Tolerances for four accurate holes pointing at a finite virtual point

[kolvos]

Hi,

I'm a bit confused as to how to define tolerances for four accurate holes pointing at a finite virtual point. (Picture attached)
I'd also like that their depth will be toleranced between each other.

What Linear and geometric tolerances should I define for such an application and how? I'm attaching a picture that will help understand. I'd be glad if you could reply with a annotations on this picture.

I'd be glad for some help

 

[btrueblood]

Tricky one; in a past life, helping with the design of liquid propellant rocket engine injectors, we tried to figure it out using GDT and were stymied. There are now such things as "projected tolerance zones" which may help, but you still need to define that you want the centerlines of the holes to intersect at a point in space, plus or minus some tolerance. Short of describing the desired geometry with those words in a note on the print...I dunno.

 

[KENAT]

Unless I'm missing something you can probably do this with basic dimensions and position control.

First thing is to work out what your functional datums are and then go from that.

May be a better question for forum1103 in future.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[kolvos]

Hi Kenat,

Could you perhaps elaborate a bit? Maybe annotate an example of a functional datum. (A datum that I will define as very straight) and annotate what kind of basic dimensions I should give?
How do I relate between the four holes?

 

[KENAT]

Functional datums is essentially related to/derived from how it mates with/works with what ever it is assembled to.

I don't know that - I don't even have other views of this part.

Regarding the depth relationship between the holes, now you've got be thinking. Do you have some guidance on how tightly controlled you think you need it?

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[kolvos]

Hi KENAT,
Thanks for your responses.
I re-attached here the document in which I underlined in red the reference plane I'd like the holes to be relative to.
Regarding the depth - I'd like them to be the linear dimensions +-0.05mm relatively to the virtual point annotated in the document. But let's say that's slightly less important.
I'm mainly interested about the concentricity of the axes of the 4 holes relatively to the virtual point.
Would be great if you could show me an example of a basic dimension I can give between the holes

 

[racookpe1978]

Most importsnt, what is your required accuracy - NOT of the drawing! - but of the four finsihed assemblies as they aim into the projected intersection zone?

If you will, the only appropriate question is: How accurate must the four "pipes" ?? be to aim their four beams (??) so the beams hit each other?

When I have done this machining and welding (the distances were much longer at 10-12 feet, but the "beams" were IR cameras that had to be aimed at the blade roots of a combustion turbine. Thus, the "camera aim points" could be as much as 1/4 off and still produce a useable image for the testing.

I found that laying out the holes and the drill points through our 1-1/2 thick steel exhaust walls was difficult enough, but the hardest thing to do was "aiming" the centerline ofthe camera tubes at the target point and positioning each tube for the tack welds. Once tacked in place, then the welding had to be carefully maintained so shrink and draw movements from the weld heat didn't distort the aim point.

I solved the "aiming" by getting a laser mounted on the camera housing centerlines with a temporary jig and fixture, then aiming the camera tube with one person inside providing directions by radio to the welder and "laser aimer" outside with a fourth holding the tube in place by hand for the tack welding. Once tack welded, the finish welding didn't require the extra hands.

You will need a fixture or target precisely positioned so the laser beams can actually "hit" that mid-air aiming point simultaneously. Now, if all four lasers hit each other, but the aiming point is not exactly right, will that result be "good enough"? My laser beams were 1/8 diameter - will that diameter beam be close enough to allow you to aim the tubes or drilling bores? You can mount the laser inside the drill chuck to verify where the drill itself is going to be aimed once the gadget is mounted in the vise/machine tool table. A 3D CAD/CAM mount might work as well, but expect several trial and setups.

2. How do proposed to install and build the gadget with those holes and intersections with the outside wall?

3. You have a mixture of ANSI dimensions/callout for pipe threads and metric dimensions: Do you really want to do that? It will cause problems?

4. A screwed thread 1/4-18 NPT will NOT precisely position the center instrument vertically: Do you really want to use normal pipe threads for that fixture?

5 I don't think that "8" (mm ???) deep is not enough engagement for regular pipe threads: They will likely bottom out the threaded fixture. Unless that is your intent - but if so, then the fixture will not be tight in place.

 

[kolvos]

Hi and thanks for your response.
Disregard the center NPT hole. The NPT is just for a collet that will clamp the central instrument after positioning with a jig.
I'm talking about the tolerances I need to give during the manufacturing of this part on a CNC.

Regarding numbers - I'd like that all of the centerlines will point to the same point with a ~50 micron tolerance.

 

[tygerdawg]

Am I missing something? Perhaps so, I'm fighting a virus and am quite fuzzy-headed today. Apologies if so & ignore my comments.

Is this a cross section of something that looks like a linear bearing rail? If so...

If you think like a machinist, then it seems logical to ignore your "virtual point" except in design construction. A machinist would use the bottom horizontal flat surface (say, Datum A) and one of (or both) of the vertical side surfaces (Datum B, Datum C) as fixture registration surfaces. Perhaps also use the angled surfaces into which the holes are bored as datums to determine the hole's depths (Datum D, Datum E). Then the machine tool used to bore the holes would hit the angle to the tolerance that you specify ***relative to the physical datums*** . No matter WHAT you design, you still must be able to measure your results and see if it achieves your goals.

TygerDawg
Blue Technik LLC
Virtuoso Robotics Engineering

http://www.bluetechnik.com

 

[dhengr]

Kolvos:
I can’t see you attachments in that format. Try a pdf file. But..., could you define one plane, with two of the lines being legs of a triangle off your base, and I assume this is standing vert. for the moment. Do the same thing with the opposite side/plane, two lines as the legs of a triangle off your base, and standing vert. Now, lean the triangular planes over/in until the two apexes touch. Tolerance each of these steps in some way. In affect, you are building a pyramid which may not be symmetrical, and you may want a boss, on your base, which is perpendicular to each of the four lines which make up the corners of the pyramid.

 

[Compositepro]

The most practical way to obtain a 50 micron accuracy is to provide aiming adjustments, just like just about every every optical device has. You do still have the problem of defining a datum and dimensioning your part. It seems that that will depend on how the part is to be made.