[Zumba]

Cameron, Good info!

In regards to wetting, yes you are right that small aggregate make the mix very dry. As I posted earlier, anything above a 5:1 weight ratio (using sand or finer aggregate) is quite dry. I think a 6:1 ratio is possible but it will require vibrocompaction. 7:1 and higher results in a difficult to pour mix.

As for vacuum, keep in mind that your mix will probably be 12" deep inside a 5 gallon bucket. It may be possible to de-air raw epoxy like that, but once you mix in the aggregate, there's absolutely no way. You'll need to get the stuff constantly moving to the surface, like a blender, to let the vacuum do its work.

Some people here have mentioned mixers of sorts. Keep in mind that the mixer will have to fit INSIDE the vacuum chamber, and you'll have to CLEAN IT after use. Thus far, the only practical solution I've thought of is vacuum-vibrocompaction. A 5 gallon bucket can be placed into a vacuum chamber (or made into a vacuum chamfer itself with some creativity). Then you place it on a vibrating table and let it shake.

All this theory of requiring 29.## inHg doesn't mean much until you can figure out a way to implement it.

[ckelloug]

Eson,

Thanks for posting the pictures from the German forum. Since I don't read German, I was intimidated and didn't ever go look. Taking a close look at the last picture you posted, seeing the aggregate design they're using, I think I understand why they have metal internally. From what I believe is true about the material properties of stones, stones like that get more prone to cracking as they get larger. I predict that those German E/G mixes aren't particularly strong. Even in those pictures, it looks like the majority of the metal is present for intrinsic functions of the machine: reasons other than strict reinforcement.

You are absolutely correct about epoxy not sticking to cured epoxy. It's my understanding that the epoxy may stick for 12 hours or a day after it gets somewhat hard but after that it won't bond well. Furthermore, IIRC amine blush can leave a waxy layer on the epoxy and make even relatively quick applications of more difficult under some circumstances.

Care to comment Larry?

--Cameron

[Fireflowers]

Hi not sure if this has been mentioned as this is a long thread!!
The Hermle C50 U has a mineral cast base. This is a seriouse cnc mill with upto 60kw spindle power and 2 tonne max table load!!
Not sure what's in the mineral cast..? but it goes to show the results that can be achieved with this type of cast material!

www.hermle.de

John

[ckelloug]

Zumba,

I don't have an excellent solution to vacuum for a large part yet. My first order of business is to figure out how to reliably make ASTM Flexural Test Specimens which are only a few inches long and have the desired material properties. There's no sense in making a lot of a material with lousy properties. . . if you can make only a little and know what you have.

As for vacuum deairing the material for an actual E/G part, I think I'd probably do it in small batches. I'll probably get a vacuum oven for testing purposes as I need one for some other work I've been doing. The US Composites 635 Epoxy has such a long setting time that one should be able to make several small batches. It will undoubtedly have to be in a fairly deep vessel inside of a pressure vessel to avoid the pseudo-boil-over effect discussed by larry et al. I've been thinking of making a vacuum chamber out of a decommissioned welding oxygen cylinder and using a length of 4 or 6inch DWV pipe as a vessel inside the long deep chamber. DWV pipe and caps are cheap enough that they could be considered disposable. I believe greybeard actually talks about having made a vacuum chamber out of DWV pipe.

I still think that bonding agent treated aggregates and silane compound based deairing agents are apt to make this deairing process work much much better than pure mechanical means. As I get closer to actually making something other than noise on this topic, I will also look into flow agents like the zeospheres Larry recommended to walter for improving the mixture's flow qualities.

Keep up the good work Zumba and if you have any more details about the experiments you've done, I think there are a lot of people on this thread that would like to hear more about them.

