[greybeard]

So the preliminary experiment with the de L type mix proved just how stiff the pre-cured mix is, and there was no way that sort of recipe would flow into the space I want it to, only by having an epoxy ratio of about 30%.:tired:

This would leave the upper sections of this part too 'ify', so on to plan B - cast plaster of paris into the mold, making a male cast, then two female casts from that.
Mount the two in a cylinder, and spin cast the parts as per the plan for all other parts of the machine base.

Checked out the proper recipe for the P of P, mixed it up, and poured.
OK, I thought it was too stiff. It went off half way through the pour.(chair)

So onto plan C.
I shall build an actual part from scratch, probably from wood and 'car-body filler', and carve it to fit the router body. Take two casts from that using the spin casting pipe as the container.
Sounds a complex procedure, but bear in mind that my mold pipe, spinning at about 1500rpm, is subjecting the mold material to about 80G for 5 minutes, before it slows down for the normal curing period of the resin.

Onwards and upwards.
John

[jhudler]

John,
If you vibrate this stuff it flows like warmed honey.

Jack

[greybeard]

So onto plan C.
I shall build an actual part from scratch, probably from wood and 'car-body filler', and carve it to fit the router body. Take two casts from that using the spin casting pipe as the container.

Edit .... New Plan C

As the second half of the original mold assembly around the router body is still intact, I shall make use of that to produce the first cast.
I've also discovered an unopened container of liquid latex rubber that I can use as a release agent, so I shall next attempt to use some of my expensive epoxy, partially filled with fillite, as a casting material.
But this time I will do a small test with a plastic box.......

[greybeard]

Well, some interruptions to the timetable, and several failures under the belt, but it has given me time to think through the whole process.

What I want to do is design a reproducible method of molding, as well as make the machine itself.
To that end I have decided to stick with my original concept of spin casting each component, rather than use vibration casting.
This would give me more experience in the method, starting as I will with the smallest component.

Each piece might be thought of as having 'critical' surfaces and 'non-critical' surfaces, with dimensional accuracy corresponding.
The critical surfaces of the z-truck are the outer flat faces, and I still plan to incorporate glass as the actual surface. Rather than following my original plan to cast them directly onto the surface of the router, as mentioned above, I shall spin them as discrete components and bed them onto the router once they have been lined up with each other.

The first step is to produce the core of the mold, the 'non-critical surface, then the outer box which will both form the 'critical' surface, hold the core in place, and carry the pivots for mounting the whole assembly in the lathe.

To be continued.....

[greybeard]

After numerous distractions I now have the speed controlled lathe in place in the workshop, and the first two molds ready for spinning.
In the attached photo, the ~3" x 9"long pipe on the left contains 6 plastic extusions which will cast test samples to be sent to ckelloug for evaluation(see epoxy/granite thread for details).
The square metal tube on the right, 3"x3"x10" with a core held down the middle, will cast two L-shaped sections to form the first casting for the Z-truck.

[greybeard]

Aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaarrrrrrrrrr rrrrrrrrrrrrrrrrrrrrrrrrrghhh.

Well, having got that off my chest, herewith update.

Trying to catch up some lost time, I went straight to my recipe of a seven part mix of four graded quartz sands, G800, and G200 zeeospheres, plus cab-o-sil, a sub-micron silica fume.
I reasoned that if the guys casting a thick mix were aiming at 10% resin, then I should have a pourable mix with 20%.
Unfortunately, I had overlooked the thixotropic effect of the cab-o-sil.
An epoxy/crunchy peanut butter mix can't be made pourable by adding more epoxy, and even raising the epoxy content to 50% still left a non-pourable paste.

Back to doing some simpler mix tests, and redesigning the test mold so that I can fill from the front, not the end.

[greybeard]

Great - we move forward. :wee: I managed to spin a set of six test pieces for Cameron's machine today.

My mix is
40gm Agsco quartz #4
40gm ditto #2
40gm ditto #1/2
40gm ditto #2/0
35 gm G800 zeeospheres
35 gm G200 zeeospheres

plus 63gm(~21%w/w) DX020 epoxy from Atlas Polymers.
(I think this has now been superseded, but as this is for comparison testing, it's ok for my needs)

The mix was just about pourable at 20 deg C, and the molds were spun at 2500rpm for 3 hours.
OK, so I forgot it - (chair)I had to cut the grass, then the hedge, and then it was lunch time - you know how it is.

When I can find my calculator, I'll edit in the G forces on the mold. Should be somewhere about 90 G I think.
EDIT It turns out to be ~225G !

Getting the samples out of the molds was a bit of a pain. (Notice the chisel) They released ok, but I managed to break one of the molds, so I need to rethink the mold design again. I want to do quite a run of different aggregates, so the mold/demold needs to be a lot easier.

