PDA

View Full Version : Making a very accurate try-square



loply
02-28-2013, 07:46 AM
Hi folks,

I have a very accurate cast iron square which is (literally) a large rectangle with a wide base - much like a precision box level.

It's square to two tenths and is 400mm high, but it's big, heavy and sometimes cumbersome to use as a reference when testing a machine.

I'm thinking about using it to make a large, very accurate traditional type try-square. My proposed method is:

1) Lightly clamp a precision ground flat stock 'blade' circa 500mm x 3mm x 40mm to the edge of my cast iron rectangle such that it's held vertically on the surface plate

2) Take a scraped cast iron 'base' of the try square with a milled slot, fill the slot with Araldite epoxy resin, and introduce the slot to the base of the blade

3) Put a weight on the 'base' so it can't move and leave the two parts sitting there to set

4) After 24 hours remove the weights and the blade should be left suspended exactly square to the base, subject to the same accuracy as my master


My motivation is that this method should allow me to make a fairly large (500mm) try square with an accuracy equivalent to my cast iron master square. To get large try squares that accurate is very expensive, most of them are permissible up to a thou or two out, whereas my cast iron thing is less than 2 tenths out.

My query is about the epoxy resin - Is this going to work? Will the blade simply 'move' when the clamping forces are removed? Will it creep or fail over time?

Would appreciate any input on this!

Cheers,
Rich

loose nut
02-28-2013, 08:47 AM
Pictures would help.

Mcgyver
02-28-2013, 09:30 AM
there may be speciality stuff that would be better, but in general i wouldn't trust epoxy to be rigid enough. otoh most other ways i can think of would required you to create the squareness after the fact, eliminating the idea of build to your square

check out the pic below; I've seen these a few times - adjustable shop made squares. Might be worth thinking about

I also think that if striving for that accuracy you might not be able to rely ground flat stock to be straight or parallel - might have to rework the edges. I know that the starrett claims .001", but that's its thickness. Flatness is not specified and width 0-.005". What they don't say is what the acceptable variance is in width from one end to the other but taken at face value it could be as much as .005" out and be in tolerance.

Flat, parallel and square to a high degree of accuracy over large distances its not easy which is why they cost so much :)

http://i20.photobucket.com/albums/b201/michael0100/DSC_4881-large_zps462dbc92.jpg

TGTool
02-28-2013, 09:58 AM
By coincidence the same question has been on my mind. I scanned a picture from the Moore book Holes Contours and Surfaces showing one being tested in their shop. It looks like it was also made there and is fabricated from a cast iron beam and (probably) a hardened blade. I can't quite make out how the two are attached. There is one section showing fasteners that might be a separate piece - that is, a rabbet (rebate) in the beam with a cover plate. Squares of this type can be adjusted and the following photo in the book shows it being checked. It's always good to check and I'm not surprised Moore does that regularly, but one function of checking would be to correct if necessary.

I'm sorry the picture is not as clear as the original which shows the scraping marks on the cast iron. Also note a recess machined in the beam to make it easier to grasp, and not just any old recess, but a bit artistic. More tool porn.

http://i48.photobucket.com/albums/f244/TGTool/Moore1.jpg

loply
02-28-2013, 09:59 AM
It didn't really occur to me that the edge of ground flat stock might not be straight, I'll have to check. I just presumed it would be done on a grinder of such size and precision as to not be an issue.

Of course I don't care what the dimensions are, just the flatness is required.

Instead of epoxy I might just rigidly affix the blade with screws or suchlike and scrape the base afterwards, for the time it takes to scrape/measure/scrape/measure a few times it may just be easier, now that I think about it.

Cheers,
Rich

metalmagpie
02-28-2013, 10:57 AM
One of the TMBR books has a series on making up 3 squares against each other, which can be corrected arbitrarily accurate depending on your patience. It is an analogous process to making 3 flat plates.

metalmagpie

Mcgyver
02-28-2013, 12:02 PM
It didn't really occur to me that the edge of ground flat stock might not be straight, I'll have to check. I just presumed it would be done on a grinder of such size and precision as to not be an issue.

it may be, but because its ground doesn not mean its flat. depending on shape size and setup getting things very flat on a grinder is not a given. keep in mind .0002 over that distance is a shop grade surface plate level of flatness, so even if double or triple that is acceptable, that stock may not be to that level. A piece, shims, blue and a plate will reveal all :)

doesn't have to a show stopper, but imo it should be checked and potentially the two edges reworked for flatness and parallelism. Sounds like an neat project. I still like the adjustable idea though as it lets you get things perfect before assembly and assembly doesn't risk any of that accuracy (ie via riveting)

TGTool
02-28-2013, 12:08 PM
One of the TMBR books has a series on making up 3 squares against each other, which can be corrected arbitrarily accurate depending on your patience. It is an analogous process to making 3 flat plates.

metalmagpie

And he describes how to lap the blade to get is square and flat if necessary. This would be the alternate to working down the beam to get the same goal. This might be a separate article - it's not generation of squareness, but just fixing a square that needs correction and has been tested by other means.