Regards,

Cameron

[lgalla]

Cameron,secondary bonding is a problem if the pours are too far apart.Epoxy does not stick to cured epoxy.Cured epoxy is so inert or chemical resistant alternate layers can not "bite"into the cured epoxy if you go beyond the recoat window.
I did not want to post my solutions,as I consider them Trade Secrets.
A small % of milled glass,1/32 or 1/16 in the mix,or kevlar pulp helps bonding.When you sand the cured mix,the epoxy sands away leaving micro fibers exposed to assist in secondary bonding.I assume the fibers will also up the E/G properties.To much fiber added will only "thicken"our other problems.
In answers or observations to the last 10 or so posts.
Vacuum negative pressure is very powerful.An air compressor tank will collapse under full vacuum.I think ABS pipe can withstand the pressure,but the surface area of a 4" pipe is not great enough to de-air.More surface area is a benifit.28Hg and mixing like concrete,tumbling or rotating should do the trick.Paint tanks I have mentioned before might have a built in agitator or mixer.E-Bay is good to find one
Rebar pre tension?Somewhere in the early posts this was discounted as a good idea for E/G but good for concrete bridges.If re-bar is necessary on a gantry,why not use a steel tube filled with E/G instead?
I f I lacquer one side of a wood table top it will warp.I assume if re-bar or carbon fiber is added to a beam,it must be the same on opposite surfaces to balance.
To me carbon black is a pigment.I have used extreme amounts and had no curing troubles.
I am very poor at explanations and re reading my posts I hope members understand.Cameron is good at translating my point.Thanks,Cam
Larry

[Mazaholic]

I would think a paint pot would stand up to a vacuum.
But some tanks are designed to keep their structural integrity while pressurized as apposed to a vacuum.
I also found a ventury vacuum pump capable of 28 inHg at 20 cfm.
I'm wondering how fast the cfm falls off as the vacuum increases.
Also wondering if it's a good idea to put the aggregate in the pot first and draw the vaccum..the aggregate will release it's air almost instantly.
Have a ball valve installed near the bottom of the tank and put the epoxy in some kind of large syringe that will screw into the ball valve.
While the aggregate is in a vacuum state,open the valve and release the epoxy into the tank....Tumble for a few minutes then vibrate to insure the aggregate is settled to the bottom and completely covered with epoxy,then release the vacuum.
Does this sound like overkill?

[PlasticWorker]

Sorry to contradict but:
I use an ASME air tank for full vacuum and it has not collapsed in 20 years. I have collapsed a 50 gal steel oil drum when I tried to use it for an extra tank

The reason wood warps when you only paint, laminate, etc to one side is as the wood gains or loses moisture it expands or shrinks and since only the uncoated side does this the stress causes warping.

A 10" x 10" steel tube filled with E/G is probably the best construction, look at Komo Routers. I don't know if their tubes are filled ( I would guess not) but they look like square steel tubes. But this thread seems to be focused on not using much steel, if at all. I thought it was an expense thing but maybe not.

Mazaholic
If you are talking about a commercial steel paint pot, yes it will hold the vacuum. we use one to degas resins all the time.

Subjecting the aggregate to vacuum before the introduction of the epoxy seems like a good idea. If there isn't any air it obviously can't get trapped.
Thorough mixing might be a problem. One other advantage, as you introduce air back into the tank the pressure should help force the epoxy into the aggregate. You could carry this idea one step further and go to Resin Transfer Molding (RTM). Create a sealable mold, put the aggregate in the mold, compact it and close the mold. Pull a vacuum from one end / high point. When evacuated introduce the epoxy into the other end / low point. When the epoxy comes out the other side, stop. You could also put whatever other reinforcements you like in the mold

[ckelloug]

PlasticWorker,

I haven't worked with vacuum or done the calculations but I'd figure an ASME air tank would be decent. I've heard some stuff on another forum I visit where some import model air tanks have fake ASME stamps and are not particularly safe for either compressed air or vacuum. I'm interested in using the bottom half of a decommissioned oxygen cyllinder which in it's previous life would have been rated to 2500 or so psi with a safety factor that is probably mind boggling.

