[harryn]

Been working on using surface plate technology to save money and machining costs.This relates to the Y carriage and Z plate.Typically aluminum plate is used for this purpose.One sq/ft 1" 6061 is $212.00 and the weight is 14lbs and thickness tolerance bad.
I propose to use A36 hot rolled steel channel,C12X25 which is 12" wide and 3" legs,3/8 thick weighing in at 25lbs/ft.The cost per ft is $39.00.The thickness tolerance is very bad also.
Place the channel web down and use tape as a dam.Pour table top epoxy and let cure.Now the backside is perfectly level.Next turn over and tape the ends of the channel as a dam,level and pour epoxy.Now we have two precision surfaces to mount rails and blocks and ballscrew support bearings.Repeat the process for the Z plate.Any comments?
Larry

Hi Larry - This sounds like an interesting idea. I had an interesting slightly related experience last week - a firm I consult for has a process to epoxy top coat silicon wafers (semiconductor applications). Basically, the wafers are single crystal, high purity Si approx 8 inches in dia x maybe 1 mm thick.

The part that was interesting was just how much the wafers can warp from the coating and curing process. The cure was done approx 200 C IIRC, and the difference in CTE was significant, and obvious.

I guess the point I am trying to make, is that the garage floor epoxy system I looked at used a 3 step process - probably for a reason. It MIGHT be helpful if you assume that your "leveling of epoxy on steel" concept will take more than 1 coat on each side.

Just curious - do you have a good idea on how to use this concept to level and square some angle Al maybe 3 x 3 x 1/4 in x 6 ft ?

[greybeard]

Hi Harryn.
I, too, am looking at using this approach to "true" up the surface of component pieces.
The only idea I can offer is along the same lines as yours, of multiple pours, but how about doing a quick brush coat on one side of the component, then turn it over, drop it onto whatever jig you've set up to get it level, then do the "table-top pour" on the other side.
While this reduces the warping stress, familiar to people veneering panels, what I can't figure out is what happens when you do the second pour over the brushed coated side to true that one up ?
Perhaps the only answer is to always fabricate anything that needs two opposed but true faces from two layers with their bad sides inwards.

On second thoughts, is this really a problem, given the difference in bulk of Larry's channel and that of the resin coating ?
John

[jhudler]

Harry,
Three steps are (if I remember)
1. seal the floor to prevent it from out gassing bubbles, and perhaps add color and/or decals.
2. Apply intermediate layer to level the system, and repair imperfections from seal coat
3. Final level coat to achieve desired results.

If you want to test the straightness or flatness of something you can use a laser interferometer (measures the angle of two parallel beams that are 90 degrees out of phase), an autocollimator (essentially a telescope with cross-hairs, light source, and very flat mirror), or....
Hook up your high MegPIX digital camera on a level and steady tripod and take pictures of a fixed target (90 deg block with a line inscribed on it) at precise intervals (say and inch) along the surface. Then count the pixels of deviation of the scribed line. This gives you the displacement in pixels (which have to be adjusted for distance). (note: huge amount details left out here, but I'll leave it to the reader to fill in)

You measure the whole surface to find the low spot, then polish down the high spot... this assumes it's pretty close to flat already, otherwise it's sandpaper time!

I use a laser interferometer only because I have one. I've used this to take a B Grade Starrett granite surface plate and lap it with Cerium oxide to +-.000015 or 15 millions of an inch.
Has to be done in a constant temperature environment and takes a long time.

[lgalla]

Thanks for the input guys.
The floor plate requires 3 coats as the concrete surface is sanded and may have cracks or expansion joints which may have air which would expand and rise ruining the surface.This applies to table tops also as you need a prime coat.Steel is non absorbing which would only require one coat or a first brush coat then pour as John suggested.An1/8 " of epoxy is not going to warp a steel plate.If this works we could build accurate machines with minimum tools and no trips to a machine shop$$$.
Harry have think a bit on the Al angle.
I have THK HRW-21's for the Z axis.These are low profile wide blocks making it necessary to use spacer blocks to raise them up to fit the ballscrew between the rails.Possibly I could pour some E/G into the channel under the rails as the spacer and leave the centre clear for the ballscrew.Any ideas?
Thankyou
Larry

[walter]

Jack,

Thank you for that awesome movie. Seeing is believing!

Which pump did you use? Like this http://www.cnczone.com/forums/showpo...postcount=1365 ?

On the subject of laser flatness. Will you be able to investigate the epoxy surface plate meniscus? There was a discussion about whether it will cause any problems http://www.cnczone.com/forums/showpo...postcount=2037
This seems like a serious concern.

