Slope Curves measuring

dexta · Jun 16, 2018

Good Day
have a question for the veterans. How do you measure the geometry of a inclined slope geometry? Assume that the geometry is a oval cone to machine , and don't have a Probe ( Renishaw ,etc ) are there any tricks to see if the geometry is actually what it should be?
Tryed different aproches but it's not easy? Any idea?

Thank you kindly
Peter
Dexta

cameraman · Jun 16, 2018

dexta said:
Good Day
have a question for the veterans. How do you measure the geometry of a inclined slope geometry? Assume that the geometry is a oval cone to machine , and don't have a Probe ( Renishaw ,etc ) are there any tricks to see if the geometry is actually what it should be?
Tryed different aproches but it's not easy? Any idea?

Thank you kindly
Peter
Dexta

That's a really interesting mathematical/technical problem.

Do have prints and tolerances for said "Oval cone"... ?

How are you currently producing them or seek to produce them ?

AND... What material ?

Ta

Eric

______________________________________________________________________

P.s. How good is your math and what do have in terms of inspection slabs/ jigs and various measuring tools etc. ?

Also rough indication of budget and time might be helpful too. I.e. at what point you say "f*ck it"/ never mind :-)

dexta · Jun 16, 2018

Hi Eric.
4340, CNC 3_axis , Kitamura , AlTiN Tools. Ball most of them.
I was thinking to get a probe and design the ball size in the cad module and see it the stylus is positioned at the cordinates , it should touch.. if not positive or negative value would be the theoretical difference from the real world. My solution has to involve the machine since I cannot remove the molds/ parts until the values are exact ( avoiding reposition and cummulative errors. )
Piece in question is a glasses frame mold with conical inserts ( undercuts )
The yellow area is where the problem is.

cameraman · Jun 16, 2018

dexta said:
Hi Eric.
4340, CNC 3_axis , Kitamura , AlTiN Tools. Ball most of them.
I was thinking to get a probe and design the ball size in the cad module and see it the stylus is positioned at the cordinates , it should touch.. if not positive or negative value would be the theoretical difference from the real world. My solution has to involve the machine since I cannot remove the molds/ parts until the values are exact ( avoiding reposition and cummulative errors. )
Piece in question is a glasses frame mold with conical inserts ( undercuts )
The yellow area is where the problem is.

That's ironic I met with Jim Jannard three weeks ago (founder of Oakley).

Let me put my puzzler / under-mind on that one for a bit...

What are your xyz tolerance for any point theoretically on that surface ?

Are there features / fillets connected to that feature that have higher tolerance… Or is it more about surface finish and draft angle to perhaps "release".

I'm trying to get a better "fix" for go/no-go conditions / set of condition equations.

You mention "Slope angle" as being more critical (shrugging shoulders).

lowCountryCamo · Jun 16, 2018

You could pick coordinates from cam and use a tooling ball in a collet with feeler and check for fidelity, or haimer taster and do the same.

Vancbiker · Jun 16, 2018

Cast it with a resin and send the casting to a CMM.

Reprorubber(R)

HuFlungDung · Jun 16, 2018

Why would the machined shape be incorrect? Is your machine not qualified accurate enough to begin with? What are you going to do to fix it if something is wrong about it? If you are trying to determine if your tool is cutting to size, then you can machine a simple straight wall pocket somewhere with it (doesn't even have to be a pocket in this part) and qualify the tool diameter from that.

Ox · Jun 16, 2018

I would think that you could put your indicator at 12:00 and program a Y/Z move of the taper expected and see if you have zero run-out?

Try aggin at 3:00. (X/Z of course)

I am ass_u_ming that those points would be the apex of the oval of course.
Edit as required.

-----------------------

Think Snow Eh!
Ox

White Lightning · Jun 16, 2018

With a flat floor you can drop a ball into each end and measure x-y location on each end as long as the z depth is correct.

Sent from my XT1585 using Tapatalk

litlerob1 · Jun 16, 2018

Ox's and the guy for the Aland Islands recommendations are about as close as you can get without a CMM. But they only tell you whether what you expect to know, is right or wrong, as Hu pointed out it gives you no real data for correction. The answer is a CMM.<that is a period.

If the callout on the Print is an All Around Profile like this one, but with a box around it--you're fucked without a CMM;

White Lightning · Jun 16, 2018

litlerob1 said:
Ox's and the guy for the Aland Islands recommendations are about as close as you can get without a CMM. But they only tell you whether what you expect to know, is right or wrong, as Hu pointed out it gives you no real data for correction. The answer is a CMM.

So I guess you would rather pull the part off the machine. Wait for someone to tell you that it's wrong and set it back up to make adjustments. Unlike machining parts there really isn't tollerances in molds. It has to be right the first time. No second shots. If the tooling balls used to check the walls are in the wrong location, simple math along with the cad you are using will show how much you are off. Always shoot for metal safe and adjust if in question.

litlerob1 · Jun 17, 2018

White Lightning said:
So I guess you would rather pull the part off the machine. Wait for someone to tell you that it's wrong and set it back up to make adjustments. Unlike machining parts there really isn't tollerances in molds. It has to be right the first time. No second shots. If the tooling balls used to check the walls are in the wrong location, simple math along with the cad you are using will show how much you are off. Always shoot for metal safe and adjust if in question.