I would agree that a 10x10 steel tube is probably an excellent solution mechanically. Filled with E/G and post tensioned to get rid of any deflection, it would be almost perfect although a bit heavy if any part of the gantry needs to move in a design with it. I'm personally trying to push engineering limits on E/G or at least epoxy whatever composites to the point where a safe good performing machine can be built without steel. There are a lot of people on this board who don't want to cut big metal as part of their project.

I would want to do some kind of analysis however to determine the side loads while cutting as 10x10 is a huge cross section and most of the stuff I'd worked on with Bruno and others was more about a 12x4 cross section which uses about half as much material. It might not be necessary to go 10 wide to meet safety factor concerns; especially on the smaller machines in the 26 inch range. I'm not building routers so I'll take cues from what people think is practical.

I agree with the comment about lacquering a tabletop, it's a moisture problem, not an asymmetric reinforcement problem. I think Mazaholic has it absolutely right however when he says put the aggregate under vacuum and then add the epoxy and mix. It might also be less of a mixing problem to add the epoxy to the vacuum chamber and add the aggregate from the top. Not sure.

The big problem is not so much air introduced during mixing so much as bubbles of air around the aggregate such that each grain of aggregate adds a chance at a griffith flaw in the epoxy such that the net result with the aggregate is worse than without it! I think I just hit on another mechanism for a potential problem with large aggregate and one that would explain some of the abysmal results on the board with the small aggregate.

It would complicate the formula further, but it might be beneficial to wet out the aggregate entirely in reactive dilutant under vacuum before adding it to the epoxy. This would mean that there was reactive dilutant to keep air bubbles out of the aggregate. This is similar to the means that the folks at Nanoresins use to avoid end user mixing problems with their nano-particles. They produce them in such a way that they're already dispersed in epoxy.

Good observations, all.

--Cameron

P.S.

Walter, do you have any info on the weight of that granite devices spindle setup you wanted to use?

[greybeard]

Cameron - a couple of questions.

1. If a "glassy" type material were to be used as aggregate, like the flint chippings I mentioned a couple of pages back, would not this avoid at least one problem, the ingress of any air into the aggregate surface ?

2. The point about the 30 micron(?) layer of epoxy around each particle as the desirable thickness for maximizing strength.
If the individual particles are all irregular, in fact any shape other than cuboids, the layer must vary in thickness from possibly zero(if they are actually touching at any point) to some macro value depending on their irregularity.
Unless of course this layer is to be achieved by pre-coating the aggregate particles.
Also, a consideration of the effect of introducing particles smaller than this threshhold seems to me to not affect the issue.
I feel that this is one consideration in the long list that we are looking at that has little value other than as a theoretical notion.


Or have I missed something important with this particular topic ?

Regards
John

[ckelloug]

John,

I'm not as worried about air entering the pores in the aggregate as I am getting a layer of air over the surface of the aggregate that causes the aggregate to carry no load and just form the anchor point for a void. That being said, flint is one of the best aggregates from the quartz family according to this paper on the fracture toughness of quartz. http://www.minsocam.org/ammin/AM67/AM67_1065.pdf

To answer carefully however, I really don't know whether the glassy nature of the surface of the aggregate makes a difference or not.

Section from tried and true researchers on E/G:

As for your point about the 30-36 micrometer of epoxy, this is from experimental data, not theoretical data in Gupta's paper http://www.cnczone.com/forums/showpo...postcount=1092.
(Note that the actual paper by Gamski with this specific surface area times thickness estimate of epoxy volume cited in the paper linked in the above post is really not much more descriptive than Gupta's paper about the concept.)

The interpretation isn't what you'd think. It may still be bogus but this was from people who get paid to do this work over the course of the last 30 years. It definitely isn't me sitting around making stuff up The thickness idea goes back to the model cited from Gamski's 1975 paper in the London Proceedings.

Rather than assuming anything about the aggregate shapes, they use a material property called specific surface area (surface area per unit mass IIRC) and they multiply it by their epoxy thickness number and use that to calculate the volume of epoxy. This is nice for types of material where the specific surface area for a given sieve size of that material is known in advance. In essence, the thickness number is an average thickness of epoxy over all aggregate.

