Impact of using 4 Mtg holes vs. 8 Mtg holes in a fixed fastener instal

[apekim]

I would like to understand the impact of using 4 of 8 available mtg holes in a fixed fastener scenario.

Example:
A plate has 8 thru holes. It's mating part has 8 3/8-16 thread inserts. How much better alignment will I have if I use all 8 mtg holes vs using just 4? I am concerned about rotation only, not movement in the x or y directions.

Hole info:
.404-.411 thru
.056 M |A|B|C|
.014 M |A|

I have no tolerance info on the mating part.

Thanks (in advance).

 

[MechNorth]

Your question is kind of open-ended, and the answer changes depending on what the primary alignment of the features is (e.g. press-fit pipe & fitting, loose square components, flat flange faces, etc.). In general, it's poor design practice to use fasteners for alignment, shoulder screws being a notable exception, unless of course you want to allow limited slop. The clearance hole in the non-fixed component will automatically allow some movement; two such holes is all that's needed to stop rotation. When I've had to use fasteners for anti-rotation in this way, I've snugged the fit of one or two (depending on the situation) to restrict movement; it's a common practice taught by many GD&T instructors. In that case, the other clearance holes can be opened even further, garnering easier manufacturing and assembly.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[KENAT]

From a fit point of view of holes aligning properly then if using position tolerance correctly number of holes makes no difference in fixed or floating fastener situation.

To control rotation of the parts that have the holes in them then the spacing of the 2 most greatly spaced holes also has a significant impact, as well as the 'slop' of the fasteners in the clearance holes.

So, assuming consistent fit etc. you'll get the same rotation control by using the 4 extreme corner holes as by using all 8.

As MechNorth says, using threaded fasteners for precision alignment isn't great practice. Maybe stick with 4 fasteners but add some plain pins or something. The separates the alignment function from the fastening function. If you're stuck with the part with threaded holes, can you get some kind of spring/roll pin that picks up on minor dia of the thread.

Posting guidelines faq731-376

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

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

 

[apekim]

Jim,

First, thanks for the reply. I refer to the tec-ease site often...nice site.

Regarding the situation:
I have been discussing this with others here at work. Most believe that using all eight holes would result in an improved alignment of the two parts. I am not fully convinced though...hence the posting.

My thought (so far), was that using all 8 holes would not reduce your worst case alignment of the two parts, but it may reduce the probability of being at worst case alignment when the parts are assembled (since there is no controlled installation process).

Your thoughts??

 

[KENAT]

Your logic sounds about right.

Whichever are the 2 most far apart holes have the greatest impact on rotation. The worst case fit (small screw big hole), at the worst case max spacing controls your worst case orientation.

Using all of them may decrease the likelihood of hitting worst case rotation, but doesn't it also decrease the likelihood of hitting best case rotation?

Posting guidelines faq731-376

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

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

 

[drawoh]

apekim,

You can locate a two hole pattern with a sloppy ± tolerance between the holes. After that, you have to calculate on the assumption that each hole is located by an external datum. There is no difference between three holes and twenty eight holes.

The more holes you have, the more likely one or more holes are located at the extreme limits allowed by your tolerances, thus imposing correct alignment. This all assumes you calculated your tolerances correctly, and that the part was fabricated to your drawings. This situation is much more likely if your tolerances are close to the capability of your fabricator. This actually is not a good idea.

JHG

 

[apekim]

Drawoh,

Your statement "you have to calculate on the assumption that each hole is located by an external datum" is not clear to me.

The .014 at Max Matl is indicating that the locational tolerance between any two holes can not be more than +-.005" apart (plus up to the bonus of .007" beyond)...correct?

 

[drawoh]

apekim,

I have worked this stuff out on my website...

http://home.eol.ca/~hgibson/tolerances

If you have two bolts (as opposed to screws), separated by a single ± dimension, your hole tolerance works out to ± the clearance between the bolts and the holes. This is a one-dimensional problem. The parts being clamped can be wiggled around to assure clearance.

The standard hole tolerance problem assumes that the clamped parts are located by external datums. In this case, each hole must be located inside a two-dimensional pattern. In current GD&T practise, this is a circle. A clearance hole for a bolt should be located inside a circle equal to the clearance between the bolt and the hole.

If you have no external datums and a three hole pattern, two holes in the pattern effectively act as datums. Any additional holes are located by these external datums.

Do not convert linear tolerances to positional tolerances. It will not help you to understand what you are doing. Read my notes on linear tolerances and ASME Y14.5M. I put a question up here (thread1103-261904) and got a discussion going. I am satisified that linear tolerances to holes are not very well defined, especially if your part outline is not accurate.

If you do the geometry and the arithmetic, positional tolerances make a lot of sense.

JHG

 

[restedup]

apekim,

It appears to me that you are in desperate need of the locational tolerances for the mating part. Without that your are shooting in the dark?

My opinion.

 

[restedup]

apekim,

If I am reading your thread properly, you are using .014 dia for your hole loc. tolerances. With this being a fixed fastener assy and your hole size of .404, that means that the threaded insert will require a projected tolerance of .015 dia. I find that to be rather restrictive for a threaded insert. This bears out the need to obtain the actual tolerance on the mating part.

This not address your concern for rotational accuracy, but is critical to assure interchangeability of parts.

 

[MechNorth]

Apekim, tks for the comment on the site.

Your thoughts on the reduction of rotational displacement MAY be valid, but are subject to the manufacturing method and fixtures used. My thoughts are that if you are using an NC mill to move in the x/y plane and pop in the holes, then the holes will all be biased in the same direction, so I don't see the advantage being gained. If the mill is stationary and an index table or such is used to rotate the part about a central axis, then diametrically opposed holes may be oppositely biased, more likely leading to the situation you hope for. Of course, all of this assumes that the fixturing and cutting method are adequately rigid/accurate to produce a consistent result.

As mentioned by others, start with the locational tolerances provided for the mating (fixed) part, and look at restrictions from that point. If you can, use an alignment pin of some kind, or restrict the clearance on two of the holes while opening all the others, and use the fasteners. Tightening all holes won't add value, just cost of manufacturing and assembly.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[apekim]

I think we are losing sight of the initial question.

I am not trying find a more benefical method to ensure less rotation. With everything as stated in the original posting...is there a benefit (reducing potential rotational alignment) to using 4 mtg holes vs. all 8.

Thanks

 

[MechNorth]

I don't think that we lost site of the OP, we often try subtlety to guide people to a more appropriate method rather than directly telling them their methodology is flawed; there is a lot of reading between the lines required, rather than just looking at the direct content. In this case, it is clear from the feedback that the addition of more holes is not considered beneficial.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[restedup]

Apekim,

If you are planning to attach with all 8 screws, I dont think the question of quantity of holes enters the equation. All holes must fit simultaneously. Whether you use 2 or all 8. Again you really need the location for the mating part to calculate the tol.