[ckelloug]

I just got back from the lab. I took a density measurement on a similar mixture to the one I was discussing with Dave as of post <A href=http://www.cnczone.com/forums/showpost.php?p=578984&postcount=3534>3534</A> The only difference is that I substituted IMSIL 1240 for G200 Zeeospheres as I don't have any G200 on hand right now. The IMSIL is not spherical so I suspect that it won't be quite as good as G200 although it is in approximately the same size range.

My measured uncorrected Phi value was 81.9%. This means that the mixture was occupying 81.9% of the volume measured in my test apparatus. With corrections for container size etc, I suspect that the true packing density will be between between 81% and 88% when I finish the correction calculations.

I intend to do a test on this mixture using 18% epoxy by volume with Hexion 813 and Isophorone Diamine as the hardener to see what the properties are. My big steel vacuum infusion mold is currently occupied with a vacuum infusion test of titanated epoxy and Novec 4432 with coarse aggregate so I haven't made a sample of the E/G reference mixture yet.

My lab notes for the day are that the reference aggregate mixture compacts down almost like a brick in the density measurement apparatus. I also noticed that using the Novec 4432 fluorosurfactant is a minor disaster without a defoaming agent. The air in the mixture stays around forever and even the tiniest amount of the epoxy mixture wants to turn to foam in the vacuum chamber. My conclusion is that Novec 4432 without a deairing agent like BYK A-525 (which I don't have yet) is likely to cause serious problems: Don't bother making up a batch with Novec if you don't have a deairing agent.

On another front, I pulled a couple papers from the Journal of Applied Polymer Science and they suggest that the glass transition temperature for IPDA hardened epoxy is around 160C. This means that for maximum reaction completion, IPDA needs to be cured at that high a temperature.

The other thing I found in these papers was that high temperature cures can lower the modulus of the epoxy even though they increase the glass transition temperature. It appears that long high temperature post cures greatly improve the fracture toughness of some epoxy mixtures.

Just in case anybody thinks I've given up around here, I haven't. It's just taken a long time to get everything on track and find the time to do the tests that needed to be done.

Regards all,
Cameron

[messar]

Hi,

I'm jumping into this rather late, but I've read through all the pages, and am quite excited by the possibilities of being able to cast a machine. Just so happens that I have a perfect application in mind..

One of the big gotchas though is that I'm in Malaysia, so materials may be hard to fine. I've contacted Hexion, but not heard back from them.

I did some further searching, and found some of the sika materials designed for epoxy floors. This is the web site I'm looking at http://www.sika.com.my/construction-...mentlining.htm

In particular, I'm looking at the Sikafloor-261. It appears to be designed for use with a filler (down at the bottom of the page, there are also graded fillers available - this will solve another problem for me in finding these).

Can anyone comment on the physical properties, and if there are any obvious problems (pot life looks like it may be an issue...not sure how this really compares.)

Any comments would be appreciated.

Thanks,

Matt

[greybeard]

Hi Matt.
Big jump, in more ways than one from Oregon to Malaysia
I was in Seremban for 5 years a lifetime ago, but I can remember the sand around the tin mines very clearly.

Looking at the Sika spec sheets, it might be worth inquiring from your nearest branch if they can supply other gradings, so you can finish up with the mix you want.
I realize the final recipe is still being 'tweaked', but depending on the thickness of the components you might want to produce, it would be very convenient if you could buy in just the sizes you needed.
John

[ckelloug]

Hi Matt,

Sikafloor 261 doesn't look too bad but they don't give a number for modulus of any kind and I can see in the chemistry it has a few things that bode poorly for it. The strength number is pretty close to what I would expect but. . .

Looking thought the MSDS however, this is basically the same epoxy as Reichhold 37-127 which uses C12-C14 glycidyl ether as a reactive dilutant. Based on some initial testing but maInly reading, I prefer epoxy with a cresyl glycidyl ether reactive dilutant as such epoxy is stiffer than that with C12-C14 glycidyl ether.

I'm not knowledgable enough about hardener to analyze it thoroughly but I can say that it contains several phenols and ethanol which tend to lower strength and hardness. On the plus side, a major hardener component is isophorone diamine which we know from roach's postings and my lab work works well.

