Granite surface plate problem

[PekkaNF]

Unbelievable......

I have two surface plates, one is cheap chinese 200*400 mm and it is pretty accurate. I have faith in it, because it gives logical results.

Then I got one almost free, just perfect size, 630x400x100 mm, brand name (MAHR), but it has edges chipped, dents all over, one end rough and what is worst, it is not very straight!

I used brayer and rolled pretty thin layer of blue and armed with SIX straight edges of varying pedigree from 400-1000 mm length.

Idea was to check all sides and diagonals. I did 8 lines and union flag patterns with one straight edge a time, cleaned the edge and leveled the blue on surface plate between each.

Looks like the surface plate has low front, worn on the middle and rough spot on one end.

Pretty annoying really, because that was supposed to be my reference to check all other instruments.

I am pretty sure that only way out of misery is to buy a new or checked surface plate. Nobody here seems to have heard of surface plate reconditioning for a private person. It's only 75 kg or such I could transport it easily at the back of my car some 200 km. But nooooo.

Now the question: Chinese or (east) european? is there price break point that defines most likely good vs. most likely bad?

Pekka

[Joules]

Look up a company that supplies and polishes head stones.  You may be able to have the surface ground and polished.  It won't be as good as a precision surface but better than what you have now.

[PekkaNF]

I visited local rock mason shop and they had plenty of shiny rock. I had 0,01 mm/div. clock and straight edge with me. All had a lot of twist and waviness. Way too much. I think the polishing is is only visually fine.

I thought that they just had bad machines and went to other shop. They had a very nice saw and that produced pretty straight cut, but those plates were not near surface plates after they come out of polishing. Really weird. Maybe the rock they use is visually very interesting and proverbial to grind flat? Some of the prettiest granite you can feel the unevenes of the surface.

Pekka

[Manxmodder]

As Pekka notes,most stone mason shops have grinding/polishing equipment that is suited to producinf a good aesthetic appearance. High precision flatness isn't of any concern to their purpose so long as the product looks good to the eye.....OZ.

ETA: Mahr is a reputable brand in metrology equipment,so the surface plate you have is likely worn.

[BillTodd]

Have a search on youtube for oxtool's video of having his plates reconditioned. interesting and gives a good idea of how to go about fixing the plate.

Mahr is a top notch manufacturer so will be the best granite available and probably well worth reconditioning.

Bill

[PekkaNF]

Yesh. I have seen some of those videos and gained a whole lot better understanding what it really takes....way over my skills and tools.

You probably mean this one:

I did some reading on and something along with Repeat-O-Meter and Planikator could be a start, but buying new surface plate would be more cost effective and faster way. Maybe, if I had most of that stuff ready

http://www.starrett.com/metrology/metrology-products/precision-granite/standard-products/repeat-reading-gage

http://www.starrett.com/metrology/metrology-products/precision-granite/standard-products/planekator

http://www.benchtest.com/w-doall2.html
Pekka

[BillTodd]

Interesting to see that the Starrett repeat o meter gauge has a Mahr indicator :-)

[sparky961]

Get two more pieces of granite from your stone guy that are roughly the same size.  Also pick up some silicon carbide powder or better yet diamond paste.  In a pinch you can use a silicon carbide grinding wheel and pulverize it a bit.... but the former would be easier.

Start rubbing.

The 3-plate method of originating a flat surface (three, actually) is well documented elsewhere so I won't go into it here.  I had a thread a while back where I was doing this with granite/marble countertop scrap.  I can tell you it's boring, messy, and extremely time consuming.  BUT IT REALLY DOES WORK.

Maybe you can find two more surface plates that are in similar disrepair, lap them all together, and sell the ones you don't need when you're done.  It sounds like you can do a pretty good job of verifying your work, so you can feel good about a hefty markup for resale.

[PekkaNF]

Get two more pieces of granite from your stone guy that are roughly the same size.  Also pick up some silicon carbide powder or better yet diamond paste.  In a pinch you can use a silicon carbide grinding wheel and pulverize it a bit.... but the former would be easier.

Start rubbing.

The 3-plate method of originating a flat surface (three, actually) is well documented elsewhere so I won't go into it here.  I had a thread a while back where I was doing this with granite/marble countertop scrap.  I can tell you it's boring, messy, and extremely time consuming.  BUT IT REALLY DOES WORK.