The next stage is to give them a post mold heat cycle to make sure they're fully cured, then into the post.

[markus_detroy]

please explain... or send a link or pic describing...
zeeospheres and where you get it from?

My mix is
40gm Agsco quartz #4
40gm ditto #2
40gm ditto #1/2
40gm ditto #2/0
35 gm G800 zeeospheres
35 gm G200 zeeospheres

[greybeard]

Hi Marcus.
Sorry I wasn't so clear. I tend to assume that readers will have been following the main Epoxy/granite thread, where all the details are somewhere buried in about 3000 posts !
There is also an index thread that Walter and Cameron try to keep up to date, which has links to all the suppliers, as well as a condensed version of main topics.

Agsco Corp. is a US producer of finely graded crushed quartz in the US, and Walter was kind enough to send me samples last year, which I have only now got round to trying out in the system I'm using.
Zeeospheres, microspheres of silica/alumina ceramic are a 3M's product and I can get them in the UK through a specialist chemical supplier, Lawrence Industries.

Hope this helps,
John

[greybeard]

Attached is a polished end of a test sample. The dark band at the top is the excess resin, as predicted, but the next grey band is cuase for concern.
I suspect it may be the zeeosphes which are slightly less dense than the quartz 2.2gm/cc to 2.65gm/cc. This might indicate segregation at this high G force, so I've either got to reduce the spin rate, or change the two zeeospheres components to ultra fine quartz sands - about 40 micron and 10 micron.

[greybeard]

oops - here is photo

[greybeard]

I decided to make a simple mix as a base line to judge any improvements in the strength that I could make by varying the recipe.
1 part block paving sand, 1 part silver sand, and 1 part flint powder, all mixed with 20% resin as before.
The flint powder is courtesy of SWMBO, retired pottery teacher. I measured its density to be 2.65, so it's the same as the sand, and I hoped it wouldn't stratify in the spinning. However, the photo shows it still seems to happen.

Since doing this, I've acquired a variety of sieves, and can now analyze the typical aggregate mixes available locally.
My first discovery was that the two sands were within a gnat's whisker of being identical. The samples seemed to be pretty solid,so a set will go off to Cameron soon for testing. Unfortunately my micro photography doesn't show up enough detail to see if the flint layer at the top of the section also goes right down into the general mix. Hopefully that will be resolved soon.

I've also gone ahead and rebuilt the spinning rig so that it is now big enough to make all my designed components for the planned cnc machine. Just the adjustable tail stock to be put in place of the present fixed one.

Using 7 of the sieves, I can now extract from a bag of "Ballast Mix"(sometimes referred to as "All-in") from the local builders' merchants, 25kgs for £1.34, all the sizes for my next trial recipe - formula as per #3199 post on the E/G thread. This will be my next test sample, testing both the recipe and the rig at the same time.

Unfortunately, having sat on a radioactive hedgehog(think porcupine if in the US) over the last few days, there will be a slight delay before SWMBO will let me back into the workshop.
Ho hum.

[greybeard]

Briefly, the samples run as described in #3209 with ~16% resin w/w produced peanut brittle(thanks Bruno), with very little strength. It also showed up weak aggregate material, which I'm guessing is sandstone, but may be crushed recycled brick. There's no spec with this cheap aggregate, so there could be anything in it.

Having now separated the quartz sand, from WBB locally, I've done a run with the same 7 sizes and resin %, but using equal parts of each grade.
The mix was much easier to stir up, and seemed more compact as I pushed it into the molds.
However, after post curing it for 3 hours, the result is poor.
I hadn't sealed one end of the test molds properly, and a small amount of resin now forms a fine line up and across the workshop. This means I have no idea on the final resin content, but I don't think the loss was enough to explain the large number of voids present. Once again, I can break the sample without too much effort.

So what now ?
The strongest samples so far have been those made with only silver sand/paving sand and flint, with an excess of resin.
I think next will have to be just silver sand with less flint and less resin.
I must somehow get the simplest formula that works in the spinning set-up that is now running well, if somewhat simpler than planned.
It's a 1/6 hp motor directly driving the molds at 1425rpm - no speed control, no variation.
Gives me 75G.
If the next one shows voids then I'll have to change the motor speed and get back to ~2500rpm, giving me 200G.

[greybeard]

OK. So 70gm silver sand plus 35gm flint, plus 27 gm epoxy. Spun at 1425rpm for 5 minutes when the tailstock bearing collapsed
Undaunted, I extracted the molds from the wreckage, and got them into the oven at ~70C for two hours.
Took a two pound hammer to break them, so photo soon.
The brief spinning pulled the air out, and the sand particles to the bottom.
The flint still seems dispersed completely throughout the resin so I guess any residual micro-bubbles attached to it are slowing any sedimentation down.

