[peteeng]

Hi Paul - I admire your optimism and hope. But the meniscus is always there no matter how you look at it and fill it. Its a physical fact of surface tension, surface wetting and surface/air interaction etc. All threads that are honest about the epoxy levelling experience have poor results.

Even if you cast the perfect surface you still have a soft foundation for your mill parts and that's the major concern. Model a 6mm steel plate, 6mm epoxy and the rail with a counterbore hole. Push down on that hole with the required bolt preload and see how far the epoxy compresses.... Not pretty... Peter

[PaulAmelang]

Peter,

Just to clarify, I didn't use the Remove Rigid Body Modes feature in the last simulation. The description in F360 documentation is: "At times, it is difficult to constrain a model without limiting or preventing the deformation that would occur naturally. If you choose to leave your model unconstrained, you can select Remove rigid body modes to eliminate rigid-body, or free-body, motion during the solution." I could be misunderstanding it, but that sounds like the situation I'm in. I wouldn't use it for general purposes, because I don't know how to make the sum of forces and moments zero in a complex simulation. But I'm using it right now to model my table by itself with an arbitrary force applied right to the middle of the table, and gravity is turned on. If I am constraining the feet, ideally I would just constrain one of the edges or vertices to slide on the floor, right? Since it is unrealistic for the whole flat bottom of the table legs to remain attached to the floor.

About using PU: you've scared me away from using PU. If I were going to use it, I would probably want to let it cure in a rough mold and then cut it down to the exact shape that I want. The other thing is that I think that it will be more economical to achieve the properties I need for the core by using steel ribs. Getting PS foam wedges to couple will be harder with steel ribs, but I think if I cut the foam and steel accurately, I will be able to use epoxy and fumed silica to fill in any gaps and mate all the surfaces well enough. I have some polyester and vinyl ester resin that is going to go bad and I've been looking for something to use it on, this could maybe be a way for me to use it up. Or I could just do the job properly with some epoxy.

I appreciate your experience and wisdom. Switching to a concrete design would put me back weeks if not a month, which is not really feasible for my business right now. I feel good about the potential of this design: it seems like it will be a great deal for the work area and rigidity. There are just a few things that need to get worked out.

[PaulAmelang]

Hi Paul - I admire your optimism and hope. But the meniscus is always there no matter how you look at it and fill it. Its a physical fact of surface tension, surface wetting and surface/air interaction etc. All threads that are honest about the epoxy levelling experience have poor results.

Even if you cast the perfect surface you still have a soft foundation for your mill parts and that's the major concern. Model a 6mm steel plate, 6mm epoxy and the rail with a counterbore hole. Push down on that hole with the required bolt preload and see how far the epoxy compresses.... Not pretty... Peter

Then I will use a torque wrench to tension all the bolts the same amount. Otherwise, the simulations seem to show that a layer of epoxy on the table, y-beams, and gantry are all relatively okay. For what's worth, Precision Epoxy says that "3/16 in. is the minimum pour for large surfaces to get the required leveling characteristics" (I'm paraphrasing). So I might be able to get away with 3/16" or even 1/8" for the Y-beams and gantry.

p.s. Yeah I know the meniscus will still be there. I will have to measure how large it is and build within its perimeter.

[Waynekofuco]

Hi Paul - I admire your optimism and hope. But the meniscus is always there no matter how you look at it and fill it. Its a physical fact of surface tension, surface wetting and surface/air interaction etc. All threads that are honest about the epoxy levelling experience have poor results.

Even if you cast the perfect surface you still have a soft foundation for your mill parts and that's the major concern. Model a 6mm steel plate, 6mm epoxy and the rail with a counterbore hole. Push down on that hole with the required bolt preload and see how far the epoxy compresses.... Not pretty... Peter

And I'll add one last niggle with epoxy wo filler. My exp with epoxy is that post cure hardness/stiffness is inverse to temp with a narrow window of stability. In ambient temps < 70F epoxy is glass like brittle, >90F is very pliable. This is months after cure and unloaded, just blanks I made for testing...

I think you're just gonna have to perform personal tests with the intended product to see if you want to fully invest in that solution.

[peteeng]

Hi Paul - You have introduced a new variable in the equation by requiring a machine for a business. You think you maybe a month behind but you are actually years behind. If this machine is required to make a profit then it probably won't. If you look at the machines in the forum and from my own experience of designing machines professionally for over say 30 years you still have 2 years ahead of you to develop the machine to a production quality level. This sort of machine will let you down at a crucial moment when you have committed to supply stuff and that will put you out of business.

You have a couple of ways to go: 1) contract the parts to a jobbing shop, do you have a few quotes for the parts you want to make? - you may think that this is expensive but they have better machines, better people and a good price. If you can't make money by contracting you probably can't make money by doing it yourself. But your an optimist and live in hope. 2) Do a business plan and lease a production quality machine if the numbers stack up. 3) There are some rules to making money in a manufacturing business. One of them is do not design and build a machine if it exists in the market. Do not reinvent the wheel it is lost time, effort and $$$. You only design a machine if the market does not have the animal you need. 4) Decide what your longitudinal core business is. If your retailing a product then start contracting the parts, put your efforts into the product not the process of manufacturing it... There is a difference and there are shops out there that can make anything. You can't be a machinist, designer, sales man, office manager, machine repairer, machine designer etc all at once.
Good Luck Peter

I went to a machine trade show a couple of days ago and a 3 axis medium size HAAS was making engine pistons in a 12min cycle.
They gave me one off the stack,,,,
Your machine can never get to that sort of thing and there are 15000 new HAAS machines out there every year to take advantage of...find a good shop and your dream will happen faster...