Maybe you can find two more surface plates that are in similar disrepair, lap them all together, and sell the ones you don't need when you're done.  It sounds like you can do a pretty good job of verifying your work, so you can feel good about a hefty markup for resale.

Have you ever tried this?

I did try to make a straight edge by scraping three this way least 18 years ago and I can tell that I'm not using it

It is not easy and it is not very automatic. Huge time (and effort saving) is available when you have known straight level to check. And it is another long trench war before you get from narrow straight edge to a surface level. Adds a whole new dimension.

And scraping will produce predictable results. Lapping will just even out minor very local surface faults. Lapping on those videos looks easy, but they don't show how the lap was straightened.

I'm still looking to have the Mahr plate reconditioned, but if that is not possible, I'm trying to buy one cheap/half decent and cross my fingers.

Pekka

[sparky961]

Yes, I've done this. I used Silicon Carbide from a green grinding wheel. I presume diamond will cut faster and break down slower (if at all).

I haven't tried this with metal, but I expect that would take a lot longer by hand.

If the plate you have is in such horrible shape, is there any harm in trying? Mess it up and you still have plenty of material for professional resurfacing later on.

[PekkaNF]

If the plate you have is in such horrible shape, is there any harm in trying? Mess it up and you still have plenty of material for professional resurfacing later on.

Well... the undulation somewhere 0,01 mm or something like it. Pitting normally does not matter, because any instrument will bridge over it. it Might be good for garage.

But it is not good enough for checking the straight edges, cast iron surface plates, laps and other references. Therefore I need one good one inside, where it is on pretty constant temperature, little drift, good light, less dust etc....

Very probably I will check it against next surface plate, see where it needs leveling, prepare as good laps I can make and try to lap it bit more true. After all I can do most of the stuff I need with it and if I have "non metrology" surface plate in garage it saves me from going back and forth.

i would very much like to hear more details how the lap was made straight, grit and measurement of the flat....that is hard to me because I have no coordinate measurement machine or planikator at disposal.

Pekka

[sparky961]

You have presented a bit of a paradox here.  In your first post you tell about using six different straight edges to check the surface plate.  Now you're talking about needing the plate to check the straight edges - presumably the same ones.

The whole point behind the 3-plate method is that you can ORIGINATE a flat surface (THREE, for those who didn't pick up on that already).  By definition, that means that by the very process you will end up with a flat surface (or three... ok, I'm done now).  With this flat surface you can check how flat other surfaces are by comparison.  The "lap" you speak of is each plate against the other two in turn.

Refer to "The Foundations of Mechanical Accuracy", Wayne R. Moore, Page 24.

They describe the process using cast iron surface plates and scraping but the procedure is identical for granite and abrasive.  When rubbed together, high points/areas on "A" will be worn down by adjacent touching spots on "B".  Likewise, those spots on "A" will remove material from "B".  Do it enough, with sufficient rotations, the same amount with each pair, and using finer and finer abrasives (getting close to flat just the two together with some water is plenty abrasive enough) and you WILL end up with flat surfaces.  You verify them against one another.

If you're a real masochist, you can do the whole process twice thus ending up with six plates.  Compare between the two batches to remove all doubt.

I'd love to make a machine that does this automatically while I sleep, and sell the extremely flat plates for a hefty profit.

..... But more to the point, is this an intellectual exercise or do you really have a need for this level of accuracy?  If you do, you can satisfy that need by spending A: a little money and a lot of time (originate a flat surface), or B: a lot of money and a little time (purchase new or resurface professionally).  "A" is the typical hobbyist approach, while "B" is commercially favoured.

With the latter, you still can't verify the flatness without using a comparison surface of equally unknown flatness.  You're just trusting that it is as good as they say it is.  You can probably trust a reputable company, but when it comes to Chinese goods...... well, we've all been there, eh?

[sparky961]

...Then I got one almost free...
...edges chipped, dents all over, one end rough...

This should have been your first clue that it was not treated like the precision instrument it should be.  Where I work, the surface plate might as well be just another table.  I expect this is fairly common except in places where they're working with incredibly tight tolerances.

Looks like the surface plate has low front, worn on the middle and rough spot on one end.

