Accurate Scrapping on MS Plates

If the Earth were perfectly spherical then it would be out-of-flat/planar to 0.0005mm in 10m. This is why asking 0.02mm in 10m for a machined surface is such a big deal, as it is only 40 times rougher than a perfect sphere the size of the Earth. It's going to be some glorious machinery that is able to achieve that level of smoothness, which probably hasn't been invented yet.

Is this proposed installation in an earthquake-free zone, BTW?

You might need to look into how large optics (e.g., telescope mirrors) are finished.

never seen a mobile milling machine that size, mobile lineboring yes but not milling

i would look at milling the top of both plates before setting them in concrete

fabricate a spreader beam with milled faces underneath that your two plates bolt up to

the spreader beam keeps the tops of the plates precisely flat while you lower them both on to the concrete plinth , when the concrete has gone off unbolt the spreader beam and lift it away , your plates should be precisely as they were positioned

so is someone making a satellite dish mounting pad or space telescope mounting pad ?

my consultancy fee on the above ......

You have an interesting challenge. I hope you are doing this on a time-and-materials basis and not on a fixed price!

This must be a military or scientific application--I agree with other posters that the flexibility of the earth may tend to be enough to allow dissimilar movement between the two plates, rendering the tolerances unreachable for any length of time. How does the end user propose to measure the flatness tolerances after you turn it over to them?

Certainly, pre-flattening both plates before placing them is a must. The flatness tolerance is just under a thousandth of an inch so should be achievable with high quality machine tooling such as a surface grinder.

Using mild steel, you will have to ensure that there are no residual stresses in the material that will cause it to move after grinding. Even day/night temperature differences will challenge the tolerances.

You probably will need to connect with contractors who specialize in alignment of large machine tools since they will have the necessary laser alignment devices to check the flatness and co-planarity of the two plates. Presumably, you can accurately preset the plates before concreting them. If I have misunderstood and you have two plates that have been tossed in the concrete and have to be flattened in place, then the problem gets a little more interesting, since you then need to bring machine tools to the plates.

It would be my guess that a combination of portable machine tools and patient and skilled laser measuring will get this job done.

Keep us posted on how you are making out!!

Jon.

That's a tall order. My best advice is to hand scrape one plate flat and level. My master level reads to 0.0005 inches/ft. Better ones do exist. The second plate can be measured from the first plate using a jig transit or a theodolite. I would set up a height gauge on the second plate and the optical instrument on the reference plate. I'm pretty sure this arrangement will get you close to the tolerance you specified. Thermally induced distortion may be your biggest problem. The amount of distortion will depend on the thickness of the plates and how they are fastened to the mounting structure.

Getting the individual plates to the desired flatness is certainly possible with scraping (not scrapping... that would be very expensive) precision grinding, lapping, etc.

However, keeping the plates coplanar to those limits over such a great distance, with concrete as the substrate, seems impossible without active controls. Day to day variability due to temperature, humidity and micro seismic activity would cause those limits to be exceeded. Connecting the two with a massive iron casting (like a huge lathe bed) would eliminate humidity effects, and mounting that bed on compliant mounts would nearly eliminate ground change stresses (this assuming that the entire structure is allowed to rock with respect to gravity, as long as the plates maintain their alignment.) Very careful design of the distribution of mass in the iron bed would help to ensure that as the bed expands from temperature changes, that it does so only in the length direction, rather than warping, as a structure is likely to do with temperature changes.

Such a structure (a massive iron bed) would probably eliminate the need for active compensation. Otherwise, you'd need something on the order of motor-driven precision jack screws under the corners of one plate (under control of CMM) to actively keep it in line with the other.

Starting with surface plates might help. In this thread, someone had one 1m x 2m.

