Any CAD users here?

[nando]

Shit, in my line of work a couple thou can be the difference between perfect mating and a no-go part. Some of the lazy machinists get grumpy with some of my tolerance callouts, but they are necessary.

heh, the tolerances on my project are like +/- 1/2" :p

[AndrewBird]

Running a part today that has a bore with +.0003 -.0002. No big deal.

[TwoJ's]

We had a customer at work a few weeks back that gave us a drawing with a dimension that was +/- 0. We all laughed at that one.

I have actually called out zero tolerance a few times at this current job (just under a year). It is often called for on the connectors I design as aerospace stuff absolutely has to be as small and light as possible. Designing fairly complex mechanisms for very small packages and out of crazy materials requires that some of the parts be extremely precise.

So do you work at a machine shop? What kind of work do you guys typically do? While it is fairly rare for me to call out zero tolerance, I regularly call for ±.001 or +.000/-.001. The machine shop I use for all of my prototype and some production machining has never had an issue with tight or zero tolerances (the grumpy machinists I referred to earlier are the ones in the machine shop at my facility... I have just given up on even trying to use them as they are slow and can only give me a part without flaws if it is pretty basic). The shop I use is not cheap by any means (I just entered a purchase requisition for 10 sets of connectors (just under 100 parts) very similar to the ones below ant is totaled to $45k. I basically have the ability to spend more than I could even try to on R&D prototyping and they know that, so they take care of myself and the other component design engineer.

I haven't worked with a ton of other machine shops, but the few that I have dealt with would all be able to produce parts with some zero tol call outs. It will obviously take more labor and likely more scrap to produce, so they adjust their price accordingly. With modern machines, it shouldn't be a problem.

Just for fun, below are some photos of a connector I designed and had prototyped. I think I posted a pic before, but this is the current rev that will go to production -- but in production all parts will be Ultem (PEI). All of the parts will be electroless nickel plated for EMI shielding aside from the inserts/insulators. The contacts on the PCB side are beryllium copper and were machined at the same shop. Those things along with the PCB inserts are intricate, but this is definitely not an example of something with super tight tolerances (sheet tolerance is ±.003 and only a few critical dims in the entire assembly are ±.0005.

This is the first project that I will be listed as the inventor on the patent. It basically doesn't mean much for me as the company owns the design, but I was pretty proud when signing the patent application.

[unevolved]

Yeah, but that zero had a tolerance of +/-0.005" ;)

[TwoJ's]

Maybe... all depends what I call out for precision. And obviously yes, no one will ever have a true zero tolerance.

[Dozyproductions]

I'm a little behind on all of this but will machine shops go from the drill to 3D printing?

[unevolved]

Maybe... all depends what I call out for precision. And obviously yes, no one will ever have a true zero tolerance.

Sorry, that was supposed to be in reply to the guy talking about the customer that supplied a print with +/-0.000.

[Thizzelle]

eh, washer dimensions were made up and I already had the selector housing piece in my hand with a set of calipers.

Its useful to put the time in to get it exactly how you want it, especially if you plan on making one. Also I will possibly make my own shifter and having this assembly helps.

I hear ya

How so? If you like to half-ass things, I can see that... but some people don't.

just extra work that is not needed. you don't need to draw every nut and washer to be able to see how it works or maybe you do.

[unevolved]

Adding in the nuts and bolts can be pretty beneficial, I do that for a lot of my stuff. I've got a good sized database built up of hardware that I've either drawn myself for specialty things or downloaded from McMaster, and I've got all the configurations set up to where SolidWorks automatically generates an EBOM for me. A lot of what I do is full cycle from conceptual design to final assembly, so the more steps I can eliminate the better.

[AndrewBird]

I have actually called out zero tolerance a few times at this current job (just under a year). It is often called for on the connectors I design as aerospace stuff absolutely has to be as small and light as possible. Designing fairly complex mechanisms for very small packages and out of crazy materials requires that some of the parts be extremely precise.

So do you work at a machine shop? What kind of work do you guys typically do? While it is fairly rare for me to call out zero tolerance, I regularly call for ±.001 or +.000/-.001. The machine shop I use for all of my prototype and some production machining has never had an issue with tight or zero tolerances (the grumpy machinists I referred to earlier are the ones in the machine shop at my facility... I have just given up on even trying to use them as they are slow and can only give me a part without flaws if it is pretty basic). The shop I use is not cheap by any means (I just entered a purchase requisition for 10 sets of connectors (just under 100 parts) very similar to the ones below ant is totaled to $45k. I basically have the ability to spend more than I could even try to on R&D prototyping and they know that, so they take care of myself and the other component design engineer.

I haven't worked with a ton of other machine shops, but the few that I have dealt with would all be able to produce parts with some zero tol call outs. It will obviously take more labor and likely more scrap to produce, so they adjust their price accordingly. With modern machines, it shouldn't be a problem.

