Designing and building ball roundness tester

[iusedtoberich]

In my observation of an older (15yrs) set of centennial balls , the area within the black circle and # is proud of the rest of the surface.
Seems that thru polishing on my machine that the # core is harder than the white area . You can feel the difference, has any one else noticed this condition?

I have seen similar even on a brand new set. I've also seen the opposite. I bought a set of "new old stock" German Rashig balls. The black lines for the number were sunken in a mile.

That's actually one of the reasons I wanted to make this fixture. Its a relatively easy way to measure those differences on specific spots by simply rotating the ball a hair and see how the needle moves between the white of the ball and the strip (or number, or line) of the ball.

 

[iusedtoberich]

You could measure diameter if you had a test standard ball, say a ground steel ball bearing that is 2.250 then you could measure plus or minus on your test indicator. You could also use this to test how well your fixture works

Or just put them in a CMM machine

I actually did that at the end of the second video. It doesn't work because the test ball rests on 3 points, and none of those points are opposite the indicator tip. So there is not a true diameter reading. Its similar to a cosine error.

 

[iusedtoberich]

Will these roundness test be used by players to declare faulty equipment at matches?

Or by Sanctioning bodies to implement strict regulations on tournament quality billiard ball sets?

The answers to the above will decide how much participation I can volunteer.

No, none of that. This is purely me nerding out, on two topics that interest me: pool and machining. Has zero real world application, unless other fellow pool nerds are like minded. It would never be used in a tournament or by a sanctioning body.

 

[Black-Balled]

No, none of that. This is purely me nerding out, on two topics that interest me: pool and machining. Has zero real world application, unless other fellow pool nerds are like minded. It would never be used in a tournament or by a sanctioning body.

You are a kind soul to humor his thoughtless question.

Literally there is nothing more fair than opponents using the same ball set.

Allowing one to appeal the validity of a contest on such trivial measures would open up a world of potential MINOR deviations to scrutiny.

 

[fastone371]

I'm working on a ball sphericity (roundness) tester for fun. Below are a couple of pictures, and a YouTube video link describing the process. I'll update as I go.

My motivation is to see how different ball sets and brands compare both when new, and when worn. Also because I like making things.

View attachment 556942

View attachment 556943

Youtube video with a lot more detail. Part 1:

I dont know about this application but when I use a dial indicator in automotive applications and I have to rotate/move the object that I am measuring my indicators give erratic readings. Even if its its only a couple thousandths thats probably about the tolerance I would expect to see ball to ball.

 

[iusedtoberich]

I dont know about this application but when I use a dial indicator in automotive applications and I have to rotate/move the object that I am measuring my indicators give erratic readings. Even if its its only a couple thousandths thats probably about the tolerance I would expect to see ball to ball.

Yes, I ran into that with the 3D printed base. Its definitely a consideration if I take it to the next level and build a tester out of metal.

 

[3kushn]

I didn't read every word in this thread. Paid attention to Bob Jewett and you mostly.

I guess you know some of the variables, but just in case, I can't help myself.
Sort of a nerd too. Maybe more than sort of.

Maybe you have all this under control, but it seems to me that your base (3 balls) needs to be ultra machined to perfection. The spheres that sit in the pockets need to be 5X tighter in spec than pool balls, the pockets? IDK, maybe 5X better than the supporting balls. If they move? For Sure. If the resting bearings don't spin, then you need to look at chatter/friction. Each position may not be the same as the last. Can't assume anything. Lubricate.

How do you Center the dial indicator? What instrument tells you you're on center, 100%. I suppose center isn't absolutely critical but ...

Used to be into torque. Some know but I'll bore you again a minute.
Gotta torque wrench..... Is it accurate/repeatable?
Test it on a transducer and loader. OK Cool
Is the transducer accurate/repeatable .
Test the transducer against physical weights, adjusted to your geo position, temp and Humidity.

Test the Arm??
Ask NIST or NIST 17025 Certified Lab

We don't need NIST Certified Perfect Spheres. Heck, as soon as you play, they might be out of spec.
First hit?
DONE, OUT OF SPEC!!!

Not the point. Hope you get it.

 

[Bob Jewett]

... For balance you could just float the ball on Mercury.

I tried a mercury float. I got a fluid ounce of mercury which weighed almost a pound. It didn't work well. I think surface tension and oxidation were the problems.

I think a rolling test is going to be the best. It's what you really want to know about. A long, very gentle ramp and try the ball in various orientations.

 

[Black-Balled]

I tried a mercury float. I got a fluid ounce of mercury which weighed almost a pound. It didn't work well. I think surface tension and oxidation were the problems.

I think a rolling test is going to be the best. It's what you really want to know about. A long, very gentle ramp and try the ball in various orientations.

Caveman.

 

[iusedtoberich]

I upped my game courtesy of eBay. $20 for the indicator and $38 for the stand! I have 2 more stands coming and 2 more indicators, that are even finer resolution. This works well though. It easily shows .0015” variation of the ball when it turns.

 

[iusedtoberich]

And here is the same stand with a new indicator magnetic holder I bought that I've been eyeing up for a couple years. Paired with another old Ebay special .0001" indicator that was about $25. No one wants these things anymore. They still work.

 

[iusedtoberich]

And here is a modern top of the line indicator on the same setup. The arm was not long enough for this one, but its close enough to test. I bought this one for my machining a couple years ago.