I'm going to make a guess that this Sikafloor 261 is going to make a product that isn't quite as stiff as the pure isophorone diamine I've been working with although I don't know how to balance the effects of the C12-C14 glycidyl ether in the epoxy, the methylamine and the phenol accelerators in the hardener, and the short pot life in the datasheet to know for sure what the properties are like.

The killer problem is that the 20 minute pot life at 20C is so short that I'd suspect that you don't manage to get the E/G out of the mixing pot and into the mold in time to avoid gellation. I'd be worried that it set up rock hard in the mixing vessel before you were organized TO POUR IT based on the time it takes me just to vacuum deair my epoxy.

Unfortunately, the big three: Hexion, Huntsman, and Reichhold don't usually sell in less than barrel quantities and then it is through the local distributor. I've managed to convince them of sufficient seriousness to get enough samples to continue my research but I'm eventually going to have to pick one and order a bunch. They want to sell rail cars full so it may be tricky to find a way to buy less than 55 gallons of something though you might get lucky and obtain enough samples to debug your process before trying to build the machine.

Based on a quick look on the web, it looks like Hexion and Huntsman both do some business in Malaysia. I'd suggest calling Hexion in that e-mail doesn't always work well with epoxy vendors. Beyond that, I don't have any contacts in Malaysia so other than my standard offer of free testing for any samples you make and send me, there's not much else I can do to help you.

Good Luck,

Cameron

[messar]

Hi Cameron,

Thanks for your exhaustive analysis. I'll contact Sika on Monday, and see if they have any more information about the physical properties, but more importantly the pot life. If they sell it here, I can't imagine people actually use it with that short a pot life - it's hot here all the time

My initial attempt to contact Hexion was through their main web site - I did look more closely yesterday, and found the distributor into, so I'll be contacting them next week as well. I'd really prefer a small quantity of the material, but I guess if 55 gal isn't horrifically expensive, I could go that route.

Do you know what Huntsman materials would be most appropriate? I'll contact them as well.

Finally, I will take you up on the offer for materials testing if/when I get things going - it's very generous of you.

Thanks,

matt

[semi-lucid]

Cameron

A long time ago in this thread, T. Zietz stated "forget vacuum".

http://www.cnczone.com/forums/showpo...postcount=2561

I was wondering if you've had the opportunity to test any samples from him, or if you have any thoughts about the quality he might be achieving?

Best regards.

[ckelloug]

Hi semi-lucid,

I have unfortunately not tested any material from Thomas Zietz. He has the advantage of having way more experience actually making stuff than I do and as usual, I tip my hat to him for getting where he has gotten. Of the two of us, he is also the one that actually has built machines. See http://www.thomas-zietz.de/

Ending my preface, I agree with harryn from <A href=http://www.cnczone.com/forums/showpost.php?p=299633&postcount=1330> post 1330</A> that degassing the epoxy before use is good practice. If you are casting large parts that are plenty strong, it may not matter. I have up to this point cast test specimens which are very thin and which could be spoiled to the point where they do not produce valid data by a few bubbles. Thomas also said he is using commercial material which may have self-compacting properties that the materials I have worked with do not. In short, Thomas's machines look great and I can only assume that he's overcome any challenges he might have.

Being careful not to detract from anyone else's work, It is my opinion that vacuum degassed material with zero percent voids will be the absolute stiffest.

In addition to degassing, there is also vacuum infusion vis-a-vis the discussion by bloefeld on <A href=http://www.cnczone.com/forums/showthread.php?t=30155&page=279> page 279</A>. The experiment that I did yesterday with titanated epoxy and fluorosurfactant has demonstrated that a void free plate can in principal be cast using vacuum infusion at 64% aluminum oxide loading and no vibration at all. I don't think vacuum infusion is applicable to all parts or helpful to the majority of readers here as it is difficult and proper hard molds are likely to be expensive if the difficulty in producing my sample mold is an indication. It does provide a demonstrated mechanism to produce void free parts up to at least 64% filler loading and also provides an excellent way to cast samples for testing the mechanical properties of raw epoxy.

In short, Thomas found vacuum wasn't good for his production process. When I have something to produce and a production process, I might find that I agree. In doing laboratory scale experiments, I've found that vacuum has its place both for degassing which I think is always good practice and infusion which may or may not be a good way to make all parts.

