GD&T to Control 4X flats on a lathe tool holder sleeve

[Cncgotoguy]

I'm drafting a sleeve for a screw machine tool. I'm having a little trouble getting the dimensions and tolerances for the four flats around the outside.

I set the primary datum -A- to the tightly held outside diameter which will be sliding into the tool block on the lathe. This will hold the axis of the tool holder to the axis of the part when it is in the spindle of the lathe

In the tool block, there is a set screw to stop rotation of the tool holder and to orient the tool holder, so a tool tip will be correctly oriented with the center of the part. This is critical for lathe tooling such as boring bars, threading tools, etc...
I put flats on four sides for versatility.

Finally, I will be inserting a boring bar into the tool holder, and I need to keep it located to the flats on the tool holder which keep the orientation to the lathe tool block.

Apologize for the long back story,

The amount of flat on the cylinder -A- isn't too critical, but they do need to be very parallel to the axis of the cylinder -A- and the angle relative to each other needs to be held tightly to 90°. I want these flats in their pattern to be my datum -B-, so I can dimension and tolerance my 8-32 threads to these flats.

Some advice on calling out the flats according to the design intent and form, fit, and function would be much appreciated.

I am attaching the current drawing and a picture of the model.


Thanks

 

[CheckerHater]

Something like this.

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[greenimi]

I would say that CF symbol should be added to the outisde diameter of Ø.9835 to lock the 2 diameter together (or even you can use position of zero at MMC with no datum reference frame)

If you don't have the diametrical symbol Ø on the position callout for the 8-32 thread (lower segment) because you want more tolerance on one direction (I suspect more tolerance on horizontal direction than the vertical one, but I might be wrong here) I think you should allign the position callout (lower segment) with the direction (horizontal)

 

[CheckerHater]

@greenimi: Where do you see two diameters .9835?

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[greenimi]

CH,
So you are saying that 2X in front of the Ø.9835 is a mistake?
If yes, than I agree with you.

 

[CheckerHater]

I'd say so. There is clearly one single feature with that diameter.

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[chez311]

CH,

I am wondering about the controls on the 4x flats. It was stated that the orientation of each surface relative to each other (90deg) is very critical - do you think the position control on the 2x width (midplane) is sufficient or perhaps an orientation refinement is necessary? I mean I understand the size/position tolerance could be set tight enough so it did control it to OP's requirements, however it was already stated the actual size/depth(relative to the OD) is not terribly critical, only their orientation. I was thinking maybe arbitrarily setting one of the flats as a datum feature and holding the other three to a perpendicularity/parallelism requirement to avoid self-referencing, however it is perhaps not ideal because you could potentially get 4x different results depending on how its measured because they are all symmetric (ie: not unique/identifiable - any of the 4x flats could be set as the datum feature).

 

[CheckerHater]

@chez311: As flats are controlled by position tolerance they basic 90 deg between the flats is implied.

Unfortunately OP did not provide actual numbers, how tight flats have to be sized and how tight they have to be oriented, so it is not clear if further fine tuning is required

"For every expert there is an equal and opposite expert"
Arthur C. Clarke Profiles of the future

 

[Cncgotoguy]

Thanks for the replies!

I agree with the mistake, there is only one diameter. I will fix that.

The flats are just for a set screw to orient the tool holder at 90° increments to the lathe (the 8-32 threads that hold the turning tool orientation to the flats). This is very important. If the tool is rotated, offset changes won't correlate to changes in diameter of the turned feature. There would be cosine error.

The center plane of the flats does not matter at all and I want to make sure through dimensions and tolerances that this is conveyed. All I need is at least enough flat to have a good bearing area for the M6 set screw from the lathe.

I was throwing around the idea of using one flat as a datum and holding the other's to this, but I was hoping there was a control that does all of this at once.

The location of the 8-32 threads need to be perpendicular to the flats and along the axis of A

I'm attaching the updated drawing with the recommendation of CheckerHater, one note on the Position tolerance you put on the flats. I believe the only time there is a diameter symbol before the tolerance is when the feature is circular. Since this is a plane, we would leave the diameter symbol out of the feature control frame.

With this tolerance for the flats, I'm imagining a plane floating in space and the position tolerance holds it perpendicular to the axis of datum A only. I think it is the right control.

 

[greenimi]

Never connect/attache datum feature symbol to a centerline or centerplane....
Make the dimension between the flats and not between the centerplane and one flat.

 

[Cncgotoguy]

Agree with that. Fixed and attached.

 

[greenimi]

I want these flats in their pattern to be my datum -B-, so I can dimension and tolerance my 8-32 threads to these flats

Is datum B driven from the pair of two opposing flats or is datum B driven from one flat?

Why on the lower segment of the positional callout for 8-32 UNC-2B thread a diametric symbol (Ø) has not been added?

 

[Cncgotoguy]

Nice catch, I fixed it. Attached a new drawing.

I want datum B to be driven from all four flats as a pattern and keep them oriented together tightly and perpendicular to -A- tightly as well. but the center plane of them or the size doesn't matter to me. Then when I tolerance the 8-32 holes, I don't want them to be toleranced tightly to the center plane of any of the flats. The 8-32 holes need to be held perp to the flats and along the axis of -A-. The distance from the end isn't too important, because the flat on the tool is pretty long.

 

[chez311]

cncgotoguy,

Even in your latest revision, you still have the dimension on the flats from the centerline (0.442) - since this is no longer a width, they are no longer a Feature of Size and therefore cannot be controlled with position. You must dimension these across the width (0.884 - a FOS must have opposed features) as CH had in his mock-up in order to utilize position. You could utilize a patterned profile control however even with a multiple single segment or composite tolerance this will hold the width to an even tighter requirement (as it must have basic dimensions), even if it allows it to float around the central axis.

Do you know how much angular error you can tolerate for the flats in relation to each other?

greenimi is also correct - you have changed the way your datum B is defined now so that a single flat is your clocking datum instead of the entire pattern. If you only care about orientation of each set of 2x #8-32 holes to its corresponding flat you could utilize a "4X INDIVIDUALLY" notation similar to 7-26 or 7-37, however you just have to make sure it is extremely clear as to how that special note is applied to the features/datums. You could utilize a detail view with 4X notation like 7-37 showing a flat and the 2x holes.

 

[chez311]

After reading through the comments again, I am unsure if my suggestion about "INDIVIDUALLY" notation is helpful or harmful. I guess it depends on what is more important, that all 8x holes are held to a pattern tightly controlled perpendicular to each other (may allow slightly more angular error relative to each individual flat), or each set of 2x holes held tightly controlled perpendicular to each flat (may allow slightly more angular error relative to the other sets of holes). My guess is the first one, so "INDIVIDUALLY" notation may not be useful in this case.

 

[Cncgotoguy]

Chez311, for sure, can't have a position tolerance without out FOS, I was kicking around the idea of holding the surfaces of the flats with a profile callout instead. That way, I wouldn't have to deal with the center plane of the features, which I don't want to be involved.

 

[Cncgotoguy]

I believe you can use profile without referencing datums in the feature control and without basic dimensions to control the orientation only

I could add datum A in the profile callout to hold the surfaces parallel to A and to themselves by calling out 4 surfaces in the callout?

 

[pylfrm]

Cncgotoguy,

It sounds like the took block only has one set screw, so only one flat on the sleeve would be used at a time. Is this correct?

It also sounds like the tool only has one flat, so only one pair of threaded holes in the sleeve would be used at a time. Is this correct?

Will each pair of threaded holes only be used with a certain flat, or are multiple combinations expected?


pylfrm