How to refine parallelism on two self datum plates

[bxbzq]

Question see attached pictures. In the welding assembly drawing in last page, the function allows large position tolerance between the two plates, but requires tight parallelism tolerance to each other. How to make it?
Comments on other callouts especially the datum sets are welcome.

 

[cjccmc]

Do you want only the two faces that red arrows contact or all four faces to have the .5 parallelism? If all four then you may want to control parallelism on the support plate detail. I would put a .1 flatness on the datum A for that and .1 parallelism on opposite surface wrt A. Then at weldment assy level make one face a datum then call out a face of the other support parallel .4 to it.

I found the datum B for the slots and holes could use some clarification. Did you want just one hole/slot to be the datum or the entire pattern of 14 as the datum?

 

[bxbzq]

cjccmc,

I would say I want the centerplanes of the two plates to be 0.5 parallel to each other. The two plates are equally important in terms of function. Calling thickness of one plate datum feature A and the other relative to datum A implies one plate is more important than the other.

What about this?
2X (8)
2
POS
0.5

From the std, I don't see it is not allowed. However, the problem is it only does part of the work.

About the 14 rectangular holes, again, they are equally important. The way I call them out indicates the 14-hole pattern is datum feature B.

 

[cjccmc]

Not sure that POS is legal, it could be, but my suggestion is to use a callout that makes your intent more clear. I think the datum A on one face the way you have it is good, from there just add parallel .5 to A on one of the other faces of the second support plate.

 

[pmarc]

I am just wondering, what would happen if:

1. On TROUGH drawing (page 2) datum feature A stayed unchanged, width 126 became datum feature B, length of trough or one side of the length became datum feature C. Other features, including rectangular holes, were controlled to A|B|C.

2. On SUPPORT PLATE drawing (page 3) datum feature A stayed unchanged, top surface (the one interrupted by 7 pads) was assigned as datum feature B, length of the plate or one side of lenght of the plate was selected as C. The true contour of the plate was controlled by profile all around wrt to A, and 14 holes were positioned to A|B|C.

3. On WELDMENT drawing (page 4) datum features were the same as on TROUGH drawing. Two callouts were applied to pattern of two support plate widths: 1) position 2 wrt A|B|C; 2) perpendicularity 0.25 wrt A|B (perp. 0.25 for both plates wrt A|B would result in max. 0.5 between the plates). 28 holes were controlled by two single-segment positional callouts: 1) |pos|dia.X|A|B|C|; 2) |pos|dia.Y|A|B|. Y value would probably have to be equal to the positional tolerance value applied to the pattern of 14 holes on SUPPORT PLATE DRAWING. X value would be greater than Y and chosen depending on maximum allowable (from application point of view) shift between two plates and the trough in horizontal direction (normal to datum plane C).

 

[Vimalmechs]

bxbzq, Can you clarify some minor doubts. PFA.

1) Dimension references (Keyway-slot,31 & 41) are different in assembly and part level drawings. Why it is so?

2) Your functional importance is only in between plates.No relation with ur weldment.

3) Can you provide the "female part-channel" dimension details.

4) There can be a gap of 1.18" in Worst case. Is it acceptable.? Are you fastening the parts with separate fixture arrangement ?

 

[bxbzq]

First of all, the tolerance values shown on the drawing are just from my gut feeling. Not to function, not to manufacturing capability. I learned recently many of those tolerance values are much tighter than manufacturing process can hold. The OP was to figure out proper datum set.

pmarc,
On the weldment drawing, the support plates are to assemble with slotted parts. This is the reason I select them as primary datum. The function of the trough is carry billets. The sheet metal bending process making the trough gives quite large variations on the 126 width, much larger than the tolerance currently shown on the drawing. I tend not to assign it as datum feature.
Maybe use the bottom face of the trough as primary and the two plates width as secondary would do the work. So the callouts to the two plate width would be
1) position 2 wrt A
2) perpendicularity 0.25 wrt A
To make the drawing manufacturing friendly, one end of the trough can be tertiary datum.

Vimalmechs,
I have problem to download the file in your link.

 

[axym]

bxbzq,

Regarding the tight parallelism tolerance of the plates to each other, what you need is a "mutual orientation" control that requires the two tolerance zones to be perfectly oriented to each other without being oriented to a datum. This would be along the same lines as a Position tolerance with no datums, to control mutual coaxiality or mutual coplanarity. The standard currently doesn't define such a control - you can approximate it, but not get it exactly. With the current rules, orientation controls must have a datum feature.

Another way of looking at this is that the two plates must both be oriented to the same datum, but you can pick the datum. As long as the plates are both parallel to some spatial direction within a 0.5 tolerance, they meet the requirement. This is a bit too metaphysical for a lot of people, so it hasn't reached the standards yet. But I think that it will eventually - there is a need for this tool.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[pmarc]

Evan,
How about something like this:
- one width as datum feature A;
- second width as datum feature B;
- basic 41 dimension between centers of the widths;
- position FCF |pos|2|A-B| applied to the first width;
- position FCF |pos|2|A-B| applied to the second width;
- parallelism FCF |par|0.25|A-B| applied to the first width in addition to the first position FCF;
- parallelism FCF |par|0.25|A-B| applied to the second width in addition to the second position FCF.

Granted, there will still be some open questions with this scheme, but won't this conceptually be similar to fig. 4-25 in Y14.5-2009?

 

[fsincox]

axym,
Are you thinking along the line of the old ISO notation of a parallel callout with 2 leaders to surfaces and no datum reference?
Frank

 

[pmarc]

Frank,
If you are thinking about method presented in the attachment, notice that still one of the features is datum feature in this kind of callout.

http://files.engineering.com/getfil...8ff-74853fb0c009&file=ISO_old_parallelism.PNG

Or are you thinking about another ISO "datumless" FCF method?

 

[fsincox]

pmarc & evan,
Yes, it was different. I have an old book "Modern Engineering Tolerancing by Rowland Hill & Cecil Jensen" Printed and bound in Canada, published by McGraw-Hill Ryerson Limited, 1976.
It clearly shows this and says under the heading "ORIENTATION TOLERANCING WITH NO SPECIFIED DATUMS" SIMILAR SURFACES: It sometimes happens that a part has two similar surfaces which require control, with no ready means of distinguishing between them. If one of the surfaces was designated as a datum it would be impossible to distinguish it from the other surface. It is common practice in such cases to direct the feature control symbol to both surfaces, WITHOUT DESIGNATING A DATUM. This creates two similar tolerance zones for both surfaces(looks like a simultaneous requirement then). I thought I had copies of old German DIN standards that showed it also but I have yet to find them. They were in German and I could not read German so this Canadian book was very insightful to me in studying and contrasting ISO.
Why only two surfaces? I do not know, it seems like a needless simplification by the “lets over simplify GD&T to sell it/feed it to the masses”, stuff.
Anyone else out there have a copy of the old ISO stuff?
Frank