BTW, I am assuming that the tight tolerances are really necessary. It is not at all uncommon that a drawing shows up with unworkable tolerances (the result of an inexperienced designer), and the machinist dutifully works on the project, only to find that the tolerances could have been looser.

yup . checking with the engineers who requested it might well show an error and a little communication might open up vastly different ways to obtain their intended outcome.

just a thought on the use of either a cast iron one piece bed imbedded in the ground for example , or my idea of using a machined one piece spreader beam to hold the two plates in position while the concrete set........

how much influence do you think it would have using nitrogen to fill the interior space inside such a beam or bed ?

the changes in outside air temperature is still going to cause expansion , but to what degree ?

just thinking hypothetically if such a job came up , you could coat the outside of a beam or bed with a thermal insulating material and fill the inside with nitrogen to stabilise thermal fluctuations.

What.

not worth the expended oxygen

What just happened????...

Methinks that http://cr4.globalspec.com/member/46045/zen has escaped again!

If it's allowable, you could build a frame around each plate, put a bulkhead through the side of each frame, connect them with a hose of some sort, and fill them with a layer of liquid that will harden flat. They will find their own equal level.

After reviewing past threads, and reading yours, I think you deserve a GA. You were on the right track, but lacked the knowledge to carry through. Nothing beats experience.

Sweet solution.

It would be nice if the OP could tell us what his coplanar surfaces will be mated against.

Hm.

GA from me. We use the same methodology for large wing jigs, tooling, etc. Plates are machined to tolerance separately. Holes in the plates are provided for laser tracker targets and leveling studs. Structural epoxy used to bond it permanently in place. All the major aircraft mfgs use the same technique for accurately locating and fixing large structural tooling components.

I have can't see why anyone would want to machine something like this on site. Makes no sense to me.

Okay. I don't usually do this, but this time it's rather blatant, and I'm without coffee and in an ornery mood this morning.

But.... with all due respect to ronseto.... is my answer to this problem invisible? All ronseto did was add a brand name to the epoxy solution that I came up with. Every post prior to mine was talking about taking away mat'l, til I suggested building up to achieve co-planarity.

Sheesh

Quote: "If it's allowable, you could build a frame around each plate, put a bulkhead through the side of each frame, connect them with a hose of some sort, and fill them with a layer of liquid that will harden flat. They will find their own equal level."

I read all the responses before adding my own. I didn't understand what you meant by a frame and bulkhead through and a hose of some sort. The Chockfast method doesn't require the levels to be coplanar to each other. Only the top plates will be coplanar. The thickness of the Chockfast will fill the gaps between the plate and support. What you were implying was; the two plates had to be machined on both sides and coplanar, because the top surface of the liquid would be coplanar. In the Chockfast method, only the top plate will be coplanar; any other surface below the top plate doesn't have to be coplanar, or even smooth or level.

No, I'm not faulting your suggestion at all. It's a good idea. I was just rather proud of my moment of inspiration, that happened to go opposite to the train of thought up til that point; which you then built upon. My suggestion, with the two pools of liquid, prior to hardening, being connected by the hose would have made them coplanar to each other... as the OP was looking for. My approach appeared to go unnoticed.

I guess I'm just a genius in my own mind

Just to set the record straight; I didn't have to build on your suggestion. My answer was based on 50 years shipbuilding experience. This just happened to be common knowledge to me. I didn't expect anyone else to come up with a similar solution. This has been around a long time and anyone who has worked in shipbuilding knows this firsthand. Have a nice day.

Okay then.

Record straightened.

To further straighten the record, Ronseto's solution and yours are entirely different, with about the only similarity being that resins can be used in each. In his well-proven solution, the resin is used to fill a gap to provide continuous and uniform support under a plate that has been leveled using conventional techniques. In his solution, the OP's desire to use mild steel plates (we can only assume that is what is meant by MS) is retained.

Your solution could provide a very flat surface onto which plates could subsequently be placed, in which case, both sides of these plates would need to be machined to high precision, and the tolerances on the individual faces would need to be tightened even further, because the out-of-flatness on one side will affect the out-of-flatness on the other.

Which is not to say that your suggestion has no merit, only that it has little to do with Ronseto's solution, which is not to use a resin as a water level, but instead as a gap filler.

Got it. But I wasn't suggesting a plate on top. Just the level hardened, co-planar resin surface. Since OP hasn't returned, we don't know for sure that a plate was needed on top. Perhaps just two "perfectly" level hard surfaces.