This is the typical wear pattern, except for the "rough" spot.  Ideally a surface plate should be turned occasionally and the enlightened craftsperson will try to use different areas of the plate to spread out the wear.  Instead, most shop monkeys will have their coffee sitting in one corner, a greasy rag smeared across the front, and their 2lb ball peen hammer sitting in the middle right next to that big missing chunk.

[PekkaNF]

Yeah, it should have been a hint.....it still surprices me that such a precision instrument is treated as a kitchen table or worse.

There is no paradox....I haven't checked the straight edges conclusively yet....I have two each length camelbacks and one square flat and one trianangllasr one....easy to cross check that all but one are close to straight, but each two does not agree completely. Now if I had a absolutely good surface plate it would be easy to veryfy which way the slight error exists and what sort of action it warrants.

Originally I used all traight edges (some pretty good and the pretty good ones agreed close, but not exacly on each others) to see if they print consistely on the surface plate. If best traight edge prints the same on all sides, crossed and diagonals then i would have thought the surface plate is ok and probably is good enough accuracy.

But the prints did not stay constat on every straight edge.

Therefore my conclusion is that surface plate is not good enneoug to check those straight edges. One side of 800 mm straight edge printted fairly constantly union flag and poor on the better side of the plate, I deem that side sligtly convex, other side aproximed more straight.

I think.....

Pekka

[PekkaNF]

Some pictures of the hangar queen.....

[sparky961]

Yup, you still have a paradox.  You're comparing multiple unknown surfaces to one another.  You have no idea when the straight edges are being printed whether the surface plate or the straight edge isn't flat.  Even if three of the straight edges printed perfectly in multiple positions on the plate, as unlikely as it is you could still theoretically have a spherically concave or convex surface.  Desirable for optics, not desirable for machining.  Try printing a concave to a convex and you won't see it.  Take 3 concave/convex in any combination and you will.

While the chunks out of the edges are nasty, you've still got a nice chunk of quality granite there.  Of course, I'm going on the assumption that Mahr wouldn't put their name on crap, but I have no evidence for that.  Even it it has some nicks, it still has potential to be a very nice plate with some work.

If you haven't already, check out this video.  They're using diamond charged lapping plates, but I'll call your attention to a very interesting exchange at about 27 min into the video where the technician is asked how the laps are reconditioned.  He replies that they're doing it as they're reconditioning the plate.  "I'm making my equipment better and making your equipment better".  One wears the other - on a much smaller scale than most people are able to think.

(Direct link to time mentioned above)

=1591

[PekkaNF]

RE: video...I was wondering that too, but I think that FIRST you must have pretty damn "perfect" lap to start with. I was thinking of buying three new disc brake rotors, lapping them against my lap and checking them against known good surface plate (probably have to buy new....thinking of ordering one today) and when I'm satisfied that they are close to charge one with coarse diamond, other finer and last even finer and attack the bad surface plate. if the surface plate is about 400*630mm in surface, about 300 mm rotor appears big enough to even out if you know where.

RE: contradiction....you are right comparing unknown to unknown is bit shot in the dark, but I had two 1000 camelbacks to compare with one 800 mm straight edge (two sides, both looks fine, but other one is more logical to other). I printed them against each others and they do agree close (not perfect). Now when I print them in order against surface plate, there should be same sort of print on each straight edge/camel back.  Right. If the straight edge is just a tad convex it should make the concave print cross the surface plate. Well, it did partially....And most straight camelback appears to be a bit convex and that should print whatever and it does but on the middle it just prints "better".

And same thing with two 600 mm camelbacks and 400 mm camelbacs. And dial indicator on height gauge shows that something is not quite right.

When I put another well kept 500 mm ULTRA brand straight edge on top of the 0,05 mm slip gauge strips on marks and slide 0,03 to 0,05 feeler gauge strip I can feel that not all over the plate feels the same. Far long side feels markedly most constant (knowing that this straight edge has a droop spesified) and on the middle "droop" seems to disapear....hastily I would conclude that at the measured point the "rock" has a valley close to droop.

I still does not KNOW how big the error is but I can say that I am pretty sure it not quite close enough to check all the camelbacks and straight edges.

pekka

[leg17]

.......
...... Nobody here seems to have heard of surface plate reconditioning for a private person. It's only 75 kg or such I could transport it easily at the back of my car some 200 km. But nooooo.
Pekka

Find out who does it and contact them directly.
Some of them are 'on the road' and do it on-site.
You might be able to work something out.