Once again thank you for an excellent job. You're officially admitted to Mad Scientists' Club!

[greybeard]

I have THK HRW-21's for the Z axis.These are low profile wide blocks making it necessary to use spacer blocks to raise them up to fit the ballscrew between the rails.Possibly I could pour some E/G into the channel under the rails as the spacer and leave the centre clear for the ballscrew.Any ideas?
Thankyou
Larry

Thin wall plastic pipe down the center, back fill 3/4 deep with resin, remove pipe with big heave ?
John

[jhudler]

Jack,

Thank you for that awesome movie. Seeing is believing!

Which pump did you use? Like this http://www.cnczone.com/forums/showpo...postcount=1365 ?

On the subject of laser flatness. Will you be able to investigate the epoxy surface plate meniscus? There was a discussion about whether it will cause any problems http://www.cnczone.com/forums/showpo...postcount=2037
This seems like a serious concern.

Once again thank you for an excellent job. You're officially admitted to Mad Scientists' Club!

LOL I didn't see... errr... remember that post, but yes it's one used by refrigeration techs and pulls down to 25 microns. There's no need to go lower than this for two reasons; price and contaminating the pump you paid dearly for. I got mine for $121 which make it a throw away pump in my books. I do recommend placing a filter in-line to keep the Zeeospheres and other aggregate dust out. Mines down to .1 microns and could be made out of 12 inches of 1 inch PVC filled with soft and dense foam or layers of a cheap HEPA filter with the edges glued to the walls.

Buy the pump new, don't try and get one of those Edwards or Varian pumps cheap because repairing them is not cheap.

I guess the rule should be... if you spent more than $200; you spent too much.

On the meniscus issue: Why all the concern about turn up or down edges?
From my experience it's always going to be turned up.
Might be a bother if the work piece over hangs, but this can be filed/sanded/scraped/machined down (polished) and move on.

Testing flatness in an area is pretty easy using a 1/4 wave flat, which should give us resolution of about 100nm or 4 micro-inches. I've got an 8 inch one.
Seems to I remember using a ronchi grating to gauge a flat... wonder if that's easier.

On Laser testing the over all flatness, I could do that in time. It's a very time consuming test.
What size of plate do we need to test?
What is the proposed resin?
Does it have any mineral content? (Zeeospheres would be perfect)

Jack
fellow "Mad Scientists' Institute"

[walter]

On the meniscus issue: Why all the concern about turn up or down edges?
From my experience it's always going to be turned up.

Some think it might go all the way to the center. Like a giant dish.

[the4thseal]

scale is an issue but i think that it does...... i am looking through my old engineering texts for a technical example. though i think it becomes so shallow that it does not matter. the question is how much and does it matter at the scales that we are working on. also what are the means for mitigation...i think some satisfactory ones have been discussed. the picture is great.

[lgalla]

John,thank you,good idea.Any ideas how to drill through to the steel below?
Jack I would not worry of abrasives getting into the pump if you manually mix the aggregate and resin first,then vacuum and vibrate.I have a rather pricey Busch vacuum pump which ran 8hrs a day,10years pulling a vacuum on polyester resins with nepheline syenite fillers.Polyester's actually lose some of the styrene under vacuum and this never was a problem.Epoxy will not vaporize any components.The pump doubled as a hold down on a pin router.Regular oil changes will extend the life.
Jack I would also file off the meniscus.If the meniscus went all the way to the center[which I doubt]It stands to reason it would only be out.005 or whatever flatness Precision Epoxy claims
Larry

[jhudler]

Some think it might go all the way to the center. Like a giant dish.

I was refering to the edges which will turn up or down.

IMO, I don't think it would dish up or down in the middle otherwise, those level floors using self leveling epoxy would not work. From the picture in this article I see no dishing at all. Which would be apparent in the reflected lines from the walls and ceilings, see this page.

Jack

[the4thseal]

if the boarders are the same material as the epoxy than the critical surface tension would be the same or so close that it would not matter. therefore no or very little meniscus. it does not mean that there are not techniques to counter the problem. also if we are talking about a couple 5 thousands over 10 feet or so... that is very very good, and would not show up in the type of photos that are linked.overall i think that it is not a problem, just something to keep in mind and to deal with. do not get me wrong .i love this stuff....like everything there are properties and tolerances that have to be taken into account. i use this stuff... so i am not saying do not use it.

[jhudler]

if the boarders are the same material as the epoxy than the critical surface tension would be the same or so close that it would not matter. therefore no or very little meniscus.

Surface tension is molecular attraction, so what happens at the edges depends on the charge state of the epoxy and the material used for the border.

This is off the cuff, so I'll go off and do some research.