Regards all,

Cameron

[semi-lucid]

Cameron

I didn't mean to put you in the position of critiquing Thomas's work.

I seems to me that you are taking a scientific approach to finding the best possible material property's, where as Thomas may be taking a pragmatic approach to getting work out the door. There's nothing wrong with either one of these approaches, and they are sort of two different things.

I was wondering if Thomas was accepting material property's that were adequate, but less than the theoretical optimum, and making his cross sections adequate for the for the strength that he is achieving.

The only way to know this, would be to test samples.?

It's not my intention to criticize Thomas or anyone else.

Thomas has been generous in offering his comments in this thread, and showing pictures of his molds, and detail of his machines. Looks like nice work.

Best regards.
John

[ckelloug]

semi-lucid John,

Most anything can work if you can make the cross section big enough and the machine small enough. The problem is not in the strength, at least not for metalworking tools. As near as I can tell, the problem is that if the material is not very rigid then things deform under their own weight which is not very appetizing for machine tools.

I'd be the first to say that my own scientific approach probably constitutes over-engineering. I would like to believe however that all this insanity will eventually produce an easy to use material that is stiff enough and strong enough to make the smaller cross section parts that are found in home shop style small machine tools.

I'm not worried about comparing stuff to Thomas Zietz's work, but after some flaming I took a long time ago here, I always make an effort to state clearly that I am doing lab work which hasn't yet led to a usable product and that Thomas has a product. I suspect also that Thomas has performed standard German engineering and his gear is as accurate and precise as it looks but I haven't tested it. (I don't actually know Thomas other than our few interactions on this thread but I respect him as he was a predecessor on the same path we're on here.)

Thomas pointed out a long time ago that he wasn't interested in material development and that he had had bad luck developing his own materials. I'm here to point out that it is a difficult path but with the hundreds of people who have contributed to this thread, we are actually making real progress.

So the final answer is that I think Thomas's material is likely to be less rigid than the optimal material. With the right engineering, this is merely an interesting fact with no relevance to actual machine performance. I believe that for quartz that the maximum achievable flexural modulus is about 6000ksi. I know that Accures and Zanite achieve about 4900 ksi. I assume Thomas's material has a modulus of between 3000ksi and 5000ksi. My last set of calculations makes me believe that materials with moduli less than 2000ksi are fairly useless.

What I really want to know but haven't worked out yet is whether making 6000ksi material is any more costly or difficult than making 4000ksi material beyond the fixed cost of designing the material. I also want to know whether materials with the highest achievable modulus and density are unworkable and what the tradeoffs between workability and modulus are.

Most of all, I enjoy learning all the new things people here have either taught me or forced me to learn and the camaraderie here.

Regards all,

Cameron

[SolarSystemsLLC]

Hey guys, I've been lurking this thread for a while now and thought I'd give you a heads up. I've got a thread going on an Epoxy Granite application that you might want to follow. http://cnczone.com/forums/showthread...348#post586348

It didn't seem like something that should pollute your thread, but you might be interested to watch this in parallel. My goal is to make a medium sized EG surface plate, but the requirements don't really line up well with the approaches taken in this thread.

The formula for the EG isn't an issue (yet), and hence it's more of a technique that might be useful for whatever formula you do come up with.

Thanks,
Thom

[ckelloug]

Well,

I just got back from the lab. My first E/G sample is curing. My first problem was that I plugged the wrong density for aluminum oxide into the volume calculation so I thought I had a higher density than I did. Phi for the mixture from a couple posts back is 76%-77%, not 81% at present.

I mixed up a batch of the stuff using 24 percent epoxy by volume with IPDA. The mix was somewhat dry. I added about 1% of the weight of the aggregate in LICA 38 titanate 3 separate times and the mix seemed less dry. I then deaired the whole mixture in the vacuum chamber (after having deaired the epoxy beforehand) and I got something like those 4th of july snake fireworks where the ash rises from the pellet. The stuff seemed to flow pretty well under the influence of pressure even though it seemed pretty dry just mixing it.

I forced the mixture between two nesting mixing containers to produce a 3/4 inch thick puck about 4 inches in diameter and it is currently in the oven at 85C curing.

