Keeping the tops of 3 tapers level - Tolerancing /Production

[LHELSEL]

We have a design situation that includes the use of 3 conical tapered parts that all fit together, and are metric parts. We have an outer housing, a sleeve, and a part that fits into the sleeve. I cannot get into specific materials, or why it is this way and not another, but suffice to say we have 3 20 degree basic included tapered parts that fit together, and we would like to maintain the top of the 3 tapered parts (top is perpendicular to the axis of the taper). According to ASME Y14.5, the best way to dimension this is to basic the diameter of a gauge point at a toleranced distance from the top, and provide a basic taper or angle. Our parts are all ~15 deep (tolerance not important for this feature at this point, and can vary widely), and going from outer to inner:
Housing: 7.5-0.2, 10.95 Dia basic (tight tol as is)
Sleeve: Housing mate 7.5+0.5, 10.95 dia basic (all the mfg can hold, strange propritary process)
Center part mate 7.5-0.5, 10.3 basic dia
Center (conical) part: 7.5+0.1, 10.3 basic dia (tightest practical tol)

As you can see, this gives us flush or within 0.2 of flush some of the time, but some times it can be ~0.6 below at LMC. We want all flush at the max, and at a min, we want the sleeve to support the center component.

We do have sucessful prototypes, but from a mfg perspective, this is not ideal due to scrap.

We also dont know if our customer is agreeable to "matched parts" wherein we define several dash part numbers that would bin the tolerances for all the parts tigheter, and have 100% inspection to get us "binned" and on orders specify a % range or ratio of each dash number, for example for the sleeve outside, a -1 would be 7.5+0.199, -2 7.7+0.399.

Has anyone had sucess with this method for a production run of ~1500 precision assemblies? We're going for a low rejection rate.

Here's some other relvant information / ideas we have discussed:

-replace the housing with a "collet" style part that would be more forgiving to the taper and allow us to "push" the sleeve and the center cone upwards towards flush?
However, due to the strength the material must have, this is most likely impractical.
-Replacing the sleeve with a poured in place material such as a hard solder.
-Parts cannot be ground down, that would cost more than scrapping 50% of the parts.
-sleeve mfg is a propritary process, vendor will not provide insight as to why this tol is so hard to control.
-once parts are assembled, there are no servicible parts, if a part wears, the entire assembly is replaced.

The questions i realy need answers to are:
-is there a way to solve the tolerances (based on the available band) to give us flush to 0.2 below flush.
-Is "binning" the parts practical for production of this scale?

-LMH

 

[cloa]

I bet someone will ask for a drawing so provision of as many details as you can will save time.

 

[GregLocock]

Conrods are binned for length AND weight, in mass production. Gas turbine blades are binned for weight. If the vendor is not prepared to do that then for a mere 1500 parts just get a workbench and a couple of contract tech inspectors for a week.


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

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

 

[LHELSEL]

Thanks guys!

@GregLocock: Not so worried about our vendor, my customer is a pain, and the prints transition to him at the end of the project, thanks for the fodder to convince him.

 

[ABritTurner]

From my understanding you have 4 parts, 3 of which have a taper of 20 degrees included angle. One of these parts is supplied as is and has a huge tolerance (20 thou). If this tolerance is a tapeered part then it seems you do have a problem.

The only way you can manage this is to match each manufactured part that you make to suit the supplied item (component.)

As you are supplying a complete assemble then it is possible to meet your customers specification but at a cost.

I suggets you meet with your customer and spell out the problems, give him a scedule of likely costs for different scenareos and let him decide how he wants to proceed.

The end result is not that ahrd to achive under normal circumstances, alomost 8 thou to play with, so there is no need for grinding unless the components are hardened.

If you press hard enough your cuntomer may be able to get the supplie to tighten up their tolerances on the 20 thou or you could remachine all the free issue parts to meet your designated size (if it is possib;e. If not you have to machine the third component to as tight a tolerance as possible so that you can supply the alst component to meet the wide tolerance that the customer has defined.

To sum up:

1. Try to get supplier to supply tight tolerance part

2. if not machine the free issue aprts to meet a common standard (if possible)

3, if 1 and 2 is not possible then make the 3rd compnent to a very tight tolerance (grind if needs be), this would be a matching operation and expensive/.

4 machine the last component to meet the cutmer's speciifcation, easily acheivable without grinding.

I hope this article has been helpful.