I'll try the same mix later, after repairing the m/c, for a full three hours, and see if there is any noticeable difference.

I have a very slow gel time I think because the resin is past its shelf life date by about 10 months.

[Gir]

Let me start this off by saying I really have no clue when it comes to the mixes that you're making. That being said, won't it always separate the mixtures unless it's slightly hardened? Or is that what you're going for...

[greybeard]

Hi Gir.
If you spin a mixture of particles of different sizes but all with the same density in a fairly viscous resin, the particles don't behave in the same way as you might expect them to if you were just shaking them together.
In the latter case, curiously, the larger particles come to the top, and the smaller ones sink. The small particle can always find a hole to drop through between the bigger ones, but the opposite isn't true. It's actually an example of probability working in real life.
However, in my case things are a little different.
First, the presence of the viscous resin slows down any movement, and reduces any effect, but more importantly, the choice of particle size in the mixture tends to reduce stratification to quite a marked degree.
For a full background to the theory, you will have to either read all 3000+ postings on the epoxy/granite thread, or the work of a French engineer in the 1990's, or simply take my word for it.
At its simplest, a particle of just the right size will get trapped in the space within a group of larger particles, and a still smaller ones will get stuck in the remaining spaces.
Get the sizes and proportions right, and the tendency to separate when you shake the mixture is reduced to a minimum. Fill the last vestiges of space with an adhesive to stop any movement, and you have a very stiff and strong material.
That's the theory that we are trying to put into practice, to develop a diy recipe for anyone else to use, which will offer an alternative route to building a frame for a cnc machine.
My own variation on this objective is to use centrifugal casting instead of open casts plus vibration etc.
Regards
John

[Gir]

That makes a lot of sense and explains a lot, but it sounds like by using a centrifugal method you will be taking out an important key to the equation - randomness. Instead of shaking and moving everything around and having all the particles bump around, you're simply sending the densest particles to the outside. 75 G's is a whole heck of a lot of force, so I would imagine you'll almost always get separation. A slow turn with vibration makes more sense in my head. I don't really have the time (or patience) to read that big thread, so it's very likely I'm still missing something about using the centrifugal method. Feel free to ignore my comment

[greybeard]

What you're overlooking is the history of the particles before the spinning takes place. If the aggregate particles have a size distribution that meets certain criteria, then they don't separate, but will be spun outwards to form a well packed layer. They exhibit this "locking together" even in the dry state.
During the original mixing by hand, care has to be taken in ensuring that a good mix occurs, random but uniform, if that makes sense. The finest particles are introduced to the resin first, then coarser ones, until each grade has been added, and stirred throughout, hopefully with the minimum of air trapped within.
When spinning starts, it would seem that what happens is to a great extent dependent on the relative proportions of resin to aggregate.

If there is an excess of small particles, in the presence of excess resin, the solid material can be spun out to form a packed layer, but the excess of finer particles get left behind, suspended in the resin.
Unless of course prolonged spinning occurs, when that excess will form a layer on top of the main body of particles, and leave clear resin at the top.
The excess resin is there to make it easier to get the mix into the narrow molds, but is also useful in allowing any trapped air to be replaced by resin.

These quantities are best determined by experiment, but a great deal of time and effort can be saved if the experiments are based on an understanding of what happens at the small scale of the particles.

I'm trying to gain an insight into what goes on during the spinning in order to push my recipe towards what I hope will be a very strong, reproducible material, but I'm very happy to receive anyones ideas to that end.
Explaining my own ideas is a very good discipline, and exercises the little grey cells wonderfully.
John

[Gir]

What you're overlooking is the history of the particles before the spinning takes place. If the aggregate particles have a size distribution that meets certain criteria, then they don't separate, but will be spun outwards to form a well packed layer. They exhibit this "locking together" even in the dry state.
During the original mixing by hand, care has to be taken in ensuring that a good mix occurs, random but uniform, if that makes sense. The finest particles are introduced to the resin first, then coarser ones, until each grade has been added, and stirred throughout, hopefully with the minimum of air trapped within.
When spinning starts, it would seem that what happens is to a great extent dependent on the relative proportions of resin to aggregate.

Oh, stupid me! That makes much more sense now! I'm fully behind your theory now

Explaining my own ideas is a very good discipline, and exercises the little grey cells wonderfully.

Heh, I know what you mean. I always have to explain Calculus to one of my mathematically impaired friends at college. Great brain exercise and an occasional good test of my patience!

[greybeard]

Hi Gir.
Over the years I've never found a more powerful analytical tool than logic.
But being able to explain it to someone else is the great test.
Regards
John