Yes, I do think you should pull the part. Tracking error is part of the game. I love it when Mould makers play off their skill as perfect. What are you going to adjust? You have 2 balls, assuming they are the same Radius of each of the Vertices (which are different Radii) and you measure the distance between, and the height of the balls once you can stick something between them, then you know that when you put a .5002" instead of .5001" gap between the balls, then the height of one of the balls needs to come up .001" and the other one needs to drop .003", and one of the blah blah blah blah. It isn't usable data. OR just use a CMM and build another part, no wonder plastic Moulds are POS.

You are in Finland so.... either you are a Sniper or a Hockey player (the only thing Finland produces BTW), and I don't trust Hockey players or Snipers to make parts.

R

White Lightning · Jun 17, 2018

litlerob1 said:
Yes, I do think you should pull the part. Tracking error is part of the game. I love it when Mould makers play off their skill as perfect. What are you going to adjust? You have 2 balls, assuming they are the same Radius of each of the Vertices (which are different Radii) and you measure the distance between, and the height of the balls once you can stick something between them, then you know that when you put a .5002" instead of .5001" gap between the balls, then the height of one of the balls needs to come up .001" and the other one needs to drop .003", and one of the blah blah blah blah. It isn't usable data. OR just use a CMM and build another part, no wonder plastic Moulds are POS.

You are in Finland so.... either you are a Sniper or a Hockey player (the only thing Finland produces BTW), and I don't trust Hockey players or Snipers to make parts.

R

I never expected you to understand how to check this 3d sloped surface. Actually you don't measure between the two balls you measure the x,y,z location of each ball. From that you can determine where the stock actually is. Yo can then adjust the program to cut it correctly. Mold makers don't pretend to be perfect they just need to hold tight tollerances on mostly 3d shapes. I'm sorry "litlerob" has a problem with my coments. He seems to jump in every thread I get involved with trying to make me look bad. I just commented on a method of checking the surface. Technology is great but knowing how to check your work as you go is a lost art. I'd rather be a sniper than a want to be bigmouth from Utah.

Sent from my XT1585 using Tapatalk

litlerob1 · Jun 17, 2018

White Lightning said:
I never expected you to understand how to check this 3d sloped surface. Actually you don't measure between the two balls you measure the x,y,z location of each ball. From that you can determine where the stock actually is. Yo can then adjust the program to cut it correctly. Mold makers don't pretend to be perfect they just need to hold tight tollerances on mostly 3d shapes. I'm sorry "litlerob" has a problem with my coments. He seems to jump in every thread I get involved with trying to make me look bad. I just commented on a method of checking the surface. Technology is great but knowing how to check your work as you go is a lost art. I'd rather be a sniper than a want to be bigmouth from Utah.

Sent from my XT1585 using Tapatalk

You're the one that said you don't really have tolerances.

If you changed your location, I would not be such a dick to you.

Plus you're a 'knowbie' which is irritating. 27 posts, you might learn something, instead of telling people they are wrong. AND if you quote me, then usually I will respond, crazy I know.

Robert, Third biggest mouth in the West, John retired so....

cameraman · Jun 17, 2018

Ox said:
I would think that you could put your indicator at 12:00 and program a Y/Z move of the taper expected and see if you have zero run-out?

Try aggin at 3:00. (X/Z of course)

I am ass_u_ming that those points would be the apex of the oval of course.
Edit as required.

-----------------------

Think Snow Eh!
Ox

I like Ox's basic idea for this … But one has to remove the wonkiness of the machine (in terms of repeatable wonkiness) as separate from slop/ backlash. Kitamura should be pretty tight / sound ? No ?

OP is asking for slope angles. So taking high and low points like Ox says can work. The Kitamura should be sound... However one should set up a sine plate or inspection surface set at the desired angles at the same height and position (in 3d space on the table) as the final surfaces of the finished part (obviously without the part on the table /fixture), then you can map out using DTI or similar (like 3d Taster as some are saying) if the machine really does travel in straight "Sloped" line in 3d space or not (or at least in one plane)Like what Ox is saying.

The idea being that if the machine has dimensional wonkiness "mashed" into it then the cut part will also display the same wonkiness so using a DTI or 3d taster won't help you as the needle won't move as the part reflects the geometric wonkiness of the machine. Using correctly positioned and angled reference surfaces means you can check your machine moves (in this instance) AND if need be calibrate your machine to delivery better metrology* (**).

___________________________________________________________________________

* Obviously CMM is best method as I don't fully believe in using "Mills" to measure with... [like a what LittleRob1 is sayin']

** Assumes that one does not have access to a CMM

___________________________________________________

Was wondering about some kind of test / gauging part that you could stick in there / insert into the cavity and blue... ?

