Question for draftsmen and pattern makers

   / Question for draftsmen and pattern makers #11  
It always takes new engineers a few years to learn to tolerance parts based on what they  need and not what Manufacturing can hold. Sometimes, it just isn't critical so "give them a garage door to ride the bike through". Then Manufacturing has options, costs go down, and (best of all) your phone doesn't ring!
 
   / Question for draftsmen and pattern makers
  • Thread Starter
#12  
Exactly. Too many just default to .xxx and ±.005" block tolerances on everything, when ±.01" or even ±.05" might be totally fine. I'm pretty careful with that these days, since every dollar spent on trying to hit a tolerance is either coming out of my own pocket, or going onto the product cost and making us a less competitive option.

But I'm also not shy to put ±.002 or tighter, when the tighter spec allows us to increase our spec'd performance minimums sufficient to cover the added cost or yield hit. Tolerance analysis simulations take a long time, but can be easily scheduled to run over a weekend, so it takes very little skin off my back to run them.
 
   / Question for draftsmen and pattern makers #13  
I think I understand the question now. Some of the holes have to be oriented in an exact manner relative to some of the other holes while others don’t. I’d say the way you are doing is as good as any. I’m a retired land surveyor and spent my career working for a civil firm. I did a lot of drafting but nothing like you are dealing with. I have worked a little since I retired but it’s been 2 years since I picked up a drafting pen computer mouse.
 
   / Question for draftsmen and pattern makers #14  
Exactly. Too many just default to .xxx and ±.005" block tolerances on everything, when ±.01" or even ±.05" might be totally fine. I'm pretty careful with that these days, since every dollar spent on trying to hit a tolerance is either coming out of my own pocket, or going onto the product cost and making us a less competitive option.

But I'm also not shy to put ±.002 or tighter, when the tighter spec allows us to increase our spec'd performance minimums sufficient to cover the added cost or yield hit. Tolerance analysis simulations take a long time, but can be easily scheduled to run over a weekend, so it takes very little skin off my back to run them.
I do that all the time to make sure not to overspecify tolerances. Our standard tolerance block has 4 entries in it:
0.0 +/-0.1
0.00 +/- 0.01
0.000 +/- 0.005
0.0000 +/- 0.0005

(I forget the exact numbers on the 1 and 2 place, but you get the idea)

So I control most things just by the number of decimal places on each dimension. Everything default to 3 places, but I will knock them down to one or 2 everywhere I can. If I need 1/2" diameter rod for a hand tool I just drew up, I list is as 0.5 not 0.500.

So perhaps you can modify your standard tol block to give more options and save yourself time down the road.
 
   / Question for draftsmen and pattern makers #15  
For years I believe JD Horicon drawings were typically using 3 sigma tolerancing (99.7% in control). Geometric Tolerancing is now fully embedded in Engineering and does have its advantages: bonus tolerances.

For example, a true position callout controls location, orientation, and form on a feature of size... but also allows for  bonus. GD&T callout talking: "Ok Manufacturing guy... you've got choices. If you can hold size, orientation, and form on this hole (or pattern of holes) we'll give you a bonus tolerance on location (cause we know it will still assemble). You decide and route the part the most effective way."

I'm retired 10 years but I believe they're honestly past 4 sigma tolerancing now (99.99% in control) on there way to the holy grail: 6 sigma tolerancing.... 3.4 defects/million opportunities! Good stuff.
 
   / Question for draftsmen and pattern makers
  • Thread Starter
#16  
I do that all the time to make sure not to overspecify tolerances. Our standard tolerance block has 4 entries in it:
0.0 +/-0.1
0.00 +/- 0.01
0.000 +/- 0.005
0.0000 +/- 0.0005
Yep, I have the same, except just 3 lines in my title block. Sheet metal gets the x.x thru x.xxx, and precision machined parts get the x.xx thru x.xxxx, with the 4-digit dimensions typically at ±.0002.

So I control most things just by the number of decimal places on each dimension. Everything default to 3 places, but I will knock them down to one or 2 everywhere I can. If I need 1/2" diameter rod for a hand tool I just drew up, I list is as 0.5 not 0.500.
Same.

So perhaps you can modify your standard tol block to give more options and save yourself time down the road.
Oh, I take it a step beyond that! I don't even produce fully-dimensioned drawings, only inspection drawings, on machined parts. Essentially, we provide a 3D model (usually .step) with every part, with implied ±.005 global tolerance on the model. Then the drawing only receives hole call-outs and dimensions on anything that differs from ±.005, either because it needs to be tighter or because we believe money can be saved by loosening it. In most cases, multiple operations on a single setup don't benefit from looser than ±.005, that only becomes a cost advantage when having to move the part to a new setup.

Once in awhile, I'll run across a shop that still wants fully-dimensioned drawings, and then I have to decide if it's worth the effort to even work with them. Working with shops that import our model directly into their CAM software yields fewer mistakes, saved time in CAD, and even much more saved time on inspection (both at fab and incoming to assembly). For these reasons, I really shy away from shops requiring fully-dimensioned drawings the last 25 years, although we still have a few, mostly screw machine manufacturers, sheet metal, and plastics.
 
   / Question for draftsmen and pattern makers #17  
Yep, I have the same, except just 3 lines in my title block. Sheet metal gets the x.x thru x.xxx, and precision machined parts get the x.xx thru x.xxxx, with the 4-digit dimensions typically at ±.0002.


Same.


Oh, I take it a step beyond that! I don't even produce fully-dimensioned drawings, only inspection drawings, on machined parts. Essentially, we provide a 3D model (usually .step) with every part, with implied ±.005 global tolerance on the model. Then the drawing only receives hole call-outs and dimensions on anything that differs from ±.005, either because it needs to be tighter or because we believe money can be saved by loosening it. In most cases, multiple operations on a single setup don't benefit from looser than ±.005, that only becomes a cost advantage when having to move the part to a new setup.

Once in awhile, I'll run across a shop that still wants fully-dimensioned drawings, and then I have to decide if it's worth the effort to even work with them. Working with shops that import our model directly into their CAM software yields fewer mistakes, saved time in CAD, and even much more saved time on inspection (both at fab and incoming to assembly). For these reasons, I really shy away from shops requiring fully-dimensioned drawings the last 25 years, although we still have a few, mostly screw machine manufacturers, sheet metal, and plastics.
Pretty much the same here. It's all STEP or native CAD sent. All of our prints have only inspection dimensions and a few ref dims like overall size, etc just to help you get oriented. Some prints have only one or two dims on them. When I get someone who wants a fully dimensioned print it is time to say bye bye. Been a long time since anyone has asked for that. Prints often come quite late in the process for us. We mostly do molded plastic parts but a lot of mold makers want the print at time of quoting to see if there are any crazy tolerances or other difficulties, so we often end up scrambling to make prints for that.
 

Tractor & Equipment Auctions

Central Pneumatic Sandblasters (A45336)
Central Pneumatic...
Wolverine 72in Skid Steer Brush Cutter (A47809)
Wolverine 72in...
Allis Chalmers WD 2WD Tractor (A47809)
Allis Chalmers WD...
Allterrain Self Loading Mini Crawler (A47809)
Allterrain Self...
2009 PETERBILT 386 FLAT TOP SLEEPER (INOPERABLE) (A47001)
2009 PETERBILT 386...
5 FORKS W/ QUICK COUPLER (A47001)
5 FORKS W/ QUICK...
 
Top