How to measure "square" for scraping? Angle plate

[PekkaNF]

Hello,

I'm going to have a go scraping and need to proceed on baby steps, because ultmate goal is to get my old and tired milling machine closer to acpetable level than it is. Not perfect of über accurate, but closer to HMS standard.

I have a small granite surface plate and standard equuipment. So far I have used good known square and geeler gages to check other square pieces. I have also used a DTI and parallel to indicate dips and walleys.

But how do I check error on two planes that are supposed to be square to ecah others. I can spot each plane indvidually and I now that both of the planes are about 0,02 mm concave in the middle....but how to indicate which corner to scrape more to bring both planes closer to square?

Is there any other way than:
1. true DTI on master square
2. get comparison readings from the angle plate (plane that is expected to be most off)
3. Scrape offending corner down
4. Spot of flatnes and race for next offending bigger spot
5. Indicate again and cycle until "square" enough
6. Scrape flat

I don't have a master square, I have two pretty good flat squares.

This is probably a daft question, but I had pieces on small granite surface plate and was wondering "how on earth I make it sure that I know which way these planes are skewed without making assumptions I can't verify?".

Pekka

[PekkaNF]

OK, found better search word for google and found this on YouTube:

How To Check A Block For Squareness

Modified surface gauge and DTI. Looks like a project.

Pekka

[sparky961]

Someone posted a link to this a little while ago, maybe it was even you - but I don't recall.  Here is is again:

http://www.totallyscrewedmachineshop.com/documents/FoundationsofMechanicalAccuracy.pdf

My take is that this book starts to become a bit of an advertisement for their measuring machine as the book goes on.   However, along the way they do provide some very useful tips for achieving a high degree of accuracy with relatively simple tools, lots of time, and even more patience.

I have little experience in this area, but have decided that you can't really "measure" square.  You can only compare it.  Perhaps this is an over-simplification, but that's the way I've been looking at it.  Have you used a "squaring gauge" before?

My thought is that you can zero the gauge on the most square item you have; be it your angle plate, square, or cylinder square.  I like the idea of the cylinder square because you can verify the accuracy by checking the consistency of the diameter, finding one position that is truest to square, then turning it 180 degrees to verify you get the same reading with your squaring gauge.

After you know the gauge is set to read "0" when two surfaces are truly square, you use that to check your machine ways and such.  Unfortunately you can't change the height of the gauge as you're measuring a surface.  Instead, you'd have to mark off the height increments of interest and go back and forth with the gauge to check the various heights.  Record the data and what you should end up with is a map of the highs and lows, indicating where you should be concentrating your scraping efforts.

A bit tedious, but you didn't expect this to be an evening project, did you?

(Edit: Seems you beat me to the post...)

[sparky961]

Here are some pictures of the gauge I made from scratch.  I have a SolidWorks model and dimensioned drawings that I'll share if anyone's interested.  I haven't made the clamp for holding an indicator yet, and have just been using another one that's offset a bit too much for my liking.

[Manxmodder]

Sparky,I really like the look of the gauge stand you've made. I'd be very gratefull to see some dimensioned drawings if it's not too much trouble.

Pekka,I'm following along and watching here as I'm very interested in how you resolve this problem.

Any chance you could post some images to get a better idea of exactly what you're looking at?....OZ

[PekkaNF]

Thanks Sparky,

That book is new to me and pretty exhaustive, I paged it trough and it looks very interesting. Although most of the methods and equipment is out of my reach.

That stand looks nice, how does it works?

Pekka

[DavidA]

I understand that somewhere, in a pleasant, tranquil part of the country, is a big house that holds (gently, but quite securely), Madmodders who thought (and still think) that they can get the accuracy of their lathe, milling machine etc, 'just a little closer' to 'spot on.

They drift off to sleep in the early hours but wake up suddenly muttering' Ah, now I see how to do it'.

Poor souls. There is nothing we can do for them.

Dave

[sparky961]

Sparky,I really like the look of the gauge stand you've made. I'd be very gratefull to see some dimensioned drawings if it's not too much trouble.

New thread with gauge drawings and details here: http://madmodder.net/index.php/topic,11054.0.html

[PekkaNF]

Sparky,I really like the look of the gauge stand you've made. I'd be very gratefull to see some dimensioned drawings if it's not too much trouble.

New thread with gauge drawings and details here: http://madmodder.net/index.php/topic,11054.0.html

Thank you for sharing it.

I have been reading about this method and talked yesterday to one mechanical engineer about this method. We concluded following:

1) To be sure DTI needs to be zeroed to a good standard, cylindrical square would be nice.

2) Arch a the toe of stand is not critical, but it must be of continuos curve to high degree. Also large radii is preferable, small one is a bit more critical (smaller radii than stylys radii is a bit counterproductive).

I don't have any other type of base than magnetic bases. I haven't got any threepoint or flat base that would fit to my price range.

I'm thinkking of building one.

1) Base: Shape and fabrication?
* round with thin rim? lapping flat
* Trilobular with balls on the corners? Only front with continuous smooth curve?
* Square, thin lapped sides, large radii on front?