There is a possibility that the original question was an exercise in logic, and not something that actually needed doing. For instance, to see if anyone considered building up the surfaces instead of machining (scrapping ) them down; which I can't see ever happening, in the environment that was described, and the distances between the plates. It sounds a bit like a professor's challenge for extra credit. Especially with the added statement of "use of shims below the plate is prohibited :( "

I say this because theoretically, there may be several solutions. But practically, the circumstances and the tolerances described, given the natural minute fluctuations of most any surface on Earth (i.e. "granite bends")... it seems nearly unachievable, without some type of laboratory grade equipment and procedures.

And to maintain that precision while actually using it in some way, such as putting some weight or equipment on it, again seems improbable; at least without some formidable large scale foundation plan.

Theoretically, and practically, the solution I proposed may be the simplest way to get two surfaces, that far apart, to be co-planar to each other, within those relatively extreme tolerances.

IMHO

A suggestion that uses leveling screws and lasers to manually individually adjust all four corners of both plates to be in a co-planar position, at the high precision that's spec'd, might be easier said than done.

"A suggestion that uses leveling screws and lasers to manually individually adjust all four corners of both plates to be in a co-planar position, at the high precision that's spec'd, might be easier said than done"

It's common practice in aircraft tooling such as large wing jigs. Not only is it the only way to achieve that kind of precision, it's the only way to measure and be sure. I know because it's what we do. BTW, we typically use 3 leveling screws (not 4) because no matter the adjustment, all three make contact. With 4, one will always be "loose" and therefore serves no purpose in the adjustment. As I stated earlier, almost all aircraft mgfs (large and small) use this technique. According to Ron it's also common practice in shipbuilding. It's easier done than said

Cheers.

Yes, I meant three points of contact. I misspoke. My bad.

0.0007" co-planarity between two surfaces that large, and 33' apart seemed like it was unattainable, given it was an in-situ situation and not a lab-type, highly controlled environment with a substantial foundation. But if y'all say that's common practice, I stand corrected.

Uh...0007"?. Yeah okay maybe not! I misread the OP's tolerance as .2mm.

.0007" (.02mm)???....What the hell for?...Whoever specified that is smokin' crack!

As they say on Shark Tank (Dragon's Den)..."for that reason I'm out!"

Peace!

A suggestion that uses leveling screws and lasers to manually individually adjust all four corners of both plates to be in a co-planar position, at the high precision that's spec'd, might be easier said than done.
I agree. The bearing beds for ship propeller shafts, aircraft wing and control surface alignment, most machine tools, etc are not held to tolerances like this over such great distances.

i did note your suggestion as being very useful in similar applications but could see some small issues that would dissuade me using it in this case

will still giving you a GA just so you know the kilojoules of your energy were not wasted

:)

best wishes Mister Out Of Box

LOL

Well thanx, Mr. Impossible. The GA didn't seem to register, but I appreciate your words of aknowledgment.

OoBE

I think he means "bulkhead fitting". He suggested building a frame (metal?) around the plates the same way you suggested it with putty. The "bulkead fittings" would pass through the frame wall and the hose would connect the two fittings. Fill the hose with "liquid" and it would find it's own level meaning both plate bases will be "level" between each other.

My impression is that OBE is assuming the requirement is for the plates to be "level" relative to earth and not just coplanar to each other. His idea would work if the compound used is something like self leveling cement like you would use to do a tile floor. Won't work with epoxy because it's not viscous enough to be self leveling or flow through a hose. As you said, I think only the top mating surface of the plates need to be coplanar ("level" or any other alignment that's required).

BTW, just noticed OBE and I are 1min apart in a reply...

That's exactly how I was thinking TerraMan. I was also taking into consideration the OP statement that shims were not allowed, which I assume, the leveling screws that were recently mentioned, would qualify as.

I didn't have a specific liquid in mind, but was thinking of a two part formula, one of which may be liquid enough to flow through the hose and achieve level and coplanarity... and then the hose would be blocked off both ends and the hardening catalyst would be added to both pools. I'm not familiar enough with ronseto's suggested compound to know if that would work with my method. If not, perhaps something else exists that would fit the bill.

