Drawing/Model discrepancies disrupting our schedule

[mhajicek]

If you commonly have discrepancies between print and model which cause delays, it might make sense to put a note in your quotes that such discrepancies may effect lead times. That could get the attention of management at your client.

 

[Donkey Hotey]

Though I guess all this simply creates more confusion for the OP. All I will say is that I have found many times more mistakes in solid models than I have on drawings, provided the drawings are intended to be used to make the print, and not non-technical sketches such as dstig is describing.

I guess at the end of the day, it is best to establish with the customer up front whether the drawing or the model is the "master" document.

Yes, not only is there the expectation that the parts will be made on modern CNC with modern CAM software but, there is also the responsibility on the engineering side that no funny business takes place between the 3D and dumping 2D views on a drawing.

I know on Catia that anything that got monkeyed with on the drawing changed color and was no longer generative (tied back to the 3D model). When drawings went through review and drawing check, anything like that would bounce the package back to the engineer. Not allowed, period. Make it match.

On the other hand, in our corporate BS world, the IT people didn't understand or want to manage things like thread standards. So they stripped out all of that information from the software. I couldn't just grab a 1/4-20 hole spec out of a drop-down menu on the Hole wizard. We were stuck between modeling the hole undersize (for the tap drill) or modeling it at nominal size (1/4"). They settled on nominal size so when the programmer sat down, they had to get the hole spec from the drawing.

Flip side of this: I think everyone considers a drawing to be a contract. If it's covered in 3 place dimensions, then everything needs to be checked and verified, one at a time. The goal was to say: we all agree that CNC parts are good enough for common manufacturing and assembly problems, let's save us all a bunch of time and headache and distill this down to what we really need and notate what isn't common part tolerance.

I personally find GD&T to be meaningless, overly complicated garbage. It's engineering legalese to attempt to outsmart poor manufacturing processes. It sucks on both sides of the process and it needs to die.

 

[guythatbrews]

But no way of telling that from just the Solid. Had to bring those back In re-drill and Heli-Coil them. Not the end of the world. But still cost time and money.

Exactly! And it costs time and money whenever the part is made by a new vendor, and often by the same vendor.

The design intent should be included in the model. That intent can then easily be imported into a drawing, creating an associative system. With the exception of complex surfaces, which may be very difficult to define with a drawing, the entire part should be defined by the drawing.

With a comprehensive drawing everyone is on the same page. Designer, manufacturer, quality. Without a drawing everyone interested in the part must search the model for pertinent details, each and every time those details are needed! And I for one sure don't know how you extract feature tolerances from the model.

I absolutely hate having to take time to query a model when a drawing is a straightforward method to communicate infornation for all to see forevermore.

Lack of a drawing or a slack and incomplete drawing is simply the designer offloading his responsibilty to ALL that follow. The lazy, unprofessional designer is the only one who saves in this scenario.

 

[Donkey Hotey]

The model, OTOH, defines the geometry for non-trivial contours that a drawing cannot (reasonably), and all of the various sundry dimensions of the part that would make a drawing overly complicated. Something like this, that I'm working on today:

View attachment 416302

If the drawing had to call out dimensions for every single contour, it would be super-busy and hard to read.

That is an awesome example. We can look at that model and quickly understand what they want. We can refer to a minimum dimensioned drawing and see what might not be obvious.

Anyone fighting the transition to reduced dimension drawings is already falling behind. The thing that will replace the drawings altogether is already coming (and I think its out there in current Catia and NX): annotating the actual 3D model. It shows as floating text call-outs, pointing directly to those features on the 3D model. The hole specs are built in to the features. The CAM software reads things like a thread spec and creates the operations and locations for you.

The only thing holding it back: just like all the push-back we see on the manufacturing side, the engineers are also pushing back. There are old engineers who don't want to do it differently either.

Most of those old guys started on drafting boards and fought the transition to 2D CAD drafting. Then they fought the transition to 3D wireframe design and "cleaning up" 2D views on a drawing. Then they fought the change to solid modeling because the methods were completely different. Then parametric solid models came along and they didn't understand why you'd want to store design parameters in the model. Then they tried to stop computer aided analysis because they always did things as strictly structured statics problems.