2) Arm. I sort of like comparator type heavy upright pole.
* Would be nice to have it fairly straight, but I can't see why it MUST be nearly perpeticular, if the DTI is rigidy mounted.

Other way would be to make it all square and sweep DTI instead of the base, but that would make sense only if the contraption would be BIG.

What do you think?

Pekka

[sparky961]

I think you're on the right track, sort of, with pursuing the fabrication or purchase of a cylinder square.  I say "sort of" because I think you might be taking it to a level of precision slightly beyond what's required.  I think a lathe that's painstakingly set to cut with taper less than you can measure with your most sensitive mic (so likely within 0.0001) is more than acceptable.  Get a nice fine finish on it and check along the length for consistency.  Unless you're doing something way beyond home shop stuff, or even precision fabrication, I think lapping is a bit excessive and may cause you to get stuck in the process instead of producing results.

Beyond my additional comments in the other thread, the squaring gauge I built has proved to work very well.  I've reproduced measurements to fractions of a 10th interpolating between graduations on my 0.0001" DTI.

Many, if not most quality angle plates I've seen are square within 0.0002 over 6".  A few I've checked are dead on to the best of my visual discrimination.

The only advantage a cylinder square has, in my opinion, is that it is self-proving.  Even if the base isn't perpendicular (they make them on purpose like this too) or there's a slight taper, it is still a functional tool.  You just have to be careful where around the circumference your line contact is, and know the exact difference in diameter where your gauge contacts it so you can mathematically adjust.

[PekkaNF]

That is very correct. Sometimes I'll get too much into the process itself, because I want to learn stuff before I actually (and often desparately) need it.

Angular error is a killer, as you note it's possible to reach good accuracy, but it is much of a mental process as well. I had very small accurate square and all shop squares tested differently. It was a pain to work with them and I know that there are ways to true them, but they were different sizes and different models. I needed a standard I could trust to some degree. Threrefore I bought two bigger and identical pretty accurate squares to DIN standard and class #1. I tried them and they are nearly identical

I can tested this angle plate with them and surface plate using feeler blades. I can "feel" the hills/valleys, but can't really put number on it. Here it is not much, but for rudimentary "mapping" it would be nice to get a value.

My lathe is not that accurate, it needs adjustment during turning, therefore I need to figure out which parts to make and in which order to get acceptable results. I have noticed that as soon as I move tail stock on the ways, it moves sideways too a little unpredictable way. When I know more about this subject I will fix the actual problem, but I'm not too keen on much of the quick/dirty methods (like much of the "in situ"  lapping on internet).

Pekka

 

[PekkaNF]

I digged out my souvenirs from last trip to USA: Brayer and Dykem spotting blue.

Dykem tube had seeped out a little even before I opened it. It was a bit lumpy and harder to work that Stuart mictometer blue. Then again I don't know much this stuff.

This time I used cast iron surface plate, because:
1) Now I can keep it clean and I know it is good
2) Cast iron palte is bigger and it apears to be good, but I haven't check it all over yet.

The suspect angle plate does not rock, but it does not feel "tight" against the surface plate.

I checked it with with square and feeler blade, can feel it is not all straight and it feels like it has a dip in the middle both sides.

Then I spotted it twice and got similar results, although I'm not confident about it, I clearly can feel and see that spotting needs some practice. I'll lower it down on it's fulcrum then lay it flat and move 3" back and fro.

I'm not sure how much pressure to apply.

Also I need to think which high spot to start. Not onlyt to get it flat, but also square.

Pekka

[chipenter]

It's own weight will be enugh press to hard and you can bend it , to get it square a cylindrical square and a back light will work .

[sparky961]

<snip> I have noticed that as soon as I move tail stock on the ways, it moves sideways too a little unpredictable way. <snip>

You should be able to turn the outside of a cylinder square without moving the tailstock on the ways.  Set it up with a live center in the tailstock, or slow speed with a greased dead center if that's all you have.  You could set it up between centers but you still need to drive it somehow.  Better would be to hold some extra length in a 3 or 4-jaw that you'll discard later.  Then take successive cuts dialing in your tailstock alignment using the adjustment screws your lathe hopefully has. 

You may have to loosen the chuck jaws while dialing in the tailstock to let the work move around.  But once you get the diameter consistent along the length you don't want to do anything except take your finish cut.  I also suggest letting it cool to room temperature and measuring before your finish cut and before you take it off the machine.

Depending on what material you have access to, starting with CRS (cold rolled steel) or TGP (turned, ground & polished) will let you get things really close with an indicator before even taking the first cut.

[Pete.]

You don't need a large micrometer to measure the cylinder either if you're making a larger diameter one. Take a normal scribing block/surface gauge and loosen the clamp, then push the rod through so that it projects downwards instead of up. Now set up a tenths-reading DTI on that rod to read across the diameter of your part with both the base and the rod touching on the OD and slide it along to check for taper.

Don't forget to break the corners of the cylinder to help prevent burrs being raised in handling.