Over complicating simple part 1

[KENAT]

I just got a really good question from an intern and it made me start to think I’m over complicating things, or maybe having a mental block, either way here goes. Attached is a sketch similar to the part he showed me but with the numbers rounded.

http://files.engineering.com/getfil...3-4c85-a973-127c67d85916&file=simple-part.tif

The part is a simple clamp plate, nominally symmetric. The only important thing functionally is the size and spacing of the 2 holes. The part gets secured in place by fasteners through the 2 holes to holes in mating part. The location of the holes relative to the edges etc is not very significant, so the pattern can move quite a bit on the part.

How would you fully dimension the part, bearing mind the need to physically identify datum’s (to avoid any ambiguity during manufacture or inspection) but not wanting to introduce excessive inspection requirements etc.

One idea I considered was adding a chamfer to one corner (shown in phantom), just to provide some kind of orientation to allow allocation of physically identifiable primary datum (and perhaps secondary & tertiary). However, it seems almost wasteful to incorporated a feature to help with dimensioning that serves no real function.

Once that’s done though, is using the 2 holes pattern as a datum (4.5.8 & figure 4-22) the logical choice? Or even though it’s not really driven by function is making edges secondary & tertiary more likely to be understood by other users while still meeting function, perhaps with composite position to allow movement of the pattern?

Working to ASME Y14.5M-1994, tried a search of this site but didn’t find satisfying answer, even though I recall some aspects of this being discussed before.

Thanks,


KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

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

What is Engineering anyway: faq1088-1484

 

[PeterStock]

I would just give a tolerance to the .25 thickness and identify one the sides as datum A. Then make the 2.00 and 1.00 dims basic and provide a profile tolerance of .035 or so. You can add basic locating dims for one of the holes to the sides just to avoid making the assumption of symmetry. The only datum you need is the A.

Peter Stockhausen
Senior Design Analyst (Checker)
Infotech Aerospace Services

 

[KENAT]

But how do you tell which is A and which isn't as required by 4.3 in the standard?

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

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

What is Engineering anyway: faq1088-1484

 

[PeterStock]

That does not make much difference as long as the part functions. If you make a functional gage, if the part fits in one way, it usually will the other way. If it fits in one way but not the other will it still work?

Peter Stockhausen
Senior Design Analyst (Checker)
Infotech Aerospace Services

 

[TheTick]

Perhaps the centerplane of the outside surfaces could be the datum. This noted by attaching the datum symbol to the 2.00 and 1.00 dimensions. (Please correct me if I am mistaken.)

Perhaps a composite tolerance? Keep the hole pattern tight but give the pattern plenty of room to float.

Drawings are for communication. Adding the rest of the necessary dimensions takes all of a minute and can save many times that in 'splainin'.

 

[MintJulep]

Couldn't you work it out so that it doesn't matter?

 

[ctopher]

You indicate that the part is symmetric; then there is no point of the chamfer.
Just add datum A to one of the .25 surfaces, datum B to 1.000, then datum C to 2.00.
If not comfortable with this, and it is important, then remove the true position from the holes.

Chris
SolidWorks 08, CATIA V5
ctopher's home (updated Aug 5, 2008)
ctopher's blog
SolidWorks Legion

 

[TheTick]

I agree that ctopher's suggestion is technically correct. However, I believe that it may be stretching the limits of prevailing levels of GD&T literacy and possibly of shop math skills in many quarters. At least add some reference dims.

 

[drawoh]

KENAT,

My primary objection to using the length and width as secondary and tertiary datums, as TheTick suggests, is that they are sloppy. On the other hand, it sounds like your requirement is sloppy. A composite tolerance block would describe your requirements accurately, even though it might alarm the fabricator.

TP Ø.02 A B(MMC) C(MMC)
TP Ø.014(MMC) A

The MMC on the datums makes inspection tooling easier to design, and it controls the maximum extent of the outline upon installation.

JHG

 

[pmarc]

KENAT,

Please take a look at attached picture:

http://files.engineering.com/getfil...4934-bd93-f7bce4826ea4&file=plate_2_holes.tif

With such dimensioning:
- You don't have to use length and width as secondary and tertiary datums,
- Spacing of the holes is controlled by assigning one of them as a datum feature and controlling the position of the second one relative to the first,
- There is no difference which side of the plate would be a datum feature A (I would just add the same value of flatness control to both sides of the plate, e.g. .01). So you don't have to make any additional features to undoubtfully show which face should be a primary datum.

Im a curious what you are thinking about it?

Regards

pmarc

 

[TheTick]

The 1.00 dimensions from the imaginary center are useless. The dimensions should come from one of the holes, preferable "B", to relate the outer profile to the holes.

 

[ctopher]

The pic from pmarc is much too complicated for such a simple part.

Chris
SolidWorks 08, CATIA V5
ctopher's home (updated Aug 5, 2008)
ctopher's blog
SolidWorks Legion

 

[powerhound]

I think if you take pmarc's drawing, eliminate both the 1.000" dimensions from the center to the outside edges like Tick said, and both the .500" dimensions, and leave pmarc's profile of a surface callout (with the value modified to what your requirement actually is) but change the DRF to be WRT datum A only and on the hole position callouts as well, you should be in good shape. This is similar to what was discussed in a previous thread in which a tec-ease tip on simultaneous requirements was posted. Here it is again:

http://www.tec-ease.com/tips/dec-07.htm

Upon re-reading what I just typed, it seems rambling. Let me know if I was clear.