Using Gamski's model, no individual particle necessarily is coated by the average epoxy thickness. The model seems to have been shown empirically in papers over the years to let you correlate the "dryness or wetness" (volume ratio of epoxy to aggregate) of your mixture with how they will stack up against published strength numbers for similar mixtures.


Idea from looney systems engineer posting on the thread:

The conclusion that I drew, the part where the flaw may lie, is this: As specific surface area of items gets larger which generally correlates with particles getting smaller, increasing volumes of epoxy are theoretically required to have the same volume of aggregate to volume of epoxy ratio upon which most models for material strength are based.

I interpreted this to mean that particles above .1 micron which don't have a dispersion hardening effect need to be coated in 30um of epoxy just like rocks for the best epoxy bond which probably is some type of chemistry and bond length and polymer chain thing. Since the Specific Surface area of small particles is very high, it raises the epoxy fraction which is bad for the predicted strength according to the rule of mixtures calculation of expected strength. Since there is a point at which higher aggregate packing is actually a bad thing, I figured that the sanest interpretation was to leave out the particles which were smaller than 30 um but bigger than .1 um. There may be other or better interpretations.

Measurements of Specific surface Area:

I'm not sure how specific surface area is measured in practice for arbitrary rocks other than this paper I saw where a national lab was scanning the aggregates with a laser scanner and then computing the surface area of the polygon models. http://ciks.cbt.nist.gov/~garbocz/paper144/index.html

I have seen a chart of specific surface area for aggregate passing different sieves however in some U.S. Army Corps handbook for erosion abatement around dams although I cannot find it now.


In conclusion John,
Thanks for the criticism. I've tried to explain clearly here what I took from published sources and what I came up with on my own. Either may be wrong and if you have any more specific thoughts on the matter, I'd be interested to have a dialog if you post them.

P.S. The TR104 paper from www.niss.org that you posted some time back talks about this concept as well which is known as the "binder skin theory".

Regards,
Cameron

[walter]

re: Solid E/G base vs steel structure

Our discussion and focus on 'not trying to force' everyone into the same 'mould' had good efect and brought a lot of new people into E/G. People who wouldn't necessarily come together otherwise. You have to keep that in mind.

E/G seems to be pretty versatile, and that is, in my opinion, it's main benefit. (Finally something that can be tailored exclusively to your needs) You see a machine base or filled weldment? Well.. I see a kitchen sink. Or a garage floor. I see many different uses for E/G.
You've got to think outside the box (and outside the tube). Why limit the possibilities?
Especially when it comes to machine tool.. You have to allow full customization or no one will use it. It is a highly individual thing, and you have to decide for yourself what will be most effective for you. In another words- anything goes.

Tube steel structures aren't that easy to manufacture. E/G on the other hand, is simple, quick and everyone can do it. No tools necessary. (This works for me, as my current tool collection consist of one screwdriver and one extension cord...Well, two screwdrivers actually- I recently aquired the Phillips).

I know E/G lacks strength. But I'm sure it can be worked out.

re: filling 10"x10" steel tubes

I don't think you need E/G for filling weldments. Just use portland cement in rubber bags- it costs like a dollar per bag and works better than anything else. Pros have been using it for years.


re: my motors

These are ac servos meant to power each axis. Spindle could be anything from Porter Cable to small HF Elkstrom Carlson- depending on budget. Z axis total weight below 50lbs.