A square can be self testing with an indicator. That is, if you've determined that the blade is flat and parallel, you can then use an indicator setup, approaching from each side, to show a double error for out of squareness so it's quite sensitive. I'll see if I can boost the Moore picture of this test being done.

http://i48.photobucket.com/albums/f244/TGTool/Moore2.jpg

Okay, photo added. The indicator stand touches the blade at the bottom and the indicator at the top. Coming round to the other side, any out-of-perpendicularity shows as a double error. This looks like a T-square they made and use in house.

dian
02-28-2013, 02:46 PM
[QUOTE=Mcgyver;833876]there may be speciality stuff that would be better, but in general i wouldn't trust epoxy to be rigid enough. otoh most other ways i can think of would required you to create the squareness after the fact, eliminating the idea of build to your square

check out the pic below; I've seen these a few times - adjustable shop made squares. Might be worth thinking about

I also think that if striving for that accuracy you might not be able to rely ground flat stock to be straight or parallel - might have to rework the edges. I know that the starrett claims .001", but that's its thickness. Flatness is not specified and width 0-.005". What they don't say is what the acceptable variance is in width from one end to the other but taken at face value it could be as much as .005" out and be in tolerance.

Flat, parallel and square to a high degree of accuracy over large distances its not easy which is why they cost so much :)

why should it be rigid? there wont be any forces applied, right?

i wouldnt use araldit though, but low viscosity epoxy.

Evan
02-28-2013, 03:07 PM
I wouldn't use epoxy at all unless it was only to hold a metal to metal fit. If there is any gap the high coefficient of expansion of most epoxies may be a killer of accuracy. There are epoxies with nearly zero CoE but you would need to be certain to use that type. It can work, Mitutoyo uses epoxy to hold the anvil in place on their micrometers and there is a gap between the ground carbide and the frame. I know because I have reworked about a dozen of their mics.

dian
02-28-2013, 03:17 PM
hm, epoxy is known to neither expand or shrink. am i missing something?

Evan
02-28-2013, 03:18 PM
Not all epoxies are the same. There are large variations in the CoE of various types.

TGTool
02-28-2013, 05:03 PM
That Moore square has been nagging at me. Why did they build it the way they did? Specifically, what's up with that tab sticking up on the one side? It's more trouble because it means more material to machine off to leave the raised tab, and it means the square only has a full flat on one side though that may not be a serious problem. With a simple right angle it might have been, but with a T it could be flipped either way.

One possibility struck me. Have they put a dowel pin through beam and blade right up there in that tabbed area? In the first picture one can just make out something between the two more obvious holes to either side. Given that they check squareness as they show in the photos and adjust as necessary by loosening screws, tapping to get a null indicator reading and re-tightening, having a pivot point there corresponding approximately with the lower contact point of their indicator setup would mean that if there's any deviation on the indicator they should just be able to coax it over to zero and expect to be right. No nudging by guessing how much, checking, guessing again and so forth. It's almost the equivalent of having something you can dial in right to a point.

Maybe I'm making this up, but I think not. They obviously went to some extra trouble to build it this way and there had to be a good reason for it. I'm speculating about what might have been a good reason.

darryl
02-28-2013, 07:25 PM
As I understand it, there are several additives which epoxy might have in it. For an application such as yours, you would use ones which don't have any additives. I don't know how you would determine this just by looking at a tube of it, unless there's specific details listed. You would probably find a metal filled one to be more stable, but that's just a guess. Makes sense to me, though. I use PC-7 when I want the ultimate in bond strength and workability. It has never let me down- aluminum to aluminum bond strength is not easily achieved, but PC-7 seems to do it. Does it warp, creep, or otherwise change dimensions- I don't think so.

I would not use any of the fast curing types if I wanted the ultimate in stability.

lazlo
02-28-2013, 08:18 PM
That Moore square has been nagging at me. Why did they build it the way they did? Specifically, what's up with that tab sticking up on the one side?

The whole book is a sales brochure for the Moore Jig Bore. That, along with all the other marvelous measuring devices they show, were specifically for measuring/indicating/scraping the No. 2 jig bore.

ETA, if you look on the picture on the left side of page 35, they show a slotted gauge block that's pinned in place by that stud.

TGTool
02-28-2013, 09:44 PM
The whole book is a sales brochure for the Moore Jig Bore. That, along with all the other marvelous measuring devices they show, were specifically for measuring/indicating/scraping the No. 2 jig bore.

ETA, if you look on the picture on the left side of page 35, they show a slotted gauge block that's pinned in place by that stud.

I don't quarrel with the book promoting Moore Special Tools. That's hard to miss but they have some genuinely good stuff between the covers anyway.

I'm not sure I follow what you're thinking in terms of the square with the stud and the gauges in that photograph. There's one that looks like a Cadillac Gauge column next to the square, then a lead screw, then a step block gauge. In fact, at the bottom of that page there's an inset of an electronic indicator head and the step gauge on the table of a leadscrew checking machine that's also shown on page 47. But I'm still not sure how you're picturing the square being used.

I don't have a theory on what the stud is used for, but if it engages something regularly it will be subject to rapid wear and a rectangle would have been a better choice which doesn't seem like the Moore way of doing things.

TGTool
02-28-2013, 10:08 PM
Wait, I do think I know what the stud is for. It's just to hold the blade up when you lay the thing down on a surface. It's only as long as the thickness of the beam on that side and can be removed easily if it's in the way (screw slot on the end).