Drawing Standards

[zoey]

My question is pertaining to Drawing Standards. I have been making chips for ever and customer drawings are less than to be desired. Now with solid models it seems customers can't make up their minds as to what they want the shop to do. I am tired of trying second guess these second rate designers. Has anyone come up with any kind of current standards for everyday questions that everyday lazy machinist's have to deal with?

 

Zoe

[crazy^millman]

Not really need to make up a template you can all decide on. If you are in a ISO or AS9100 shop then this is required IMHO. Then depending on the nature of the parts you may have to be 100% detailed or as little as possible. Think everyone involved needs together and come up with some standards and go from there. That way you make sure everyone on the same page.

 

HTH

[Guest CNC Apps Guy 1]

The "standard" I'm seeing these days is that you get a solid model. If it has holes and or critical features, the blueprint defines the features and positional tolerances. All other features and relationships not noted are a "standard" tolerance call out (like .010 True Position A-B-C), etc...

 

Otherwise the model is gospel. Puts the onus of correctness on the engineering dept. - where it belongs.

 

Anymore, PC's are inexpensive enough that you can have a few out on the shop floor and you can buy a couple of seats of Mastercam Design LT and your machinists can interrogate the models themselves when necessary.

[Matthew Hajicek - Singularity]

[RANT-ON]

 

Here's what I usually get: The part was obviously modeled in metric, then dimensioned in English, and rounded to use the number of decimal places to indicate tolerance. That way, what was .3937 (10mm) with a +/- .010" tolerance becomes ".39". Acceptable parts should range from .3837 to .4037, but according to the print it's .3800 to .4000, which means they're accepting parts that should be bad and rejecting parts that should be good. That just chaps my hide.

 

[RANT OFF]

[Sambo]

I feel your pain. I had to deal with something similar to this a couple months ago. I redrafted in Pro/E a print from 1982, that had been marked to death. Some of it, xerox of xerox of xerox was illegible. The QA manager kicked it back to me three times, each time saying, "it doesn't meet company standards."

But when I asked what the company standards are, he didn't have an answer. Plus, I'm looking at several customer prints, each customer with their own standard. Oh well.

[YoDoug®]

We have a few engineers that use JIS standards. The solid models are created to the nominal and the print tolerances for clearance/press, etc. Example; 7MM shaft and 7MM hole in solid models. The print will Show Shaft as 7MM -.025/-.05 tolerance for clearance. Without the blueprint the machining to the model will be incorrect.

 

We also have Metric ISO prints, and Inch ANSI prints.

 

If I had my way I would prefer the JIS standard metric.

[Bruce Caulley]

The method that James described is quickly becoming standard. The only condition that I would make is that ALL un-specified features are modeled to mid limit.

 

Bruce

[gms1]

Bruce if you could get the engineers I work for to model like that it would make my year

 

I have never, ever received a cad file modeled to nominals and in some cases completely different then print dimensions.

[3R!< W3NG3R]

Anyone else also get the "its real easy. will only take a couple of hours."? Whenever I hear this, I feel like running away.

[YoDoug®]

quote:

---

if you could get the engineers I work for to model like that it would make my year

---

In ProE if the solid was created with tolerances then you can change the output of the solid to be nominal, low, high, best fit(if it was created in an assm). The hard part is getting the engineers to model solid with tolerances. It makes a huge difference if you want to use the models for analysis like kinematics, FMEA, fit checks, etc.

[Guest CNC Apps Guy 1]

quote:

---

...The only condition that I would make is that ALL un-specified features are modeled to mid limit...

---

+1000

[mcpgmr]

code:

---

 ...The only condition that I would make is that ALL un-specified features are modeled to mid limit... 

---

I totally agree as well. Just went through this with a part. The solid model was constructed to nominal print dimensions. The print specified +.02-.00 on most of the wall thicknesses. I assumed (xxxx) that the model was at mid tolerance and programmed by extrcting edge curves and went with it. After Q/C checked the first part I was asked to change the program. Part was still good (tool deflection can sometimes be you friend) but it just wasn't the right way to do it. The thing that really sucks though is I usually Verify against the customer model as a final check of my programming and it really didn't represent the real part. The one thing that the customers designer did right was give me .130 R corners in the pocket features so I could interpolate with a Ø.250 e.m.