[greybeard]

Thanks Cameron. I've just started sieving my various buckets of sand etc so I should be able to see what the dry mix is like soon, then try out the new set up to produce some more test samples. Slight delay next week visiting the medics, but back soon.
Take care,
John

PS How many samples of each run do you need - was it 5 or 6 ?
If it's 5, I can keep one as a reference, or send bits off for microphotography.

[ckelloug]

The sample regimen for ASTM D-790 is 5 samples. I just finished welding the initial steel rectangle to make the perimeter of the table to convert the tile saw into a sample saw for your last set of samples. . .

Regards all,

Cameron

[greybeard]

Well, I'm back from sitting on the radioactive hegehog, and can now drive the nearest airport security crazy. The downside is that SWMBO is keeping a very careful eye on me, and I can't sneak out to the workshop for a while.
However, time to think and plan. HEheheheheheheaaaaaaaaaaaaargh.

EDIT
Just been to the BBC news site for an update on the Phoenix lander and the soil problems.
The first test of the sprinkling method on Monday produced a layer of fine particles.
Quote <
"This is good news," said Professor Ray Arvidson of Washington University in St Louis, lead scientist for the robotic arm.

Professor Arvidson said the clumping tendency of Martian soil at the Phoenix site and some earlier landing sites might come from extremely fine particles filling in the gaps between coarser, sand-size particles, perhaps acting together with an unidentified ingredient to cement particles together.
> Unquote

Wonder if it's epoxy ?

[greybeard]

OK folks, it's mixing time again.
I shall now try the P7 mix as per recent posting.
So do I -
1..mix all the grades together, then add to the resin mix
2..add the resin mix to the aggregate mix
3..add the finest to the resin mix, then go up through to the coarsest
4..add the coarsest, then go down through to the finest ?

EDIT
I've had a couple of hours to think about it, and it seems to me that the sensible thing to do is to mix the finest particle in first, then work up through each size.
The logic is that the smallest particle, which is also present as the smallest quantity, will have the least effect on the viscosity of the resin. Likewise, each fraction in increasing size and increasing quantity.
While this might increase the amount of air trapped between the particles, more than mixing in the reverse order, spinning the mold should remove all the air anyway.


Anyone like to state a reasoned opinion, or their own experience ?

John

[probinson]

I'd say your best to mix the fine particles first, not necessarily one size at a time but, all the sizes that can be easily stirred. Saving the largest ingredients for last should make thorough mixing easier.

Pete

[ckelloug]

I'd agree with pete that mixing the small particles first is good although as the thread theoretician, I have made epoxy samples but no rocks and epoxy samples yet.

I've been out in the shop working on a special saw table for accurately cutting samples on a tile saw. I'm still picking swarf out of my shoes.

Best of luck with the medics, John. I should be back on the road to accomplishing something soon.

--Cameron

[greybeard]

Thanks Pete and Cameron.
Apart from the pair of "plums" I'm sitting on I'm recovering OK.

[greybeard]

Well, that was fairly useless. :tired:
Though I suppose even failure can generate knowledge.

I tried the P7 mix with a reduced amount of resin - 300gms P7 aggregate + 60gms resin.
Gave it a much shorter spin 5 mins@1500rpm, followed by 3hours@300rpm.
This morning, the samples were given 2hours @ 70 C in the oven.
After cooling to room temperature, I could break them in my bare hands without much effort.

1. During the grading out of the seven sizes, it seemed to me that the ratios were very light on the fine particles, so much so that a dry mix was very difficult to get into any sort of uniform dispersal. The slightest vibration caused the finer particles to fall to the bottom very easily.
After re-grading, I combined the three finest, the next pair, and the largest pair to give me three parts to add to the resin, starting with the finest group.
Mixing was easy for the smaller five sizes, but adding the largest pair turned the 'paste' into a very open, loose mix.

2. After demolding, it was obvious the a great number of voids existed throughout the samples. The reduced spinning time at high speed was insufficient when combined with the lower resin concentration to get the resin to flow down through the mass, to fill the lower voids. The larger particles remaining on what is the inner surface of the spinning samples were still 'wetted' with resin, and well glued together.
However, I would guess that the resin ratio should go up to 20%, rather than the 16% in this test.

3. Perhaps most crucial of all, the aggregate was breaking. A close inspection of the break shows quite clearly that the larger particles are cleaving. This leads me to believe that my 'Ballast Mix' contains a lot of weak material - sandstones etc and makes it a bad choice of aggregate source.

Fortunately, I've got a bag of quartz sand to try next, so combining that with silver sand will be the next test.

[Peldor]

How about other materials for example I just found in my local distributos homepage sutch SicaDur 42 (Rapid High Strength Pourable Epoxy Grout) www.sika.com
in datasheet USES section was:
Grouting of bearing plates, machine bases,
mechanical bridge joints, bridge bearings and rail
base plates (I like this one becose I have THK reils which need base ).

