scraping an angle plate

what is the standart procedure to scrape a 90° surface? do you put it on the surface plate and blue it up with a master square (angle plate)? what if you cant blue it up, because you dont have a square or its much smaller that what you want to scrape? do you check it with a beveled steel square? and what if that is not long enough? or do you check it with an indicator, that you have set with something that is square?

Not speaking from much experience, but when I did my cast iron square from Martin Models, I scraped one side flat, then used my Brown and Sharpe 20" try-square (machinist square) to do the other side... I scraped the other side rough flat... Then I put it against the try square and step scraped until it looked good...

I did read I should have put slips of paper between the try square and the cast iron square, and then pulled them out noting the difference pull pressure required...

In any case as I planned all along the cast square went to Melbourne in 2011 and I scraped it against a known master square.. It turned out I got it pretty close, and I had it very good against the true master square in about half an hour...

Speaking with zero experience. The "indicator set to something square" sounds like a winner.

Edit:-
This kinda style.
Super simple.

Ideally you need something like a master square, that is flat and square to blue against.

But, what if you don't have a master square, and you know how to measure square with a dti, then you could shim another angle plate to square and blue from that. When you get too close for shims, you can step grind on a surface grinder.. that is grind a few microns, but leave one edge.. which is equivalent to having a precise micron sized shim.

Square is one of those self checking measurements by reversal, but like flat surfaces, you need three squares to agree with each other before you can be sure.

Ray

Funny this should come up now, as was doing this this morning as a matched pair of angle plates. I was going to put it up but figured compared to what most guys do here my feeble efforts didn't account for much so didn't bother.

In fact you don't need 3 angle plates to self check, you only need two and a dial indicator with stand. I bought two cheap Chinese ones just to get the castings and considered them nothing more than rough castings. Rather remarkably however they were actually quite square straight out of the box. Shock of horrors as I haven't come to expect much from this type of thing in the past. Anyway, to get accurate 90 degree angles the process is quite straight forward and self generating.

Basically what you do is blue one and use that as a master against the second one. Scrape until the two plates have full bearing. However this doesn't tell you anything about whether the angle plates may or may not be square. Since they're angle plates it's not possible to check them as cubes, and difficult to fixture such as you may check a square with a dial indicator. I haven't looked at the above video but presume he's using the indicator on a stand to check around a block. In this case, do the same thing but instead of checking one side of a cube against the other side, compare one angle plate against the other. When both angle plates indicate zero with respect to each other, have full bearing against each other and to the surface plate, they're square to the accuracy that you choose to indicate them.

Hopefully that makes sense. To simplify things (as I'm a simple guy) I explained it as if you do each sequentially, but in fact you're indicating and scraping to each other at the same time.

Pete

My solution was to do 3 together in the classic manner. I followed the sequence of checking one against the next, scraping to one and and proceeding to the third... I wanted one, so the other two were just "L" angles...Basically manufacturing fixtures for the "real " one. The one I wanted I did with a solid web.

I added a flatness check against the granite flat periodically, so that I didn't have to scrape the 3 for flatness and angle simultaneously, which shortened the job in terms of number of cycles total. I was able to do the two "fixture" pieces with just an area top and bottom, since they only needed to set the angle, flatness came from the granite flat.

Luckily these were small ones.... about 4 or 5 inches, so it didn't take *that* long.

Be aware that since they can't be rotated, if you just scrape 3 angle plates to each other they will also need to be checked flat against the surface plate itself.

This is common problem to address with many ways one can proceed. I like using an indicator and when setup as shown it indicates twice the error :-)http://www.practicalmachinist.com/v...oree-nw-2014-a-275498/index6.html#post2232969
Cheers Don

There are other ways you can use your steel square to check the square and not use the papers. The paper method is a good method but you have to guess at the feel. What you can do is to hold the base of the square against the angle block and gently clamp it to it. be sure the angle block is not high in the middle. Have the thin side (blade) of square pointing up.

Then take a surface gage and .0001" indicator and check the length of the blade on the top and I usually take a Sharpie marker and write on the side of the blade the reading. Then lower the indicator and check the bottom of the blade. If the square is accurate the top and bottom should be the same. if it is say plus on top .001 on far end the bottom of that same end should be minus .001. You need to use you math too. if the CE angle block is 12" high and your square is 6" long then the reading would be double the error. I take my blade square to customers shops and that have a CMM and have them test my squares. I also use the method the you tube guy shows by setting the surface gage on my granite square first and then check my angle block I am scraping. I also have a magnetic cylinder square I put on the angle block and use a surface gage and .0001" indicator. When you use a indicator you know exactly what it is and not a guess the paper method is.

I will search my pictures and the net for photo's and add them.I would bet Jeff T or Robert G has pictures of this as I show all these methods in the classes.

This is why I say "It is easy to learn scrape, but know where and how much to scrape off is a Trade" or the secret of success of becoming a rebuilder. Having the right tools too....;-)

Good Day. Rich

PS: Don was typing at same time....lol...he hit enter first.

What the Tool & Die Guy show in his video is a perfectly good procedure, but what he doesn't mention is an inbuilt assumption he's making right at the start. That is that the two sides are parallel. Since he's done this on the grinder it's generally a good assumption but one can get caught short. Suppose, for instance, that the part is pyramid shaped, the same angle on all four sides. As he checks with his indicator it would suggest that the block is very square. In fact it only shows that the two opposite sides he indicates are equal. The error will be evident when he flips it over, but if the top and bottom aren't parallel he'll get all kinds of wonky readings.