It would seem to achieve all of the OPs goals. But he seems to have disappeared, as is often the case.

Who needs an OP, we can still have fun with this...

Liquid levelling is fine I suppose. Why the hang up with resins and hardeners and viscosity issues through the balance tube when you can just use plain water and lower the temperature till it freezes.

How practical is that?!!

I noticed your suggestion but didn't give it a GA because I don't think it will work. Once the steel plates are attached to a floating frame, you have no clue as to whether they will be parallel to the liquid surface. The only way I can see this working is if the liquid is denser than the steel so the plates float on the liquid. This is possible,but involves complexities that we need not discuss. That being said, the lack of feedback from OP leads me to believe that this is strictly a mental exercise and not a real problem.

No, I apologize. I'm not explaining myself properly. I wasn't describing a floating frame. I was saying build a frame around each plate... in the same way that ronseto described the putty dam. Same thing. In fact, making the frame out of putty would be fine. The liquid (epoxy) gets poured on top of the plates. It will be surrounded on four sides by the frame/dam. The flat co-planar surface would be the top surface of the hardened epoxy, or whatever concoction was used. The plates themselves are still sitting on the concrete or whatever is originally under them.

We don't know what this set-up is for, so the smooth layer of epoxy on top of the plates may have fulfilled the purpose.

In fact, the hose I spoke of could just as easily been a narrow trough from one plate to the other. The resulting span of epoxy between the two plates could then have been broken away after it was hardened. This would allow the layer of epoxy to be even thinner. But just thick enough to form the co-planar surfaces on the two plates.

1. Check if the tollerance is correct, or it was some extra zero added (or it was in inches!).

1. The plates, machined previously with such accuracy, should be strong enough to withstand their own weight without deformation.

(No chance for tooling at site within reasonable cost, IMHO.)

2. Set a tent around the area you are going to work, and try to keep the temperature there as stable as possible (no direct sunlight!) and the place clean.

3. I would use plastic bags in the shoes when working there.

The surfaces where measuring must be checked, must be cleaned prior to each measurement.

4. Using 4 or 6 jackbolts, level the first plate. You can use a straight edge of 2,5m with a suitable master level for the first approach.

Later on, a laser theodolite should do.

5. After the levelling, you should have a gap of around 10mm between the plate and the foundation. I use Chockfast Orange, as mentioned by Ronseto, and it is an advisable choice to bond the plates to the foundation.

This procedure is used to level and align single baseplates of 4000x4000x120mm, within 0.06mm flatness (+/-0,03mm height) using around 12 jackbolts, and it is not easy.

Yours will be +/-0,01mm... good luck man.

PS: Out of curiosity, could you say for what are this plates?.

http://www.chockfast.com/2011pdf/659H_Chockfast-Orange.pdf This link shows how the Chockfast system works.

To OP:

When you read the instructions, pay special attention to the TEMPERATURE-HARDENER DOSING. Or you can end with a very expensive sandwich.

Let us know!!

I can't believe this - 45 post on this subject and the OP didn't respond once. I hate it when the OP doesn't participate in his own discussion.

Sorry friends,

I was exploring, proving your ideas...some actions were hit some were misses...

Still struggling to calibrate my Laser mc calibration...

Basic leveling is still under progress... I am trying to gauge the flatness of individual plates using Wyler and co-planarity using a Hamar laser...

Will post the developments soon.

Please do.

You still haven't mentioned what will be sitting on these have you?

these hamar lasers look very interesting

if you did decide to machine two plates , then sit them on top of a concrete foundation with three levelling screws under each plate , then pour a synthetic grout to fill the gap underneath with something like chocfast, after having adjusted them to perfect alignment , then tightening up some concrete anchor bolts after the chocfast had hardened , and postponed any earthquakes until you had been paid , would you then have time to let us know what sits on top of the foundation plates ??? lol

just kidding . just curious

if i was you i might be reluctant to divulge any information that might breach the privacy of the business relationship that you have with your client too

its been an interesting question to contemplate

best wishes.

:)