Most of the people on here, reading this don't relate because by the simple act of being on the internet and participating on a web forum,. you're probably not on that side of the bell curve.

 

[gregormarwick]

I personally find GD&T to be meaningless, overly complicated garbage. It's engineering legalese to attempt to outsmart poor manufacturing processes. It sucks on both sides of the process and it needs to die.

IME, every person I've encountered who holds that opinion (and there have been quite a few) are ignorant, arrogant fools who think they're machinists.

I would hope that you'd prefer to be excluded from that group?

GD&T exists to serve a specific, critical purpose, that otherwise has no way to be communicated regardless of whether it's on a drawing or on a solid model.

Geometric accuracy is equally as important as dimensional accuracy, and without GD&T is much more difficult to convey and to prove.

To the original question, here is how I handle it. I'm sure I'll get raked over the coals by some for this, but whatever.

I exlusively program from a solid model, unless the part is extremely simple. However, it is never the customer supplied model - I make the solid model from the customer's drawing. Sometimes the customer will supply a model and a drawing, and I will occasionally reference the customers model if something doesn't add up, but I do not use it for programming. 99.9% of every part that crosses my desk is modelled by me as the first step in manufacture.

A side effect of this method is that engineering queries almost always become apparent during this phase, and not when the job gets to a machine.

I can count on one hand the number of times I've been given only a solid model, so I don't really count that at all.

 

[guythatbrews]

I exlusively program from a solid model, unless the part is extremely simple. However, it is never the customer supplied model - I make the solid model from the customer's drawing. Sometimes the customer will supply a model and a drawing, and I will occasionally reference the customers model if something doesn't add up, but I do not use it for programming. 99.9% of every part that crosses my desk is modelled by me as the first step in manufacture.

This!

I don't do mold work and do very little surfacing. The stuff I make has close tolerance features. Who knows where in the tolerance band the supplied models fall? There is no rhyme or reason. How the heck do you use a supplied model if you don't know this vital information. So I make my own model just as gregormarwick does.

Maybe it's just the industry some of you guys work in, but I can't imagine making a part from a supplied model and assuming all is good without knowing any basic sizes or tolerances. How the heck do you even provide an inspection report? It sure wouldn't fly with my customers. 100% dimensionals required.

 

[Donkey Hotey]

IME, every person I've encountered who holds that opinion (and there have been quite a few) are ignorant, arrogant fools who think they're machinists.

I would hope that you'd prefer to be excluded from that group?

You can call me whatever you want as I'm here driving a wooden stake through GD&T's heart. I self manufacture now. I embed everything in the solid model that I will need, except what can't be stored there: actual material, fits and tolerances of specialty holes, etc. I even include a manufacturing sheet to document the workholding and any other notes (if necessary). What I don't need or want is meaningless dimensions that are safely stored in the 3D definition.

I've had too many cases where GD&T was overused and overly complicated the part, to the point of costing unnecessary money. Engineering management at my last job pushed back on this. It took having engineers with their own machines to prove that both Engineering and Manufacturing in a major company were wrong. We could bang out parts over and over, that fit together and exceeded the requirements. Manufacturing would produce crap and look me in the eye saying they couldn't do better without driving up cost. They were lazy.

One of the managers has/had machines at home and might read this someday. He had aluminum legos on his desk that worked (clicked together). He knew what was possible. He couldn't get that out of most of them.

GD&T exists to serve a specific, critical purpose, that otherwise has no way to be communicated regardless of whether it's on a drawing or on a solid model.

Yes, if I have a rocket nozzle with a metal-to-metal sealing joint and it must meet certain flatness and material finishes to achieve an acceptable seal, that's the exception where GD&T applies. Nothing says I can't use it on the drawing for that surface.

I'd just like to end the practice of calling out every meaningless feature with datums and min and max material conditions and whatever.

At the same time, I should not ever get a part with a 0.250" hole when I asked for a 0.257" hole. I shouldn't have to specify +/- 0.002 on that dimension to get them to use the proper drill. I shouldn't have a GD&T position call-out to not have them drill it on a knee mill while counting ticks of their hand-wheel.