Flexural strength: 30 N/mm2 @ 20°C
Tensile strength: 20 N/mm2 @ 20°C
Modulus of 18 KN/mm2 @ 20°C
Elasticity: (static)
Density: 2.0 kg/litre

There was nothing mentioned about wibration damping characteristics .

This is wery intresting material SikaRail® KC330/340 Polyurethane Flexible Resin Based Grout

USES
SikaRail KC330/340 provides a flowing, vibration absorbing load bearing resilient grout for:
* Bedding and bonding rail base plates to:
steel bridge decks, concrete bridge and tunnel slab
tracks, concrete sleepers

ADVANTAGES
* Out performs cementitious, epoxy and polyester
grouts under dynamic loading.
* Excellent noise and vibration damping.

maby this is the right material to tray out ?
and in datasheet is also sayed that Additional detailed information such as load/deflection graphs, spring coefficients, resonance frequency, damping factor etc are available on request.

I start to think Do i realy need this 10x vibration damping (over cast Iron)?
I think that the main benefit of these special compozit concretes are that i can make any solid shape myself for mutch lower price

I also found a self leveling epoxy that might work for this. We use it as engine bedding one crew boats and tugs...

http://www.philadelphiaresins.com/
look under "shipbuilding and repair coumpounds" its under the Chockfast Grey link.
By the way, Philadelphia resins also had a "Castable for High Precision Applications" the stuff is called "Zanite polymer compound". It looks like some really tough stuff. http://www.itwpolymercastings.com/index.html

Here is a promo video they have on YouTube: [ame="http://www.youtube.com/watch?v=_XXATSCAaCg"]YouTube - ITW Polymer Tech - Casting Division - Capabilities Video[/ame]

[brunog]

John,
your sample does looks as there should be more smaller particules.

It looks more like epoxy peanut brittle.

Best regards
Bruno

[greybeard]

Hi Bruno.
Yes, I agree with your thought about the lack of small particles, and your "peanut brittle" is exactly the term I was looking for but eluded me.
I've now got a mix going of equal parts of the same grades as in the P7 mix, but with the same resin content (16.6% w/w), and it stirred up completely differently, with no signs of any voids. I've also gone back to a long higher speed spin at 1425rpm, which for this mold equates to 75G.
I'll add a photo later this pm, when its out of the oven.
Regards
John
EDIT
Not even worth posting a photo - another weak sample. Not sure why, though I did lose some resin via a badly sealed end of the molds.
Still a lot of voids present, so I might be forced to go back to a higher rev motor.
I'll try going back to an earlier mix, just using silver sand, flint, and resin. With an increase in the % resin, to see if that cures the void problem.

[ckelloug]

John,

How have you been measuring your components? Do you know the intrinsic density of each of your components? The formula's from de larrard's book are based on the mass of component divided by the intrinsic density of the solid portion of that component. (You were the one that pointed this problem out to me an eon or so ago.)

I'd have to agree with everyone else that the number of small particles looks low. With the grading curve you said you were following, the theory says that most of the smaller particles are supposed to get trapped between the larger particles. It was also stated in Kinloch's book that larger particles are definitely weaker than smaller ones.

We also have the problem that despite what the theory curves say about resin that there is some amount of resin below which the sample will be unprocessable via a given method.

Based on my understanding of the theory, your comment about adding the largest one making the stuff into an open loose mix demonstrates the effect that de larrard called the loosening effect in the theory where larger particles force the packing of the smaller particles apart. I suppose that I'm going to have to get around to taking beta measurements for the packing coefficients and also fix the simulator since it choked on the mixture you're working with for reasons I cannot explain.

Peldor,

We've seen the ITW Philadelphia resins page. We're trying to understand the basic technology to the point of being able to design the best use of the materials that each of us have on hand. John and I probably aren't on the fastest path to getting something built but we will understand what we've got when we get it.

To everybody else,

I'm still sidetracked on making a precision table for a tile saw to cut John's samples since it's a bit trickier to clean up samples that are almost to size than it is to cut them from a large block.

Regards all,

Cameron

[greybeard]

Hi Cameron,

John,

How have you been measuring your components? Do you know the intrinsic density of each of your components? The formula's from de larrard's book are based on the mass of component divided by the intrinsic density of the solid portion of that component. (You were the one that pointed this problem out to me an eon or so ago.)

By weight.
I've measured the intrinsic densities of the various components using a volumetric flask, weighing a sample of the sand, filling to the mark with water, and reweighing. They all came to 2.6 - 2.7gm/ml, so I'm assuming weight will give me the correct proportion.

I'm still sidetracked on making a precision table for a tile saw to cut John's samples since it's a bit trickier to clean up samples that are almost to size than it is to cut them from a large block.

