Gallery, Projects and General > How do I??
Quartz Surface Plate?
vtsteam:
Pekka, seems like you could get within your level of tolerance with a straightedge. Some long carpenter's levels are ground to a reasonable tolerance, and might be used with some blue marking compound.
PekkaNF:
I got one 500 mm long Flat straightedge, DIN 874/2 and two parallels, that are true to first one as far as I can tell. I also have 150 mm long hairline and according to that one these are pretty straight as well.
Problem is that this is essentially a line, I need plane that I can measure height of several points relative to each others and do it repeatedly.
I borrowed an accurate spirit level from a friend. It has graduation of 0,02 mm/m inclination and leveling the milling machine took some time. Took like a minute for the bubble to creep into final position (last few digits). Scary sensitive. Machine foot are 700*900mm grid and a fraction turn of the leveling screw sent the bubble often 5 divisions away.
I tested an 10 mm thick float glass piece and it's surface variation was a whole lot better than old granite kitchen counter top. Short distance accuracy was good, but it seem to need very stable/straight subtable to keep it straight. Placed it on top of the counter top and put only 600*50*10 and DTI /w base and it was not straight at all in 400 mm distance. Any 100 linear distance was pretty good tought, but it looked like it adopts shape of the table under when we are talking 0,01 mm or so reading when any load it applied onto it.
I have to test few objects and see how reasonable results could be obtained with far from ideal eguipment.
If the float glass or granite countertop turns out to be good enought, maybe I can build each measurement setting such way that the weight of the object (E.G. milling machine table of 50-100 kg) is suported on parallels, discs or such and the measurement table is supported right under these spots to prevent the flimsy measurement table from bending under the full load????
Dodgy?Maybe, but I don't think I have the skill/time to make three really big surface plates and then start scraping the heck out of the mill machine after that anytime soon.
The three palate method is something I should to in a small scale to make some support "discs". Is there a good guide where the method and logic is explained plainly and clearly?
Pekka
jcs0001:
Pekka:
Regarding supporting the float glass. I read some years ago about race car repair bays where they laid down a thick layer of epoxy which self leveled to give them a very flat surface to use for measuring suspension etc. on their cars. Can't seem to find it online but it may give you some ideas for supporting the glass.
A few years ago I made a small glass surface plate and bedded it in bondo (car repair filler) on a 3/4 in. plywood base. It did seem to "bend" a bit though. I actually made 3 and used valve grinding compound and a huge amount of manual work grinding them against each other. They turned out ok but still had low spots around the edges. I have one left and use it for a flat surface for lapping and the like.
John.
vtsteam:
JCS description of 3 plate method using glass and bondo beddding on ply comes from the Gingery books. Gingery later concluded that the bondo and ply shrank and threw the plate off in flatness.
Pekka I think you misunderstand me. The ground carpenter's level is to be used as a straightedge, not a level, and it (or any other straightedge) would need to be longer than yourf milling table -- say 1200 mm, minimum. I believe that such a level can be found at very reasonable price ground to less than a .001" overall, which I believe fits your rquested precision.
You do not necessarily need a surface plate for a limited tolerance on your milling table. You can test with a 1200 mm straightedge laid lengthwise at front and back and crosswise, as well as diagonally. Even a conical (twisted) straight solid will register as nonlinear if the straightedge is not placed directly on a radial.
When i built my Gingery lathe bed I used a 4 foot long ground carpenter's level, with a guaranteed .001" edge straightness tolerance over that length. I believe it cost about $40 US at he time.
You can also make a straightedge using the method of threes from hot rolled steel stock -- say 1/2" by 2" by 4' (or similar metric).
The method of threes is simple:
Grind with lapping compound any two surfaces (called say A and B) until they conform. This may not be straight, but they will be mirror images. Then lap a third surface (let's call it C) to A.
It will now be similar to B.
Then lap C and B together. Since they will have relatively corresponding high and low areas, the lapping will produce a reasonably flat surface. This method can be repeated if extremes of accuracy are required, but generally one go is sufficient for home machinist's needs.
I made a straightedge for the new lathe I'm building, but used another method of comparing it to a surface plate I already had, and scraping.
Scraping or lapping a straightedge by the method of threes would be much easier and quicker than doing the same with 1200 mm square surface plates, and because of the low mass, a straightedge needs far less stiffening, for your limited accuracy puroposes.
PekkaNF:
Thanks.
I'll get that. Works well on simple planes, but gets a bit more convoluted when there is a dovetail and other features stikking out.
See CoC.
I know there is a way, three points makes a imaginary plane, but it's easier to measure from the real plane nearby than from points of any real distance.
I was thinking of checking the table (worst part) first. It looks like table surface and dove tail planes should be parallel. This should be easy to check with micrometer. I hope there is few original/straight/pristine spots left I can establish this. Then I'll mount the milling machine table on top of the "plane" from these spots. Short paralles or such.
Then I could check the wear on dovetail guide surface that should be straight and coplanar to milling machine table and surface table. This measurement should be fairly straightforward with a comparator/scriber foot and DTI.
Primary goal is not perfection, but correction to the right direction.
Pekka
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version