I exlusively program from a solid model, unless the part is extremely simple. However, it is never the customer supplied model - I make the solid model from the customer's drawing.

Not all customers are the same. I totally understand where you're coming from but the designer sucked if you have to recreate the solid. This is a culture shift that is slowly happening on both sides. You should absolutely be able to trust the 3D. That should have been the 3D that went into the assembly and had all the (presumably) analysis done to it.

If you're a shop, you live in a reality of having to manage the shortcomings of your customers. I totally get that. I'm talking about where things are supposed to be headed.

Sort of related: we had a design engineer that was one of the "didn't get it" crowd. If we were modeling a car wheel and tire, where would we start? Maybe at the mounting flange? Or maybe a 2D profile of the rim? We'd put the mounting flange at the origin and do a body of rotation?

Not this guy. He'd start modeling from the valve stem. Or from one of the tread blocks. Then he might rotate the tread block and make 180 instances of the tread block. Then somewhere in there he'd create a three-point plane and try to define that as the mounting flange. The origin of the model is up there on the corner of a random tread block and you're saying WTFO?

Then he'd rig his new wheel to a 3D model. Use the mounting flange as a mate? Heck no. The tire is 10" wide so he'd constrain it from the tread and put a 5" dimension to center it. Use the axle flange as the reference? No, not that either. The flange is 2" from the housing so he'd dimension to the housing and make the total 7" from the axle housing to the tread block. No rhyme or reason to it.

Everything he ever created was referred to by his name: the Barry coordinate system. We all knew.

 

[Donkey Hotey]

Adding on to my previous post:

I learned Catia V5 in 2001 or 2002. It was not only mature at that point but, I was one of the late adopters (because the company absolutely would not be pried from their legacy mainframe CAD). That was 21 years ago. My seat of Mastercam is 2005. That was 18 years ago and it was hardly cutting edge. My Haas was delivered almost exactly 19 years ago.

What I'm suggesting isn't cutting edge. It's almost an entire career ago. How many lifetimes does the industry need to 'transition' from drafting boards, their outdated standards and hand wheel accuracy?

I haven't used a single drafting tool professionally since 1985. Haven't counted marks on a hand wheel since 1984. Gene Haas created an entire company, built an empire and has semi retired to play with his F1 team in that time. We're still strapped to GD&T designed to deal with 1970s capabilities?

 

[gustafson]

GD&T was meant to make it so you did not need to overspecify tolerances to get the part you want, but engineers who think everything should be to .0001 use it to overspecify dimensions, which is why machinists hate it
The solid model with a helicoil should be modelled with a helicoil thread, not a #6. Model was wrong customer pays to fix it. Lucky it wasn't the other way round

 

[VTM]

GD&T was meant to make it so you did not need to overspecify tolerances to get the part you want, but engineers who think everything should be to .0001 use it to overspecify dimensions, which is why machinists hate it
The solid model with a helicoil should be modelled with a helicoil thread, not a #6. Model was wrong customer pays to fix it. Lucky it wasn't the other way round

Yep, I've got a drawing sitting on my desk to quote right now. .001 Position callout to two Datums. Over 18 locations. How tight you think you have to hold location from each Datum to pull that off?

And yes, customer paid for the redrill and Helicoil. And If It was the other way around they would have paid for the job twice.

 

[guythatbrews]

I embed everything in the solid model that I will need, except what can't be stored there: actual material, fits and tolerances of specialty holes, etc.

A problem right here. Actual material, fits and tolerances of specialty holes, etc. should be in the model. If you are making design, the model and the part, you can be you. But that is not really what's being discussed here.

We did a lot of work for a guy that was great at devising complicated mechanical assemblies, but the worst at documentation. He'd get the visuals down in the model but add all the pertinent details in the drawing. Drawings and models never matched, and neither was completely correct. A never ending pain in the ass. I tried to teach him how to make correct models and detail the models. Then import everythjng into the drawing. Nope, not a chance in hell. Sad because it is so much easeir and better. He got sick and died and his legacy is a bunch of designs that are riddled with mistakes.