Powerhound, GDTP T-0419
Production Manager
Inventor 2009
Mastercam X3
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[axym]

I thought pmarc's drawing was pretty good (issues with basic dimensions notwithstanding). The only change I would make would be to define the two holes as a pattern and reference the pattern as the secondary/tertiary datum feature. This would match the described function a bit better, as neither hole takes precedence over the other.

Chris - It's a bit disheartening to see you dismiss pmarc's drawing as "too complicated for a simple part". It may seem that way, but his tolerancing scheme represents the functional requirements very well. If the inspection was done correctly, the results would accurately reflect fitness for use.

PeterStock and TheTick - the alternative datum schemes you suggested might be "simpler" in a certain way, but they don't represent the functional requirements. The inspection results might not agree with fitness for use.

Powerhound - for what it's worth, I understood your rambling post about simultaneous requirements. While I agree that referencing everything to A is technically correct, it unfortunately falls into the "almost guaranteed to confuse" category.

This is one of the eternal difficulties with GD&T - the geometric requirements of real parts are usually complicated, even when the nominal geometry is simple. It's just a plate with two holes in it, but expressing the requirements can result in some fairly complicated GD&T. Attempts to "simplify" the drawing can result in other types of complication, such as ambiguous inspection criteria or inspection results not reflecting fitness for use.

OK I'm done, I'll step off of the holier-than-thou QA-guy soapbox now.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[KENAT]

Don't have much time right now but one question, it seems just about everyone is saying not to worry about the physical identification of the datums as required by 4.3, especially as regards the likely primary datum.

Am I correct in my understanding of the suggestions being given?

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

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

What is Engineering anyway: faq1088-1484

 

[ewh]

Am I the only one that cringes seeing dimensioning crossing like that when it is so easily avoidable?

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[drawoh]

ewh,

I think TheTick and powerhound addressed it. It sounds like outside dimensions showing the total length and width would make everybody happy.

Personally, I am fine with using the two holes as secondary and tertiary datums. It is important to understand that the outside profile is dimensioned to the holes, not vice versa.

If this were my drawing, I wouod not have bothered with the FCFs on either hole. I would have applied a ±[ ]tolerance betweent the two holes. The two dimensional positional tolerance on datum[ ]C is meaningless, since it is a clocking feature.

JHG

 

[ewh]

They addressed the utility of the dimension; I'm addressing the sloppy placement. It seems that it is no longer a concern that this be avoided (as per the standard). It probably took longer to put the gap in the extension line than it would have to move the dimension.

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[pmarc]

There has been already some comments regarding my version of KENAT's drawing, so maybe I'll try to explain myself a little bit:

General remark:
This case is another very good example how difficult and ambiguous GD&T can be in describing even simple(?) applications. I think the discussion we have here is sufficient evidence.

Coming back to the drawing:
The Tick - I agree with you that 2 dimensions 1.00 are meaningless. I do not know what I was thinking about when I was placing these dims. It would be much better when they come from hole assigned as "B".

Evan - thanks a lot for your comment - I wouldn't write it more precisely, especially that English is not my native language. I totally agree with your opinion which I called "complexity of a drawing vs. simplicity of a part". Additionally I am also not fully convinced to simultaneuous requirements method using single datum A as it can produce a lot of misunderstanding.
There were 2 reasons why I didn't use 2 holes as a pattern (with perpendicularity control 0.014 relative to A):
- First of all I wasn't 100% sure if perpendicularity relative to A controls spacing between the holes despite that the basic 1.000 is given. (If it does I see no problems with using them as a pattern and as a secondary/tertiary datum. But then 2 basic dims. 1.00 look OK for me),
- Secondly variant with pattern allows to have greater variation of distance between centers of the holes when they are on MMC (0.986 - 1.014) comparing to my version (0.993 - 1.007). I somehow thought that requirement is to have this variation as small as possible, therefore I chose the second option even that one of the holes takes precedence over the other.

KENAT - as I wrote in my first post if you add equal flatness control to both sides of the plate (which I've omitted unintentionally on my drawing) there will be no difference which side is chosen as datum A. I realize that symmetry of the part might cause that two people will pick two different faces as a primary datum A however in my opinion with such dimensioning it wouldn't have any impact on part quality.

ewh - sorry for this dimensioning crossing - I wasn't focused on this kind of details. Especially that internal standards of my company allow me to make such crossings with the gaps. I realize it can be "painful" to see such things for someone who works according to ASME standard. My company follows ISO and units in milimeters so for me it was also weird to use third angle projection as well as not to place 0 before decimal point (.02) or not to remove trailing zeros (1.000).

Regards,
pmarc

 

[TheTick]

No matter how mundane the part's use or how loose the tolerances are, it still requires a certain number of dimensions and constraints to fully describe it. This would be true for a rough spacer or a precision die block.