipe.

Best regards

Bruno

[ckelloug]

Bruno,

My understanding is that shore hardness doesn't directly correlate with the strength of E/Q although it may play a role in the abrasion resistance. Hardness also sometimes implies brittleness. I assume if the surface of the material has little exposed aggregate and lots of low hardness epoxy then it may not do too well in terms of usability.

<B>One of the materials scientists on this thread might have more insight. </B>
--Cameron

[brunog]

Cameron,
I am just wondering what difference a shore D hardeness of 65 vs 85 would make in compression and tensile strength of E/Q.

Best regards,

Bruno

[walter]

Hey Walter, do you want me to make a test block with West system epoxy #105 epoxy and #206 slow hardner and send it to you? Dave

Not a bad idea. We could compare different formulas and epoxy types- and test them with the same "setup". Thanks Davo! I'll think about the sample size and send you info this weekend.

If anyone else wants to participate, just let me know. I don't really have proper "tools" but at least we can put some numbers on these samples..

[DAK3333]

Carbon black is used mostly for tint and "is" a dispersion hardener in films( where thickness is near carbon black diameter). Fumed silica is a better choice in the larger sections we are considering.

A lot has happened while I was out.

That Dow z4060 6040 whaterver doesnt seem to be to bad on the face but it still has a trimethylsilane group with a protection functionality (the methanol producing component) it is still not a neccessary component.

Here is a question. So far I have seen talk of square beams what about I beams or other geometries founf in extruded profiles. I know harder to cast but in the macro world are there advantages to be had.

Vaccum....

[ckelloug]

DAK3333,

Thanks very much for this wisdom. Plug me into your firehose of knowledge! I are not a chemist and thus tend to goof at that scale much more often. I believe dispersion hardening is a good thing and I've also been reading Gamski's paper from 1975 who wants to use 3000-5000 cm^2/g material to prevent the material hardening with built in strains from shrinkage. I assumed carbon black might be good although the specific surface area is a bit higher than Gamski asked for but, since a number of us had some in stock, I figured it would be a start. Do you have any other ideas on dispersion hardeners besides silyated silica fume like cabot TS-510 or TS-630 and the nanopox from www.nanoresins.com?

I'm still trying to understand for academic reasons why the carbon black samples are not hardening well. Do you think that carbon black is adsorbing either the epoxy or the amine based hardener?

The Dow z6040 and the BYK-A525 are ideas I got from the Reichhold apps engineer and from studying sample formulations form commercial materials. My reasoning was to see if they would produce a mixture that would deair and consolidate with either less or no vacuum. I assume that adding them will produce the absolute strongest material for experimental purposes to help gage what kinds of problems we're having in getting the material to actually hold together right now. Do you have any comments on the BYK-A525?

My current theory is that with the smaller aggregate sizes we're producing a material that has mainly the epoxy carrying the load and that the aggregate are mostly surrounded by voids and not working correctly.

As for gantry top beams, you are right about other profiles specifically rectangle and rectangle tube being better. I did the calculations with square beams mainly because it was simple and gave folks some intuitive feel for the materials for those who don't have a complete grasp of moments of inertia and flexure. (After Martin's correction, this seems to include me).

Specifically, brunog's design of a 4 inch wide by 12 inch deep beam 36 inches long seems to be good and close to what home shop guys could cast at 1ft^3 and it also has static deflection of about .0003 under 100lbs static load which could be nulled by post-tensioning. My calculations all assume a flexural modulus of 2000ksi which is half that of commercial material. Since I've been doing the sample calculations, I still need to update my model to reflect shear and the shear between the layers in the beam as it deflects. I should also point out that I've been doing calculations for simply supported beams.

Deflection is inversely proportional to the third power of the depth of the beam in rectangular beams so profile does matter much. In square beams, deflection is inversely proportional to the fourth power of the side length. Square is suboptimal for minimum deflection but it's easier to visualize. I believe I got this right in calculations I have done, just not here.

Deflection=FL^3/48*E*I where for a rectangular beam as Martin points out, the value of I=bh^3/12 where h is the depth of the beam.

Finally, my last theory is that the dow Z6040 and or the BYK-A525 might allow the beams to be cast without pulling 29.5 inches (or any) vacuum on the mixture. The vacuum requirement is rather bad for the home shop guys but it has been in the formulations I've got back to the 1970's and earlier and seems to be borne out by Walter's data that his samples are much less strong than would be expected.

Your reply is the best thing to happen all week on this thread! It's a lot more fun than hand trenching for sprinklers anyway.

Comments, especially on the role of additives would be much appreciated if you get a chance.

--Cameron