Question about dowel positional tolerance and total fit result in gas turbine engine

[giangnguyen92]

Hi everyone,
I am reading the old gas turbine engine book and I am in trouble to understand their definition of positional tolerance and “total fit” in doweled rotor.
Problem: bolts are used to concentrically locating 02 disks stacked and transport between them in gas turbine engine.
Bolt has fit mode to hole on each disk (here transitional fit) and it must go through 02 discs. And holes on each disk has positional tolerance themselves. I attached here the scanned picture for your reference.
According to the book (dimensions are in inch):
Bolt diameter = .7545 +/- .0001
Hole diameter = .75495 +/- .00045
Thus fit = .00045 +/- .00055, i.e. from .0001 interference to .001 clearance.
Hole true position = +/- .0005
Hence, total fit = .00045 +/- .00105, i.e. from .0006 interference to .0015 clearance
So my 02 first questions are:
- What is “total fit”?
- How is that determined?
I try to refer to ASME Y14.5 but cannot find anything similar to this. I am looking for your engineering comments.
Perhaps it follows different tolerancing definition in aero engine OR it is just the old way of defining?

 

[3DDave]

There is never a negative true position tolerance. Someone made up their own nomenclature.

 

[drawoh]

giangnguyen92,

I am looking at your numbers below.

[pre]
octave:19> boltmax = .7545+.0001
boltmax = 0.75460
octave:20> boltmin = .7545-.0001
boltmin = 0.75440
octave:21> holemax = .75495+.00045
holemax = 0.75540
octave:22> holemin = .75495-.00045
holemin = 0.75450
octave:23> clearmax = holemax-boltmin
clearmax = 0.0010000
octave:24> clearmin = holemin-boltmax
clearmin = -0.000100000
[/pre]

Given your dimensions and tolerances, an interference fit is possible. If you are working out the positional tolerances for for holes to assemble multiple bolts or to assemble a single bolt located from a datum, you work from the MMC[ ]condition. In your case, this is an interference fit. Your clearance is zero. Your positional tolerance is zero.

Your maximum clearance of .001" is extremely precise. Your drawing looks to me like a pitch circle.

If you are using one bolt only to fasten and locate your feature, then I assume your bolt is a location device, and your positional tolerance would be based on that requirement, and not the hole clearances.

Is it possible your clearance holes were drilled after assembly?

I have posted an article on calculating positional tolerances...

Calculating Locational Tolerances: HTML PDF

--
JHG

 

[giangnguyen92]

@3DDave:
There is an old style of positional tolerance, which is made by +/- (coordinate tolerancing)
This equation is cited from the book "Jet Engines and Propulsion Systems for Engineers" - General Electric, page 3-57.

http://library.lol/main/CB568B69136A85121F545E5AB84ECA88

So I think it is serious enough.

 

[giangnguyen92]

@Drawoh,

Thanks for your support.
Actually there are numerous bolts (for example, 72) distributed equally in the circumferential direction over a disk. And these bolts are used to locate and connect the 02 disks.

I analyse this fit hole/bolt diameter and I find out that this fit (-.0001" interference to +.001" clearance) corresponds to H7/js4 fit, which is extremely precise as you pointed out.
Do you often see this type of fit?

And because of this precise fit, the designer must specify positional tolerance of holes over the 02 disks properly so that the bolts can go through both to locate the 02 disks.
Here the designer at GE specify positional tolerance of holes on disks as +/- .0005.

What I do not understand is the term "total fit" and its calculation. I never see anything like this. Its calculation for hole MMC, bolt MMC and hole positional tolerance is not similar to formula on appendix B, ASME Y14.5 2009 (those you posted).

 

[3DDave]

Serious enough? Not sure what that has to do with it.

Thaddeus made up his own nomenclature. It's not my fault he did that. It's not "an old style" it's something he invented.

 

[giangnguyen92]

@3DDave
Yes, this is the document for training GE engineers, so just assume that it is serious.
Do you mean the Hole true position = +/- .0005 nomenclature? It is called coordinate tolerancing.
I found the comparison between old coordinate toleracing vs modern GD&T here. (Figure 3 versus Figure 4)

https://www.gdandtbasics.com/gdt-vs-coordinate-tolerancing/

 

[powerhound]

Sorry giangnguyen but Dave is right. If there's any implication that a +\- tolerance can be equivalent to a position tolerance then it is not in accordance with ASME Y14.5 OR the ISO standard and not even as an "old" method of locating a hole using position. I find that using the phrase "made up" is appropriate in this case.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech

 

[3DDave]

I didn't say anything about "serious" to begin with. You wrote that. Not sure why. Is this a life-or-death matter for you to know? It's an antique book that, given this error I would not trust too much.

I know what coordinate tolerancing is. That's why I said it wasn't true position tolerancing. Coordinate = X + Y dimensions; that is two orthogonal controls on location. You have ORDINATE tolerancing as you have only one dimension.

What is it you want?

 

[giangnguyen92]

Thanks Acosta and Dave,
Yes, it is not in accordance with ASME Y14.5 OR ISO standard. I guess that it is the brief writing of coordinate tolerancing, like the Figure 3.

I am finding the way to locate my 2 rotating together disks using dowel construction (36 holes circumferentially distributed on each disk). That is why I want to understand these terms by GE.

Of course, I believe that these formula have worked well in GE engines. The only problem is that I do not know how they are derived.

Here I question that if they applied coordinate tolerancing for holes on 2 disks like those above, would bolts be able to go through 2 holes and do this function as locating device?

Just quick analysis, I find that if holes are at extreme corner of +/- .0005 coordinate tolerancing, MMC bolt will not go through MMC hole easily, but is that normal to dowel construction?

Do you know what "total fit" mean?

 

[drawoh]

giangnguyen92,

Read the link I provided you. While [±][ ]tolerances are not equivalent to GD&T positional tolerances, you can calculate the tolerances needed to make your part work. My article explains how to do it. My explanation of [±][ ]tolerances makes a good case for proper GD&T.

Regardless of how you do it, your calculation of tolerances is done on the assumption of some clearance that is greater than zero, at MMC. You do not have this.

I suspect that your holes were drilled and reamed after assembly. This is the only way to achieve absolutely perfect positioning of the holes. Your holes will only ever line up if your parts are rotated correctly. There almost certainly is a keying feature.

Just quick analysis, I find that if holes are at extreme corner of +/- .0005 coordinate tolerancing, MMC bolt will not go through MMC hole easily, but is that normal to dowel construction?

The extreme corner of a functional [±][ ]tolerance is based on the clearance in the hole at MMC, as I explain in my article. At MMC, your clearance is zero. Your [±][ ]tolerances are zero. Your positional tolerance diameter is zero, at MMC, anyway. Even at LMC, I am not certain your resulting positional tolerance is fabricatable.

--
JHG