[PaulAmelang]

If I needed to fill the gantry with PU, I guess how I would do it is mix and pour one cup at a time. Would be a little wasteful, but I could use a spatula and disposable cups to minimize waste. It supposedly takes about 5 minutes for the initial cure. The problem with filling the gantry with 16lb/ft3 PU is that the PU would weigh ~100lb and cost ~$400. That's about half the weight and twice the cost of the steel skin. It would probably be a better use of money to go with an economy core and thicken the skin if necessary. But I'm not sure.

[PaulAmelang]

And I'll add one last niggle with epoxy wo filler. My exp with epoxy is that post cure hardness/stiffness is inverse to temp with a narrow window of stability. In ambient temps < 70F epoxy is glass like brittle, >90F is very pliable. This is months after cure and unloaded, just blanks I made for testing...

I think you're just gonna have to perform personal tests with the intended product to see if you want to fully invest in that solution.

Thanks, it will be curing in a basement and I'll have control over temperature and humidity. After it cures initially, it probably wouldn't be a bad idea to throw some heat lamps on it to help it cure fully. I believe epoxy doesn't ever cure completely unless it hits a certain temp. That may be why the casting epoxy is spec'ed lower as far as strength and rigidity goes; it is a low temperature, long cure time epoxy.

[PaulAmelang]

Peter,

If I sub every single thing out, I will lose money every time I sell a product. Spending a bunch of money on expensive machinery, wether bought or leased, is not the way to go about starting a bootstrap startup. The stuff I need to do is low tolerance. You are trying VERY hard to talk me out of this!! If I had the money, I'd probably buy an Avid. That would probably suffice for what I'm doing.

You can't be a machinist, designer, sales man, office manager, machine repairer, machine designer etc all at once.

Hahaha. Yes I can.

[peteeng]

Ah Paul - Have you actually got a quote for anything? Do you really understand your potential cost structure and production stream and limits, I think not! If an Avid can do the job then buy a second hand router, you will get to the finish line faster then building a quarter baked mill. I do wish you luck but that's what your going to need lots of. Every step of the way you are going against so much prior knowledge and experience, technically and now commercially... There is always a way to get a bootstrap going if the proposition is good. I've started a few in my time. I'll check back in a few months to see if your floating or not.

And very sorry you can't fulfil those roles "all at once" that's the issue you can do all of those tasks but it takes 7 weeks to get a weeks work done... so your project just blew out from 1 year to 7 years. Automotive, aerospace, mining, apple, every modern manufacturer subs out work its the only way to leverage time and $$$. Peter

Peter

[PaulAmelang]

Peter,

I appreciate your technical help, but your business advice is unsolicited and off topic.

[pippin88]

I've used epoxy. Didn't work out for me. Wouldn't do it again.

I've posted about in detail before.

Epoxy is too soft as a linear rail bed. And the self levelling doesn't work for the sizes / areas we are talking about.

[PaulAmelang]

Hey pippin,

Any chance you could link to threads for me to check out? I searched for “pippin88 epoxy” and it didn’t bring anything up.

Are you saying that this application is too small for self leveling epoxy to work?

[PaulAmelang]

Okay,

I can see that the low-density XPS isn't doing a good job of transferring the strain. It is better than air, but it is not fitting the bill for what I need it to do. The best option is going to be PU, because of the amount saved on shipping (a few bottles shipped via UPS vs boxes of foam shipped via freight). PU would be a bit of a hassle, but no more difficult than cutting up XPS and getting it to couple well.

My new plan for the table is to use expanding PU and a steel or aluminum honeycomb core, which I can fabricate with my brake. Here is how I would do it:

1) Put what will be the top skin on the ground
2) Glue in the honeycomb and side skins
3) Pour in PU in small batches until the foam is slightly higher then I need it to be
4) After a few hours, cut off the extra foam
5) Put incandescent lights on it for a few days to help it cure faster (I actually have a 4x4 hot box I can use for this)
6) When I feel confident about the PU being fully cured, glue on the last skin and add edge and corner protectors. (I may want to go a week in the hot box before doing this just to be safe)
7) Turn table top over
8) Add legs
9) Level with epoxy
10) Continue with next build steps

I am also considering using 1/4" top/bottom skins instead of 1/8". This will make the weight of the top and bottom skins go from about 160 lb to 320 lb. The siding, honeycomb core, foam, corner/edge protectors, and epoxy will all contribute to the weight as well.

[PaulAmelang]

I could do the same thing for the gantry, but a different bracing style would probably be more appropriate.

Here are the effects of internal diagonal bracing on a beam put under torsion, illustrated in a paper from DuPont:



CR Onsrud does this with their gantries:



Credit to Rob Taylor for posting this info on another forum.

[PaulAmelang]

Quick update:

It's very difficult to simulate the honeycomb core tabletop. I've been trying to figure out how I can simulate it so as to get useful results. In the meantime, here are some gantry simulations I've run. In my previous simulations, the top rail/carriages were taking about 1/3 of the load and the bottom was taking about 2/3 of the load. So these following simulations have -333N on the top carriage, and +666N on the bottom carriage. The left side of the beam is fully constrained.

SIM #1: Hollow beam, no core.


Results
Top: 9.5 um
Bottom: 24.4 um


SIM #2: 3lb/ft3 PU in middle

Results
Top: 5.8 um
Bottom:13.6 um


SIM #3: 1/4" internal bracing and 3lb/ft3 PU


Results
Top: 0.97 um
Bottom: 3.2 um


SIM #4: 1/4" internal bracing and 3lb/ft3 PU


Results
Top: 3.2 um
Bottom: 2.9 um


NOTES:

As you can see, the humble, low density PU is successful at adding rigidity without adding a lot of cost or weight, although the gantry performs best with internal bracing.