[Havent't totally wrapped my head around the inspection balls / Gauge balls and various ellipses and projected drafts (from an imaginary point below the part) with elliptical cross sections and varying slope angles coupled with shallow depth of feature... Maybe there is a neat trick for that I'm not aware of.. Always been impressed by Vanbiker's use of inspection balls and other fixture / metrology conundra… I would have thought he would be the "Go-to" for use of metrology grade balls ;-) ].

Also like what OX has said but maybe slope angles every 30 degrees ? (Every hour around the clock face ?)

If that's all you have what are you gonna do ?

White Lightning · Jun 17, 2018

litlerob1 said:
You're the one that said you don't really have tolerances.

If you changed your location, I would not be such a dick to you.

Plus you're a 'knowbie' which is irritating. 27 posts, you might learn something, instead of telling people they are wrong. AND if you quote me, then usually I will respond, crazy I know.

Robert, Third biggest mouth in the West, John retired so....

So just because you have 2549 posts makes you more knowledgeable than someone who has 27. Wow that's true out of the box thinking. I could care less where someone is from when it comes to learning. The only reason I even joined the forum is to learn and maybe share a little bit of my experience over the last 30+ years. I'm sorry Rob I'll leave " your little soap box forum to you and your little buddy's". You win again. What a big man you are!!!!!!!!

Sent from my XT1585 using Tapatalk

dexta · Jun 17, 2018

Thank you all for the posts.. I wasn't expecting to step into a warzone..

Yes the Kita is pretty good for an old machine 0.01mm on all axis. the spindle is about 0.02 at 150cm from nose ( 6" ). Trying to compensate all of this data into the cad/CAM ..
Will keep you posted.. Cheers to all..

White Lightning · Jun 17, 2018

dexta said:
Thank you all for the posts.. I wasn't expecting to step into a warzone..
Yes the Kita is pretty good for an old machine 0.01mm on all axis. the spindle is about 0.02 at 150cm from nose ( 6" ). Trying to compensate all of this data into the cad/CAM ..
Will keep you posted.. Cheers to all..

I'm sorry for taking the thread to the gutter. Best of luck I'm sure you've got this one.

Sent from my XT1585 using Tapatalk

litlerob1 · Jun 17, 2018

White Lightning said:
So just because you have 2549 posts makes you more knowledgeable than someone who has 27. Wow that's true out of the box thinking. I could care less where someone is from when it comes to learning. The only reason I even joined the forum is to learn and maybe share a little bit of my experience over the last 30+ years. I'm sorry Rob I'll leave " your little soap box forum to you and your little buddy's". You win again. What a big man you are!!!!!!!!

Sent from my XT1585 using Tapatalk

It's not mine, and no one is asking you to leave. But you're going to have to get used to people having different opinions. Don't take anything personally. It is in our guidelines to put your location in your profile.

If you ask 20 Machinists how to do a thing, you'll get 20 different answers, right?

R

cameraman · Jun 17, 2018

dexta said:
Thank you all for the posts.. I wasn't expecting to step into a warzone..
Yes the Kita is pretty good for an old machine 0.01mm on all axis. the spindle is about 0.02 at 150cm from nose ( 6" ). Trying to compensate all of this data into the cad/CAM ..
Will keep you posted.. Cheers to all..

Not so much a "War zone" more of a "Mosh pit" lol.

As I have said in the past LittleRob1 will learn many a "Whippersnapper" the true meaning of Christmas.He's good / mighty fine.

Thoughts that occurred to me was wondering what the parameters geometrically on your CAD model are holding this all together...

I.e. how it was modelled in the first place. Like Is the elliptical cross section projected from a point from below the part ? If so what is the XYZ location for that point of projection or was it Boolean modelled in another way, or is it more free form where there is the possibility that sides of the "cone" are not straight but are parabolic or Hyperbollic (hyper-bolloxed ? .

Is the floor of that feature orthogonal to the table / XY plane? [3 axis mill]. Is the "conic" -ish projection orthogonal ? I.e. not tilted at an angle either ? Or even tilted and rotated (compound angle)… (If it is I think that means slope angles on opposing positions on the clock face will not be symmetrical . i.e. will have different slope angles if the point of projection is off axis/not orthogonal to the machine axes. [Not sure I expressed that correctly but I think you'll know what I mean]. Tilted projected elliptical cone... Or even tilted. rotated elliptical cone.

You know what you are wanting to do better than anyone else, so after a bit of playing around and a bit math I'm sure you'll close in on something that you feel gives you piece of mind in practical way...

Cheers/Good luck (you won't need it).

Slope Curves measuring

Aluminum

Diamond

Aluminum

Attachments

Diamond

Stainless

Diamond

Diamond

Diamond

Aluminum

Diamond

Aluminum

Diamond

Aluminum

Diamond

Diamond

Aluminum

Aluminum

Aluminum

Diamond

Diamond

Similar threads