Thanks and keep up the good work!
_

[Mazaholic]

Thanks for the feedback.
The removal of air in the mixture has been going through my head for a few days,it seems to be one of the main problems that needs to be solved for strength and adhesion.
I've been trying to think of a way to deair thats not just a waste of vacuum and time,and also will not require alot of fabrication and money.
Some have mentioned using a thick mixture of epoxy,so i'm trying to think of a way to deair that would give acceptable results with different consistencies,because i think different mixtures and thicknesses will need to be tested to find the right equation,and i don't want to have to build something new for each test.
I guess you could say i needed to think out loud and get feedback.
Thats what i like about this thread.
Thanks

[lgalla]

I have asked this question before.On a cast beam,where do you get the precision surfaces to mount rails?The mold would have to be precision and costly.Anyone have thoughts or solutions?
Larry

[Geof]

I have asked this question before.On a cast beam,where do you get the precision surfaces to mount rails?The mold would have to be precision and costly.Anyone have thoughts or solutions?
Larry

If you go back many, many, many, many posts you will see this is a point I brought up a long time ago. Using polymer concrete does not remove the need for having the capability to do fancy machining. This is why I said that embedding something such as 8020 extrusions might be a viable approach. I also mentioned something about using a rigid tube to hold the extrusions straight during the embedment. But that was many posts back.

[lgalla]

http://www.cnczone.com/forums/showthread.php?t=30751

Epoxy aligned machine from Hayden.May be of interest.
Geof,I remember the 80/20 solution,Surprized no one has gone that route.
Larry

[Eson]

Igalla: You mount the rails on inserts. This requires the inserts to be straight but not leveled. You later on can adjust the height with a screw. When everything is in place you can make a second casting and make it all set. No precision surface needed.

[brunog]

I have asked this question before.On a cast beam,where do you get the precision surfaces to mount rails?The mold would have to be precision and costly.Anyone have thoughts or solutions?
Larry

Larry,

Embedded 1020 steel Flat Bar can do the work, with hex cap screws bolted from under as anchors can do the trick, this will also serve as reinforcement to the E/G cast.

Geof,
Isn't there any special treatment to be done before applying epoxy to aluminium? also the 8020 parts definitely need to be anchored to the E/G.

Best regards

Bruno

[Geof]

Geof,
Isn't there any special treatment to be done before applying epoxy to aluminium? also the 8020 parts definitely need to be anchored to the E/G.

Best regards

Bruno

I am not really up on what treatment can be done to improve adhesion between epoxy and aluminum but in this application I don't think adhesion is important, it is more a case of mechanical embedment. If the 8020 was buried flush to the surface leaving just the one slot for attaching things the mechanical trapping along the length would be adequate I think.

[brunog]

I don't think adhesion is important, it is more a case of mechanical embedment. If the 8020 was buried flush to the surface leaving just the one slot for attaching things the mechanical trapping along the length would be adequate I think.

Conceptually you must increase thickness of the E/G to compensate for the lack of adhesion.

I totally agree about attaching anchors on the T-slots embedded in E/G ,I just wanted to point out that anchoring is necessary.

Best regards

Bruno

[greybeard]

Hi Cameron,

John,
I'm not as worried about air entering the pores in the aggregate as I am getting a layer of air over the surface of the aggregate that causes the aggregate to carry no load and just form the anchor point for a void.

Good point, but I would expect epoxy to "wet" most surfaces pretty well.
Certainly that would be something to look out for in my first experiments with crushed flint.

It definitely isn't me sitting around making stuff up The thickness idea goes back to the model cited from Gamski's 1975 paper in the London Proceedings.
....... In essence, the thickness number is an average thickness of epoxy over all aggregate.

This is exactly what I meant when I asked " have I missed something important...".
I'm sure you will now understand my original concern when I say that the "average thickness" had passed me by.

.......increasing volumes of epoxy are theoretically required to have the same volume of aggregate to volume of epoxy ratio upon which most models for material strength are based.

I interpreted this to mean that particles above .1 micron which don't have a dispersion hardening effect need to be coated in 30um of epoxy ................ Since the Specific Surface area of small particles is very high, it raises the epoxy fraction which is bad for the predicted strength according to the rule of mixtures calculation of expected strength. Since there is a point at which higher aggregate packing is actually a bad thing, I figured that the sanest interpretation was to leave out the particles which were smaller than 30 um but bigger than .1 um. There may be other or better interpretations.

While I get a bit lost here, I'd be happy to ignore the dust anyway.

Regards
John