Just for fun, below are some photos of a connector I designed and had prototyped. I think I posted a pic before, but this is the current rev that will go to production -- but in production all parts will be Ultem (PEI). All of the parts will be electroless nickel plated for EMI shielding aside from the inserts/insulators. The contacts on the PCB side are beryllium copper and were machined at the same shop. Those things along with the PCB inserts are intricate, but this is definitely not an example of something with super tight tolerances (sheet tolerance is ±.003 and only a few critical dims in the entire assembly are ±.0005.

This is the first project that I will be listed as the inventor on the patent. It basically doesn't mean much for me as the company owns the design, but I was pretty proud when signing the patent application.

A zero tolerance is impossible, it's like dividing by zero. You will always be able to measure it to a high degree and find that it is off of nominal. The only limitation is your measuring equipment. I mean, yeah, holding a +/-0 using a pair of calipers to measure the part is easy, but what about a micrometer that measures down to millionths? That is a much different story. And using a +/-0 tolerance purely for the fact that you want it to be the best it can be is just dumb. Use a tolerance that fits what you need the part to do. Pretty much nothing needs to be tighter than +/-.0001, or +.0000, -.0001 or visa versa.

The reason why machinists grumble at a +/-0 is not because it is hard to make an accurate part, but because you have no range for how tight a dimension needs to be. There is no "perfect" in the machining world, it's always a question of how perfect?

I do work at a machine shop. Most of what we run is aerospace stuff, but we do some medical and then any other job that might come through the door. Here is a connector similar to yours, only a bit smaller:


For reference, the squares on the table are 1/2". Most of the dimensions are +/-.003, but some are +/-.002. This part is very easy to run.

We also recently did some thruster nozzles for satellites. Roughly 2 inches long, they were made from Inconel, took 8 hours a part to run and wore all the tooling out after one part. A lot of the dimensions on these were +/-.0002 as the flow through the nozzle was precisely calculated.

[AdironRider]

Bump bump bump it up!

Anyone have any experience with DraftSight? Specifically for Mac but I assume they are overall pretty similar.

Getting ready to go solo again in the knifeworld, and in my previous startup I did it all by hand on graph paper and paid a couple extra bucks per unit for my OEM guys to make one up for me, looking to do it right this time.

Anyone have any experience, good books for a newbie? I understand the engineering but know nothing about CAD. Draftsight seems to be highly recommended in the knife world, but I assume thats because its free unless for commercial use.

Anyways, Nitto arent you working for leatherman? Would really be interested in your input.

[LEANE30]

I'm in energy infrustructure design, i don't often get to deal with truly being creative with component design, i just model what the cut sheets of vendors that can't supply 3d models depict and then design larger process assemblies and facility/site layouts and interdisciplinary model incorporation/auditing/oversight. It's damn true about so many engineers, really really smart, but they have no idea about efficient design for hands-on construct-ability. I have become accustomed to working with my teams of engineers quite closely for designs, it can be fun and enlightening at times,... other times it's frustrating when i get left out of the brainstorming loop as often the engineers at my office often preoccupy themselves more with project managing and client RFIs than actual engineering.

for me, calling out tolerances often seems to be more of a lesson in futility than actually useful as field contractors seemingly always deviate to some degree, i design in areas for anticipated field adaptions and vendor introduced variations, out-thinking how others might manage to bugger it all up.... i'm starting to go through as-built processing on 3 separate projects i've designed and then issued for construction over the past three years. On the upside, in the meantime i've been able to dabble with Cyclone and laser scan point clouds a bit, which blows traditional survey COGO point data lists out of the water in pretty much all aspects. Might make the jump to ACAD 2015 sometime soon. 2014+ has some real power now for crunching point cloud data. Recently I find my workload pulling me more toward Civil3D lately as a means of efficiency for me to decrease the dependency of projects on having competent subcontractors. it's not by choice, but necessity. ...Top of concrete elevations can only be accurate to a point of usefulness, steel & pipe dimensions, etc., all subject to temperature expansions, ground heave, frost swell, etc.... you can only be precise to a point of usability. /rant over.

Those connectors looks cool. given access to machine shops, i'd be designing things for my m20, LOL

[lambo]

Solidfails exclusively at work and unigraphics at school. Been having a lot of fun deigning a waveguide system for an aircraft at work, though.

[bmwman91]

Creo/Elements at work, waiting to get the current project into production before (hopefully) making a switch to Creo Parametric. Creo/Elements feels freaking archaic to run compared to SW and Creo Parametric, but it gets the job done.

[george graves]

TwoJ's - Nice job on the patent. There are a few in my family. It's a nice feather in your cap! Well done!

(but insert rant on how patent law now-a-days-is-bubkis https://www.youtube.com/watch?v=Z7BL1O0xCcY )

Question: Why the custom connector? I ask cause that looks like off the shelf stuff you see all the time (and you may not be allowed to answer that - so that's ok to)

This is the first project that I will be listed as the inventor on the patent. It basically doesn't mean much for me as the company owns the design, but I was pretty proud when signing the patent application.