Still GD&T Resistance 2

[KLineDesign]

I'd like to start with the disclaimer that I really just prefer the dimensions and tolerances (be it GD&T or rectangular or a combination) to just describe what we need the part to be (meet the requirements essentially). Now the issue:

We keep getting into GD&T pissing matches here. The last one essentially was around dimensions only used for machine programming that aren't inspected. I see no reason why they can't have a tolerance, but the path chosen was to rely on the qualification of the machine capability and show basic dimensions with a note reading they are for programming only. Really we just said +/- .002 when the reality is these dimensions could be positional +/- .125 for one of them and the other whole area could be a profile callouts of .030 easily.

I kind of get the idea, tools and methods are spec'd in the setup sheets, so why not let that handle the acceptable tolerances there, but I wonder is there a good way to do this on the drawing? Has anyone instituted something like this? (Letting the process capability handle the tolerance, and the model and or basic drawing dimensions just showing perfect nominal to program to.)

Thanks,
KLine

 

[3DDave]

No reply here will solve the pissing match, because there isn't anything rational about it.

The rational thing is to evaluate the costs and chose a path that minimizes them, but that isn't easy so instead people fall back to using anything at hand as a lever to try to show how important they are.

Lots of people I have worked with have put unrealistically large tolerances on drawings and never see a problem because process capability was so much better.

 

[CWB1]

The design print controls the part, the process sheets and manufacturing markups control the tools and process. Do not mix and match these ever because there is a very real cost to doing so.

In this instance you're putting +/- 0.002 on a print when +/- 0.125 could work as well. Ok, so why not use +/- 0.125, or better yet a general tolerance applicable to all surfaces unless specified otherwise? Would your part still work at +/- 125" (no decimal point)? Its an extreme example but without a tolerance on the design print its perfectly acceptable.

 

[KLineDesign]

I agree. I want the tolerances specified we want the part to perform within.

But, at the same time, is there a way we can correlate all the non-important dimensions programmed to drawing values, models, or 2D .dxfs will run within a general tolerance based off machine capability. Each time I try to get on board with such notions, I come full circle to just specifying what is acceptable, but it seems like maybe - just maybe we could be onto something. See, the trouble driving the apprehension here (with some anyways) is the notion of general tolerances being checked and fear of full inspection of the non-important dimensions. I'm just trying to drive some consistency across part types, but maybe it's not possible and it truly is specify everything but argue to only check what needs to be checked.

KLine

 

[drawoh]

KLineDesign,

The concept I am learning at my new job is that dimensions and tolerances are placed on the drawing to show what we will inspect when the parts are delivered. Everything else is MBD (Model Based Definition). There are notes specifying tolerances for the MBD. If I don't care (much) about it, I am discouraged from dimensioning it.

The standard explanation for positional tolerances is that they allow 50% more variance than X/Y [±][ ]tolerances. Actually, if you are measuring [±][ ]dimensions and rejecting 20% of your parts, GD&T positional tolerances will help you, but making clearance holes larger will be the intelligent fix. What I don't like about [±][ ]dimensions is that the tolerance applies to each and every hole that is on the line you dimensioned. The positional tolerance goes on the hole, not on the dimension line.

If you are sending drawing out a vendor, you need to specify what you are going to accept. When your vendor or shop prepares drawings, they need to be process driven. The CNC punch places the hole to a certain tolerance, regardless of what you specified. If it is within specification, everything is good.

--
JHG

 

[KLineDesign]

This is indeed internal stuff, and that is part of the problem. Qualified tools to inspect certain things. I'm not an expert, but essentially we have a lot of tools that can measure everything we need, but getting the tools to a qualified state actually is limiting what we can currently measure. This is kind of an aside though.

If you have 10 dimensions on a print covered by model based or other general tolerances. Kind of like the punch example (which we have) say 10 holes that are punched have a positional tolerance of .010 and the punch runs within .003 all day everyday - we want to say it meets without having to actually measure it. Some have been able to simply justify this in a tech review, but it just seems like there is a better option that ties up the idea of Model Based generals and Process Capability.

 

[Nescius]

Relying on the inherent capability of a process to make good parts is equivalent to skipping tolerancing altogether and simply calling out the process by name on the drawing.

Why apply a flatness tolerance? Just call out the surface as "ground". What do you mean that's not interpretable?! Grinding is really, really accurate!

I fight these battles daily. I often find, if you really start looking, that the process is not as consistent as its defenders claim.

 

[mkcski]

I am really enjoying this thread. I too have fought this battle many times. The logic falls short with many who do not understand the difference between product definition and mfg and inspecting the part (to make sure it meets the definition)

Certified Sr. GD&T Professional

 

[KLineDesign]

I like bringing that up too: So, what are we arguing the drawing for right now? to help fabrication? to help inspection? to help procurement?....or all of these things which is the actual part definition. It can be difficult to bring the groups together, and often the change is written due to troubles in only one area.

It's unfortunate that a lot of these changes CAD is on the tail end of. We sometimes get consulted, but then through review other things get affected and we don't see the final change until it's approved. It's the way it goes I guess. Anyways, and I don't know why I didn't think of this sooner, but with the drawing that blew up yesterday (it really blew up today by the way as the dimensional changes are not mathematically possible - model discovered issues - though the calcs could've been done in 2D) I've adopted a new practice: Just think about sending it out for quote when stuck in these conversations. Some of the lame internal justification really gets highlighted when drawings and models are considered in this way.

Again, we don't always have that luxury of early consulting with changes, but I'm going try to get the good stuff in (i.e. see model, general surface tolerance, and tolerancing of only things we actually care about) much like "new files" if and when I'm ever consulted on a change in the future to try to help alleviate the internal customizations. If it forces new tools outside manufacturing or checking, so be it.

Thanks for chiming in guys, good talk! See you out there!

KLine