 

[Donkey Hotey]

A problem right here. Actual material, fits and tolerances of specialty holes, etc. should be in the model. If you are making design, the model and the part, you can be you. But that is not really what's being discussed here.

Except that the designer has to pick a number for the feature. Should they model it at nominal size? Max? Min? I try to keep the +/- 0.005" number in mind and model to the center (exactly as I'd machine it). If it's a bearing bore or a dowel location, I model on the minimum side but, notate the tolerance on the drawing. Yes, I can call out a material but, the material might be a casting or a forging. The 'material', with the specifications needs to be on the drawing because (as of right now) I'm not aware of anywhere else to store that in the 3D that survives converting to STP format.

We did a lot of work for a guy that was great at devising complicated mechanical assemblies, but the worst at documentation.

Yeah, I'm as confused as you. Every design software I've used from the past 25 years has been linked. Opening the drawing forces an update from the 3D model. If the drawing wasn't updated, it showed an error state.

I'm not at all arguing that your guy should have been forgiven. Nope, the bar got raised and he needed to meet that expectation too. Design and embed what you need in the 3D. That should reflect on the 2D. Out of scale dimensions were not allowed and would show up in a different color. I don't understand people too lazy to update their 3D. The 3D should communicate the same message to our CAM software. No discrepancies allowed.

 

[gregormarwick]

We're still strapped to GD&T designed to deal with 1970s capabilities?

This sentence makes no sense. 1970's capabilities of what? I really don't want to cast aspersions but it seems increasingly like you don't understand the difference between dimensional and geometric requirements of a component.

If I'm wrong and you know what you're talking about, you should be able to articulate some rational reasons why GD&T sucks - there are a few. And if you have some ideas of better ways to define geometric requirements I'm all ears. It is absolutely not a simple or intuitive system.

I wholeheartedly agree that working with GD&T when either party doesn't fully understand its application is a painful experience. I have seen my fair share of drawings where the engineer covered his drawing in GD&T symbols with no understanding of how to apply them or what they mean.

 

[SeymourDumore]

I exlusively program from a solid model, unless the part is extremely simple. However, it is never the customer supplied model - I make the solid model from the customer's drawing. Sometimes the customer will supply a model and a drawing, and I will occasionally reference the customers model if something doesn't add up, but I do not use it for programming. 99.9% of every part that crosses my desk is modelled by me as the first step in manufacture.

Pretty much the same here, except the more complex models I try to interrogate first and make a direct edits where needed if possible.

As far as models without drawings, thankfully only had a few requests of that ilk.
Couple were for stamping dies to be EDM-cut and everything was modeled appropriately for proper fit so there were no questions to be had.
Then there were a few where I was told by the customer that they will not waste their time creating a drawing.
Response was that I will not waste my time making the part then.

 

[Donkey Hotey]

This sentence makes no sense. 1970's capabilities of what?

Of manufacturing and the expectations of dimensional and geometric qualities that can be achieved. I have 1980s manufacturing magazines that brag about how "CNC technology can now maintain tolerances better than 0.010"". Forgive me if that's not impressive anymore. Technology moved on and we got worse.

Here's another example: supply a drawing and 3D model to manufacturing. They inexplicably screw up the location of basic mounting holes. Walk over and talk to them. Eventually get out of them that they were at the end of their shift so they broke down the machine and went home without the holes. They did them the next day. Stupid but, okay. Why didn't they indicate it back into the machine? "well, you didn't specify a tolerance and I can hit the +/- 0.020 in the title block without checking."

Was that lazy or stupid? The minute someone puts 3 place GD&T requirements, it gets sent off to the CMM and spends more time in inspection than they spent trying to figure out the crappiest way they could make it and squeak by.

Another: try to give manufacturing some breathing room by specifying wider flatness and parallelism zones on what should be a datum surface. The intent is to accept something that maybe twists during manufacturing but, will acceptably flex enough on assembly to not matter. There's maybe a center bored hole that has to be some squareness to the surface.