 

[iusedtoberich]

In conclusion, I think this set of 3 different setups shows that .0001" reading dial test indicators, and .0001" reading plunge indicators, are adequate to read the variations within the same ball. They each showed about .0015" total indicator reading on the same ball. .001" indicators are not fine enough resolution for this job.

The problem with all these setups is you can't get the ball out unless you move the whole indicator setup. Its workable, but not ideal if you wanted to quickly test many balls. I can work on a solution for that later.

The other indicators I have coming from Ebay are even finer resolution.

Once I figure out which setup I want to use, I will make a platen with the steel ball bearings again to fit it.

I may also consider building a couple of these for some of the ball nuts on here if anyone really wants one. I think I can make a design that can be Waterjet cut out of a plate of steel, with 3 holes for the ball bearings, and then a magnetic indicator base can be used. That might be the cheapest and easiest way to go to build a few, without stalking Ebay.

 

[Bob Jewett]

..., and then a magnetic indicator base can be used. ...

I assume this is so the holder can slide out with the ball rather than needing to lift the ball out against the indicator.

Have you looked at laser indicators? That ought to get you down to millionths. Well, I looked and found this micrometer. They can get down to the micron range. May be a little expensive. https://www.keyence.com/products/measure/micrometer/ls-7000/specs/

 

[kling&allen]

Great work as always. Did you machine the white base?

 

[DaveK]

How did I miss this earlier ? Good to be here now ...

A device used by opticians /lens makers is a spherometer

Spherometer - Wikipedia

en.wikipedia.org

The wikipedia shows old manual ones. I am in the process of making one something like this :

DIY Spherometer by BoByS

So...you are crazy enough to build your own telescope mirror or other optics? Well, you'll definitely need a spherometer! This design has 2 configurations, based on the central hole - 8mm and 8.3mm. When I print a hole, my printer has an accuracy of about 0.3mm so in my case, 8.3mm is actually...

www.thingiverse.com

With a tenths-reading dial indicator (easier to see movement on a dial than digits) and a surface plate to zero it out you can measure curvature quite easily assuming you know the exact spacing of the three points (and of course consistency of curvature or roundness without having a clue as to the point spacing ).

iusedtoberich your indicator stand is in the same concept, but a spherometer can be used maybe a little easier.

Dave

 

[iusedtoberich]

I assume this is so the holder can slide out with the ball rather than needing to lift the ball out against the indicator.

Have you looked at laser indicators? That ought to get you down to millionths. Well, I looked and found this micrometer. They can get down to the micron range. May be a little expensive. https://www.keyence.com/products/measure/micrometer/ls-7000/specs/

The idea for using a plate of steel and a magnetic indicator base, is so I can waterjet the steel (no machining needed), and buy an off-the-shelf indicator base that has a fine adjustment. Less effort on my part. The only machining in that scenario would be to put chamfers on the holes for the 3 ball bearings, so they have a nice surface to sit on. They would then be epoxied in place with JB-weld. If the plate was not perfectly flat (it would not be), as long as the indicator base did not rock back and forth, it would not affect the accuracy.

The alternative to this, which is the way I was leaning before, was to scratch build a plate out of aluminum (I can machine it easier than steel), that had a pivoting arm the indicator attached to, and the pivot of the arm would be incorporated into the plate. I may still go this way, because it would be more fun to design and build. With a custom arm, I maybe be able to figure out a way to have it get out the way so the ball can be removed and replaced easily.

On that note, the new magnetic indicator in the picture (noga brand), its fine adjustment screw actually had enough range to get the ball out easily. It just takes 2 seconds to turn the screw enough. I wasn't expecting that.

No lasers for me, that is way above my pay grade

 

[iusedtoberich]

I had also thought of using 123 or 246 blocks, along with the magnetic base. They already have the holes, and are perfectly flat. But I doubt the holes will be where I want them. I have a few in the office I will check. If that worked though, there would be nothing to machine. Just off the shelf 123 blocks, and a magnetic base, and an indicator.

 

[iusedtoberich]

Great work as always. Did you machine the white base?

Thanks. No, the white base is 3D printed. I drew it up last night to be able to bolt it to the indicator stand that was delivered and try a few balls. I was excited to try it. It works surprisingly well. I can repeat to about .0002". It deforms I press hard into it, but if I'm gentle when I spin the ball, it works really well. The real one would be metal, either aluminum or steel, and have steel ball bearings epoxied into it.

 

[iusedtoberich]

How did I miss this earlier ? Good to be here now ...

A device used by opticians /lens makers is a spherometer

Spherometer - Wikipedia

en.wikipedia.org

The wikipedia shows old manual ones. I am in the process of making one something like this :

DIY Spherometer by BoByS

So...you are crazy enough to build your own telescope mirror or other optics? Well, you'll definitely need a spherometer! This design has 2 configurations, based on the central hole - 8mm and 8.3mm. When I print a hole, my printer has an accuracy of about 0.3mm so in my case, 8.3mm is actually...

www.thingiverse.com

With a tenths-reading dial indicator (easier to see movement on a dial than digits) and a surface plate to zero it out you can measure curvature quite easily assuming you know the exact spacing of the three points (and of course consistency of curvature or roundness without having a clue as to the point spacing ).

iusedtoberich your indicator stand is in the same concept, but a spherometer can be used maybe a little easier.

Dave

Yeah, very similar. The last time I was at my optician I had him show me his lens curvature measuring tool. His was 2D only, not 3D. It had two contact points, with the plunger of the indicator between them. What is the scale you are working in? Will your device be something like the size of a hand, the size of a foot, or bigger still?