In short, it looks like the aggregate formula needs a bit of tuning. There appeared to be no problems with wetout even before the addition of the titanate. It's possible that the "coating the surface area of the particles criterion" from way back when will play a role in requiring for epoxy than the fill the voids criterion I have been using.

For the moment, the mixture is at 24% epoxy by volume not corrected for container size. We'll see how it turns out in modulus and strength although improvisation will be necessary since this mixing properties sample is too small to do the normal tests.

Regards all,

Cameron

[ckelloug]

Coming to a theater near you!

[romihs]

Hello all you guys out there...

I came across a very interesting CNC mill design built on an EQ table.
http://www.meisterkuehler.de/forum/w...aschine-3.html

I know that the topic of EQ surface plates comes up quite often on this thread, so I thought some of you guys might be interested in this.

I have not read the posts, my German is not that good... ;-)

Regards

Sandi

Edit:
After looking into the thread some more, I realized that the builder in fact used plain concrete and as I understand it, not EQ.
What led me to believe that it was EQ, are the obvious different layers in the base visible in some of the photos.... The design is non the less inspirational.

[the4thseal]

dammmmmm Germans!!!!!!!!! if only i could read german

[unclejoe]

Heinrich and Harbech ( VDF Lathes)
The VDF brand CNC lathes for years have been mounted on a concrete base, we specialized in removing the machine from the base refurbishing the slides and replacing the slides onto the base, we did this with two part epoxy.

uj

[brunog]

Sandi,
Very interesting thread, some guys have much more than a do it yourselfer's budget!!

I was able too browse through the thread and understand most of the info thanks to google's translation tool. The translation is not perfect but the logic is there.

Best regards

bruno

[ckelloug]

Drat.

While the puck produced a few posts back seemed pretty strong, it had a fair number of air bubbles. Upon reflection, I should have realized the software optimizer was shooting out bad results when I saw that the largest component #6 Al2O3 didn't have the largest percentage.

That mixture I was discussing with Dave back at post 3534 is wrong. It contains way too much #2 Quartz. The problem stems from having the optimizer programmed for a special case involving a small number of specially selected sizes instead of the general case where there are lots of sizes. I'll post a new mixture when I unbreak the code. Sorry for the bad data.

Regards all,

Cameron

[MaX-MoD]

Hello everybody,


today a guy on usinages.com has pointed out a Mechanics Doctor's thesis, guess what, on polymer concrete!
Yet it is in French, it is very complete (232 pages!) and covers theorical approach, tests of EG samples, cost factor study, mixing techniques, vibrating techniques etc.

it is here : http://www.usinages.com/post86232.html#86232

An important thing I noted in this document (but I went very quickly through it) is that the vibration of the material was capital, and that the acceleration more than frequency was the most important parameter. It states that vibrating over 2G of acceleration gives much better results than a long vibrating at less that 2G, even by tweaking the frequency.

as far as I see, he has not used vacuum at all.
Magic vibration?

[ckelloug]

MaxMOD,

I'm mad now I've found and implemented as many models from the literature as well as coming to the same conclusions and I'm not gonna get a Ph.D. for it.

I'm not going to translate the whole paper but I'll summarize the salient points:

The guy mentions the importance of coupling agents and deairing agents. He used BYK A530 for deairing but he decided not to use coupling agents like silanes and titanates. He doesn't say why.

He designed his aggregate mixture with Fuller's formula and by addition of water, he measured the voids present at 19.2% in the uncompacted mixture. He predicts the resin by volume required for his 34GPa modulus design requirement as being 14% i.e. 86% aggregate by his rough model which is 92% by mass. He also says that the problem is difficult if one has to meet the 34GPa modulus constraint he came up with as his goal.

He uses Huntsman Araldite BY 158 Epoxy with Aradur HY 3740 hardener. Reading the datasheet for HY3740, it mentions Araldite BY 3739 as Huntsman's EG resin with 3740 as Huntsman's EG hardener.

He measures tensile and compressive moduli of about 40GPa and an extraordinarily linear stress strain curve. He also notes that using calcium carbonate vs. microcrystalline silica as the smallest aggregate has no effect on the modulus or strength. He also measured creep and believes the results are sufficient to make the material very satisfactory for the manufacture of machine tools.