What we get back is something where they took advantage of the wide tolerance to hit the raw stock with a DA sander (because the zone was wide enough to include everything on that surface). Then they did a shitty bore job on a knee mill and barely made squareness because "your drawing allowed that and it meets the specification." Yes, but no. I didn't want a piece of crap to deliver to our customer.

These would be examples of not meeting "best shop practice". I agree that's an ambiguous note but, I don't know any other way to communicate: don't produce crap. It's the 21st century. Doing this in a CNC isn't a hardship.

 

[guythatbrews]

Except that the designer has to pick a number for the feature. Should they model it at nominal size? Max? Min? I try to keep the +/- 0.005" number in mind and model to the center (exactly as I'd machine it). If it's a bearing bore or a dowel location, I model on the minimum side but, notate the tolerance on the drawing. Yes, I can call out a material but, the material might be a casting or a forging. The 'material', with the specifications needs to be on the drawing because (as of right now) I'm not aware of anywhere else to store that in the 3D that survives converting to STP format.

Yup. Exactly why I advocate for a comprehensive drawing.

Put all design intent in the model. Import that intent seamlessly into the drawing by the designer or the detailer. Changes and revisions are made in the model and the drawing is automatically updated. Everyone has complete and correct information. Great!

This can still be done for "partial" drawings. But when the object of the designer is to create less work for himself, why bother to do it right?

My apologies to all who work in industries where a model with no drawing is the norm. My views are informed by my work experience, where complete drawings are the norm. That, and I'm sick to death of guys that think being able to create a 3D model is akin to being a competent designer and detailer. The model is a very good start, but it is not the finish line!

 

[guythatbrews]

Then there were a few where I was told by the customer that they will not waste their time creating a drawing.
Response was that I will not waste my time making the part then.

Love it!

 

[Donkey Hotey]

This can still be done for "partial" drawings. But when the object of the designer is to create less work for himself, why bother to do it right?

This might be the part being missed: the intent is not to make less work for the designer. It's to make less work for the manufacturer. It's to relieve all the inspection nonsense for features that should be pedestrian to create.

Make the part. Don't suck. It will work fine. We don't go to a restaurant and specify "Don't give me expired dairy products or rotting vegetables." It's part of the expectation.

 

[guythatbrews]

This might be the part being missed: the intent is not to make less work for the designer. It's to make less work for the manufacturer. It's to relieve all the inspection nonsense for features that should be pedestrian to create.

Maybe. If you and your customer have an agreement. But I hate this sort of thing. So anything not directly specified can be anything? Probably not. So how far can it be off? Open to interpretation it is! And I will not second guess if that .257 hole is clearance for a 1/4 bolt. By the way that is way too close for mating parts with more than one hole!

Reminds me of the guy that puts something like .05 max chamfer on the drawing then rejects the part because the chamfer is .010.

 

[Donkey Hotey]

Maybe. If you and your customer have an agreement. But I hate this sort of thing. So anything not directly specified can be anything? Probably not.

Again: we may be talking past one-another. There's a drawing note specifying tolerance to the 3D model. I gave the example of 0.005". It could be wider if the part is larger and that's acceptable.

There would be other notes to suggest something like break or chamfer all sharp edges, maximum 0.030". Again, something wide and easily achievable.

So how far can it be off? Open to interpretation it is! And I will not second guess if that .257 hole is clearance for a 1/4 bolt. By the way that is way too close for mating parts with more than one hole!

Not in my experience. I had two mating parts with twenty tapered protrusions (basically conical dowels), all the way around a 24" diamond-shaped opening in aluminum. The towers were ball-milled on a "supposedly" shitty Haas VF-5XT. The mating bores were also ball-milled on the same machine.

The parts mated perfectly in one orientation and were out maybe 0.0005" if spun 180 degrees. It was enough to jam the assembly in the second orientation but just slightly.

It was the first time we discovered the machine wasn't completely square. A 0.257" hole and 1/4" fasteners would have been so sloppy we'd have never known. It would have worked with 0.250" holes if that had been the design. The experts said we couldn't do it and we did it on the first try. That Haas did lots of things we were told wouldn't work.