Jack

[the4thseal]

that is correct.... that is why i said that if the materials are the same it should be the best possible situation. i do think that we are making a mountain out of this. the stuff works well enough for the intended purpose. i think we chased our tails on this earlier.....though i will be making a poor soon and this time i will obsess and post pictures. though as usual the more input better.:-)

[jhudler]

ditto!

[the4thseal]

one last comment on the meniscus............depending on the materials the meniscus can go both ways..... there should be a water cooler graphic to denote casual musings and the like,,,,, versus earnest proposal and debate ..... it can be hard to discern the difference.... all are great but not the same.........just a thought..........

[ckelloug]

First off, thanks all for 8 months of good discussion. It made me a lot more ready for the materials questions on the FE Exam I took yesterday. Thanks for the encouragement too. (John especially).

On the topic of menisci It's probably not a problem on a relatively large pour. It could however be a problem if one wanted to pour a layer of leveling epoxy over the surface of a relatively narrow beam while it's still in the mold for something like leveling linear bearing mounts.

The wood router guys can probably take a bye on this one but for extremely accurate metal working equipment, one must either solve this problem or diamond surface grind everything flat which somewhat defeats the purpose of using E/G.

There will always be a slight curve to the epoxy surface regardless of the meniscus problem due to the fact that gravity points a slightly different direction at every point on the casting. (curvature of the earth effect). As a result, you can't do better than the curvature of earth when gravity leveling castings.

For the wise-acres here, you might be able to look at gravimetric data for earth and find somewhere where there are iron or uranium deposits in the right places to alter the local gravitation field enough to get a flat casting. This exercise is left to the reader. (The foregoing paragraph is absolutely true however impractical it may be. I once spent some time at my day job writing software to look up the gravity field data for a given location on earth. )

Congrats to Jack et. al on all of the tests lately. I'm still standing by with my flexural tester to test any samples provided. 5 per formulation 3/8 x 1/2 x 7 inches is the biggest D-790 sample I can handle right now.

As for Segregation and John's comments, it is true that de Larrard typically measures packing density of aggregate under pressure and vibration but he also studied segregation in concrete mixes which include icky portland cement and water. I probably didn't mail the section where he's talking about segregation but the formulation that we're using is the one that his formula for segregation potential says will segregate the least.(Generally speaking, the larger the size difference between the components, the more likely segregation is to occur). If you're following the logarithmic size rule as discussed before, this is the solution to minimizing de Larrard's segregation potential. It should also be noted that if any components are dropped from the aggregate design, the proper size ratio between the remaining components changes.

Regards all,

Cameron

[svenakela]

Sven,
What you said is valid if you're not vibrating in a vacuum. Otherwise when you see it start to rise, slow or stop the vacuum and let the vibrations and vacuum liberate the entrapped air until it falls back. Then resume the vacuum and continue to vibrate until air bubbles are no longer seen on the skin.

There's no need to pull a full vacuum all at once.

Jack

You're goin' crazy guys, I haven't been checking the thread for one day and my last reply is pages away...

Jack, If you pull the vacuum really low there's no need for vibrating as there will be no air left at all. The attached picture shows a piece that is vacuum casted. Even though it's thin walled there's no risk for bubbles, the mould fills up nicely without being pressurized or shaked.
But I see the idea when you vibrated too. I don't say it's wrong or right, just giving the options.

--S

[jhudler]

You're goin' crazy guys, I haven't been checking the thread for one day and my last reply is pages away...

Jack, If you pull the vacuum really low there's no need for vibrating as there will be no air left at all. The attached picture shows a piece that is vacuum casted. Even though
--S

Really low is relative , we are at 25 microns or .025 mm which some concider really low vacuum for degassing. We are allready in a risky zone; this is *WAY* below the boiling point of epoxy (~40 mm @ 20C), leave this too long in vacuum and it starts to boil, like water does at 25 microns. This doesn't take into account the temperature rise from curing.
See this; http://www.youtube.com/watch?v=3capdSVddwY however, his conclusions are wrong.

Why vibration? E/G rises during vacuum but does not settle back very well, leaving voids were the expanded and coalesced air bubbles existed. When returning to normal atmosphere, you see it slump, this mean entrapped air still exists. See the picture of Sample 3 (bottom) http://www.hudler.org/pub/eg/Three Samples edge.png.

It was most clear from the experiments that you had to vibrate and vacuum simultaneously due to the density of E/G.

Or in other words; E/G is so damn thick the air bubbles can't get out!

Jack

[walter]

Jack, here's something to inspire you http://img142.imageshack.us/img142/1336/minimillzs3.jpg
to get back to that Rong Fu project.

(yet another gem from that German site)