He mentions as an afterthought the possibility of adding carbon fiber to utilize it's negative coefficient of thermal expansion to lower the thermal expansion of the material.

He says that an acceleration of 2g during compaction is the boundary between getting good and bad compaction. He vibrates his samples on a large 100mmx 600mm vibrating table which achieves 2g acceleration. After the mold is filled he pokes the surface with a pointed steel rod to help release air and then vibrates for 45 minutes. His resin only exotherms by about 2 degress C. 90% of the strength is obtained within 1 day and 98% of the strength is obtained within 7 days.

He says his tests with 8% and 9% resin by mass turned out with few voids and no segregation while 10% resin by mass caused epoxy to pool on the surface.

He says the order of adding components to the mixer is very important.

His mixing procedure in a large electric mixer is:

1. Mix the epoxy and deairing agent, for 1 minute.
2. Add the large aggregate. mix for 1 minute.
3. Add the fine aggregate. mix for 1 minute
4. Add the sand aggregate. mix for 1 minute
5. Add the micron sized aggregates. mix for 1 minute
6. Remove the bowl from the mixer.
7. Mix by hand for 1 minute.

Place in the mold. Poke with pointed rod. Vibrate for 45 minutes. Demold after 24 hours and test after 7 days.

My Notes:

He predicts modulus with the Hashin Strikman model which I mentioned back in <A href=http://www.cnczone.com/forums/showpost.php?p=438099&postcount=3075>post 3075</A>. I abandoned Hashin Strikman which isn't very accurate at high filler loadings for the model from Kinloch: Tell me that the graph from post <A href=http://www.cnczone.com/forums/showpost.php?p=444339&postcount=3119>post 3119</A> doesn't look an awful lot like the graph from page 78 of the Castillo German Thesis MaxMod posted (page attached as a pdf).

He predicts that 86% aggregate by volume is needed. I predict 88% aggregate by volume is needed.

He mentions coupling and deairing agents. I and others mentioned coupling agents and deairing agents. Pupa mentioned silanes back on <A href=http://www.cnczone.com/forums/showpost.php?p=280684&postcount=737> post 737</a> and I rementioned them and deairing agents after locating Dow Z6040 on <A href=http://www.cnczone.com/forums/showpost.php?p=301391&postcount=1351> post 1351</A>.

The vibration is necessary because the aggregate won't pack correctly without it whether or not epoxy is involved.

The microvoids located in figure 4.12b of the paper could have been eliminated with the inclusion of silanes or titanates from what I've read.

Regards all,

Cameron

[harryn]

dammmmmm Germans!!!!!!!!! if only i could read german

All is not lost - there are free on line translators that are not perfect, but reasonble. Babblefish was an early one, and of course google has "borrowed" this concept as well.


Altaic - Please consider to put a "cold trap" between your vacuum pump and the EG mix. There are actually two reasons for this - diffusion from both directions.

Except for a very few pump types, the majority of vacuum pumps rely on a thin film of oil to make them actually air tight. At low pressures and the nice warm environment inside of the pump, the oil will have sufficient vapor pressure to backflow into your experiment. This can be reduced by using a synthetic vacuum pump oil like fombolin or crytox (sp) but you probably will not like the price (> $ 50 / liter last time I looked), nor one other little side effect. These synthetic oils are such good lubricants and so good at getting into little cracks that they tend to leak out of even well sealed pumps - all over the floor. You will find out quickly just much science has advanced lubrication if you step on it.

For the same reason (oil vapor pressure) and epoxy "stuff" vapor pressure - make sure you vent the pump outside.

I will assume that your pumps are not being N2 purged, so I will throw in one more user experience - they can blow up. The pump oil can absorb and desorb enough of the fumes from what you are pumping to reach non - steady state air / fuel mixtures. In a hot pump, these can be auto ignited, resulting in remarkable effect. (guess how I know) Make sure to give the pump oil time to boil off these interesting mixtures you are feeding it before pumping on a lot of air. (10 - 20 minutes between cycles)

The part about epoxy vacuum degassing that gets tricky is that while it is "generally a good idea", it depends a lot on what is coming out. If you are degassing one of the important components from the mix, this is not helpful either.

Cameron - I cannot believe you remembered that old post I made about the candy bar wrapper. I had forgotten about it.