Cameron

It would be a lot easier for me, I now realise, to cast a 64mm diam tube, 15mm or so thick, than these small sections.
Could you cut slabs from that ?
Or better still, I could cast a few separators down the tube so that you would have say 90 deg segments. It would save on epoxy too
Regards
John

[greybeard]

Oops.... destroyed tailstock bearing by overtightening.(My usual standard of construction, I'm afraid)
Just rescued the current test - 70gm silver sand, 35gm flint, and 27gm epoxy.
5 minutes at 1450rpm removed the air and settled the sand, but left the flint dispersed through out. After post curing at 70C, it took a 2lb hammer to break the sample(15mm x 9mm), so I'll repeat the exercise as soon as I've repaired the m/c.
John

[greybeard]

Cameron, I still have a small doubt in my mind about the validity of testing one of the samples I have made using my spinning method on the same basis as a sample which has been 'open cast'.
I did raise the thought some time ago, but it slipped by, without any resolution.
The source of the problem I have is that my samples will not be uniform in properties as measured radially with reference to the spin axis. Lets call that my Z axis.
I would predict that you will get a different deflection under load depending on which way up it is. In one orientation, the resin rich surface is under compression, and the resin poor surface is under tenson, and then the other way up, vice versa.
A casting made with a surface that is resin rich all round, and resin poor in the neutral axis will surely give a much better performance than one in which the same cross section has variable resin concentration.
Thoughts ?
Regards
John

[ckelloug]

John,

Excellent point. A sample that is resin poor along the neutral axis will perform fine as long as there is enough resin there to carry the shear loads which are much smaller than the tensile or compressive loads in a flexure test such as the one I will soon do on your samples. Within a sample that has only a fixed amount of resin, the neutral axis is the least important place for the resin to be.

More resin along the bottom of the sample will tend to make the samples stronger as the failure mode of E/G is normally in tension in the literature. E/G is much stronger in compression than tension and I would surmise that to a point the epoxy ratio is less critical for parts of a sample that are in compression.

Since I haven't tested your samples yet and you gave me an extra one, I can flip that one and test it for a rough test of the importance of the difference at least on a sample of this size.

All this being said, I'll test your samples resin side down since that's going to have the highest tensile strength.

I'm nearly done with the table for cutting your samples. I just need to add clamping and adjusting screws to the apparatus.

Regards all,

Cameron

[greybeard]

Thanks Cameron.
Could you let me know what the maximum dimensions of cross section your set up could test are ?
For example could it hold a 3/4" rod, or square section ?
I have in mind to try casting a small set of test pieces as solid as I can make. Unfortunately, the smaller I go, the faster I need to spin it, and to get a 3/4" diam cast, I need to spin it at 3000rpm to get 100G, or 4250 to get my preferred 200G.
Now that SWMBO has cleared my to get back into the workshop, I might try over the weekend.

John

EDIT If holding a bigger sample is not possible, could you post a simplified diagram of the testing set up, so that I might lash together something that could give us ball park figures ?

[greybeard]

If I spin a 3/4" diameter plastic pipe as above filled with epoxy resin at about 3000rpm, the force on the pipe wall will be equivalent to pressurizing the pipe to some fairly high figure.
Realizing that this might exceed the pressure rating on the pipe, I can see a possible nasty ahead.

I'd be very grateful if some kind soul, wishing me not to be covered in epoxy/sand, could run me through the maths.
I'd particularly like a general formula so that I can substitute alternative figures.
Thanks
John

[ckelloug]

John,

I managed to get revision A of the tile saw to sample saw conversion running today.

It's up to your standards of construction. No bearing failures yet however the brilliant design failed to account for the clamping screws interfering with the mitre guage. I threw a couple of cobalt lathe bits behind the mitre guage since they're parallel and small and I was able to cut samples but the hope is that revision B will cut straighter and with less of a major hassle. I need to modify the clamping assembly so that the the clamping screws bear on an intermediate piece rather than the sample itself as tightening the screws causes the samples to move unless extraordinary care is taken.Also, some of the epoxy paint flaked off of the assembly. I suppose I knew in my heart I should have baked it before use but live and learn. I must also say that I don't feel that either the spindle of the tile saw or the linear bearing from the old table which I adopted are rigid enough but with a better clamping assembly, they may do.

I didn't destroy any samples and will be able to test but I am slightly disappointed with the result of the cutting.

Unfortunately, the biggest sample I can test is 3/8 x 1/2 x 7.5 due to the text fixture available. I can however cut samples of 3/4 x3/4 or less to the size I can test. I can also probably figure out a way to cut some of the strange tube shape samples but I haven't worked out how yet.

Regards all,

Cameron

[greybeard]

Could you use 11mm diam rod as a sample ?
I've just built another spinner to make them.
John