So before his check is valid, he should have checked parallel on the opposite sides and then checked for square. The test Don shows doesn't have this liability since it's checking only one side, front and back as it were, and will be accurate.

This is the way stuff gets passed around the net as gospel and can cause all kinds of confusion if the listeners don't really think through and understand the basic geometry. He's right that if you understand basic squareness you're good, but then he doesn't go ahead to educate completely.

GMTA Bebop :-)

Richard King said:

GMTA Bebop :-)

Click to expand...

why thank you kind sir
Cheers Don

I was thinking of when I was a kid...20 I think...and i was helping a Millwright set some lathes in a new shop.....he showed me something that is real handy, but not accurate to tenths, but is pretty cool and I have used it several times to get things close to square. We plotted a place on the floor and drew caulk lines on the floor by using 3 tape measures: Measured out 3' with one tape, 4' with another and 5' with another to find squareness. I have used 6, 8 and 10 too. It works with inches too...try it on some graph paper. :-) Rich

TGTool said:

What the Tool & Die Guy show in his video is a perfectly good procedure, but what he doesn't mention is an inbuilt assumption he's making right at the start. That is that the two sides are parallel. Since he's done this on the grinder it's generally a good assumption but one can get caught short. Suppose, for instance, that the part is pyramid shaped, the same angle on all four sides. As he checks with his indicator it would suggest that the block is very square. In fact it only shows that the two opposite sides he indicates are equal. The error will be evident when he flips it over, but if the top and bottom aren't parallel he's get all kinds of wonky readings.

So before his check is valid, he should have checked parallel on the opposite sides and then checked for square. The test Don shows doesn't have this liability since it's checking only one side, front and back as it were, and will be accurate.

This is the way stuff gets passed around the net as gospel and can cause all kinds of confusion if the listeners don't really think through and understand the basic geometry. He's right that if you understand basic squareness you're good, but then he doesn't go ahead to educate completely.

Click to expand...

The main point I was trying get across was the indicator set up itself. I understand that block has six sides unlike most of the guys that commented on that video . There was another video that Phil did to address this by flipping the cube on a known wonky block, but i cant find it on his channel
Bebops soloution is even more simple, same indicator setup with no master required. Perfic!

I've made three small 3x3x3" squares using the surface plate for flatness and the comparing the three for tilt. I did it for the exercise/technique rather than any real need. Later I made four 30-60-90° squares with the same techniques because I actually needed at least three 60° blocks so that I could jig my HLV bed plate on the surface grinder to re-grind the dovetails.

Square is one of those self checking measurements by reversal, but like flat surfaces, you need three squares to agree with each other before you can be sure.

Click to expand...

No, you don't need three surfaces. You only need a square and verify it. And after that, your verification instrument is as accurate as you want.
Verifying a square is not so complicated.

But first, standards allow quite a huge (I don't have the numbers here right now) deviation from true squareness. It is almost shocking. So my square standard is G00.

So how do you make a perfect square?
We need a gizmo. It looks like a magnetic stand. Just with a column quite beefy. And a rack along it. The clamp that holds the horizontal bar needs to have a perfect fit and a gear that meshes with the rack. So with a knob, you can drive a DI up/down. And a beefy foot. I have one of those from Mahr. Just paid 20 €-

Next step is, how to verify wether the column is dead at 90°:
Take a square, the precision almost doesn't matter. Put that side to side with the rack-gizmo-thing, so the horizontal legs both point in the same direction.
Crank the rack-gizmo-thing down and mount a DI sticking sideways and having contact with the angles vertical face. Zero. Crank up and note the difference.
Rotate the rack-gizmo 180°, so the legs point one left, the other right. Crank up again and note the difference. Half the difference is how much the gizmo is out of square. Adjust, repeat, until satisfied.

Now you have a rack-gizmo that is dead 90° and you can use that to:
a) adjust your square (by peen hammering)
a) check your CI square.


Maybe I need to make pictures …


Nick

I always wonder about this device

Squareness measurement - YouTube

I always wonder about this device

Click to expand...

That is the gizmo, just lacking the rack. Now they have to sweep it to get the lowest reading.


Nick

I've got the cheap and dirty version of that. I purchased a cylindrical square and it came with a little extra piece. It's essentially a wide angle V-block with an indicator mount on the back side. On this one, aluminum probably, they've also installed a couple magnets and the contact points are ground pads but a V-block with two planar surfaces is all that's required.





It's not quite as facile as the one you linked to, but the use is the same. Just bring the cylinder close enough to the workpiece to indicate as you sweep back and forth. Zero the indicator, then slide up the cylinder and sweep a different height. The two readings will compare the workpiece with the cylinder. It seems simple enough to make with magnets not strictly necessary.

RC99 said:

I always wonder about this device

Squareness measurement - YouTube

Click to expand...

Silly thought, but can't you just measure perpendicularity of a master square in your bridgeport (or any other mill) with an indicol*? Obviously it works best/is easiest for squares with a (short) leg of <5" (or whatever your z travel is). Clamp the square on its side in a vise, indicate the side parallel to z (for repeated checking, set a stop). Then use an indicol to sweep the perpendicular surface of the square. No need for fancy equipment or custom fixtures. Obvious due diligence about machine condition and not distorting the square in the vise is required. Heck, you could even use a lathe to do a similar setup with the square on the carriage.

*I've never tried this, having never thought of it. Next time though. I've always used an indicator on a surface gauge with zero set off a master square. The master square was verified by establishing blade parallelism and then using an indicator setup similar to the first video.

You can always just buy or make a cylindrical square as a master for comparing with other squares.