Manufacturing/ Machining to model - Capable?

There has been some question over the past couple of years atmy company about machining parts to model and I wanted to survey atleast people in the SolidWorks forum about whether or not they doand what industry do you work for. I am not talking about doingaway with drawings completely. Obviously functional faces that aredatums and GD&T must go somewhere and I absolutely do not thinkthat the dimxpert is advanced to the point we can put alltolerances in the model. But I was wondering with the capability ofcnc mills/lathes, cast to model, rapid prototyping and CMMs is thecomplexity of parts grown to the point where it is prohibitive todo FULLY defined drawings?

So copy the following and answer yes or no (sometimes=yes):

Machine to model:
Manufacture to model:
Inspect to model:
GD&T:
Fully detailed drawings:

SolidworksDrawings And Detailing
Answers 39
AW
AW 2009-01-15 Best Answer
We do Minimum Content or Reduced Dimension drawings all thetime. Have operated this way for several years.

We have software in place for our CNC mills, Wire and Sinker EDM's,Laser and Water Jet and our CMM's, etc to allow us to operate thisway. Works very well and keeps us competitive in today'smanufacturing world.

I have several posts on my blog on the subject.

Cheers,
CC
CC 2009-01-15
All our cast parts are to model (with some dimensions on thedrawing). Most are "swoopy shapes", driven either my spline basedsurface models or imported parts. 10 years ago it would have beendone to a physical pattern, now it is just to a "virtual"pattern.
MC
MC 2009-01-15
Our toolmakers do the square up of details and do themounting holes and simple machining. Our CNC dept. does the 3Dmachining per model. Our prints usually just have a note stating"Machine per CAD data" pointing to the 3D areas. I will includedims. like an overall height or some other ref. checks thetoolmaker can do at his bench after CNC.
DA
DA 2009-01-15
for in house machining that works to shop "standards" i thinkthat programming a part from a model and supplying a drawing thathas critical feature dims can work. yet if any inspection is doneother than linear dimensions e.g GD&T, then i still think youneed a fully dimmensioned drawing for quailty control purposes.

if one has to send a part out for quote or outside fab, a fullydimensioned drawing is necessary to protect both the engineer andmanufacturer. there are so many CAD and CAM platforms that toexpect all the tolerances etc to be properly ported and displayedusing only the 3d model is pretty lofty. Also, considering that SWcannot even follow published standards andthat dimxpert an oxy,leads me to believe that this is not a path SW intend tofollow.
SF
SF 2009-01-15
Minimum Content or Reduced Dimension drawings: Awesome techlingo!
AW
AW 2009-01-15

Paul Lemke wrote:

 

Minimum Content or Reduced Dimension drawings: Awesome techlingo!

And the ASME standard for this is....

ASME Y14.41-2003, Digital Product Definition Data Practices

A couple other terms you will see and can search Google on are

Simplified Drawing
Model Based Definition

Cheers,

SF
SF 2009-01-15
Yeah I took a look at that a while ago and it wasn't tooconvincing. Do you use it?
AW
AW 2009-01-15
Yes, we use the concepts of the ASME standard for our models.

Search Google on some of the terms I have thrown out and you willfind a number of good bits of info on the subject.

Cheers,

SF
SF 2009-01-15
Boy Have I come a long way... Here you go.

As a former machinist, I've used both paper and software to produceparts. Models are great to machine from although there is no knownCAD/CAM software that I know of to generate the tolerances forcommon machining practices. Putting tolerances and GD&T in themodel Do Not generate to CAM software. A good machinist would stillneed a detailed drawing to communicate such tolerances andGD&T. There is no lazy way about it. Now in my current positionbeing from my machining background I detail drawings for thatinterpretation, and yes we use ASME, ASTM, MIL-SPEC standards etc.Standards are in place so when one orders an apple one gets anapple. As we used to say while machining 'Part to print - Print topart'. Meaning to put it on the drawing. You WILL get what you askfor. What's your cost...right.
SF
SF 2009-01-16
I have had some problems trying to convince people that themodel is often more important than the drawing if people use it tomachine to and inspect to of course. There are a few parts I haveseen on several occasions where the model had no draft but then anote was added that said draft this surface one degree. Can youguess what the actual part looked like? That's right NO draft.Certainly drawings can be useful for specifying tolerances, datumsand areas of concern but models are much better to communicateshape. It is this fact that is the root cause that autoCAD (2d)lost and solidworks and other 3d companies won.
SF
SF 2009-01-16

 

Certainly drawings can be useful for specifying tolerances,datums and areas of concern but models are much better tocommunicate shape.

Let us not pat you on the back so soon Paul. Most of us arein the business of precision. The MODEL obviously didn'tcommunicate a draft. As you stated...

 

models are much better to communicate shape

. What is your cost? Doing it right the first time or findout you have a problem later? Designers CANNOT get lazy. Machinistare not mind readers.
DA
DA 2009-01-16
hi paul,

i would argue that the reason autocad lost was not because it was2d, for it was 3d wireframe and had 3d solid modeling long beforeSW was on the scene. It lost cuz it was a low rent 3d package. iknow cuz i used its 3d wireframe for machine design and 3d solidsfor designing instruments for the hubble space telecscope (no, notthe primary mirror). it was tailored to architects and by the timeacad decided to get into parametrics it was too late as SW was hada jump start from the ProE folks who alread knew parametrics.

w/o a print, you will end up with something other than what youwant. perhaps there will be another another medium to take theplace of drawings, but in the forseeable future i cannot relysolely on the cad model for the end product. on the bright side,there is a great benefit to creating drawings and that is i findall my errors when i detail a print like wrong tapped hole sizeetc. so i find it to be a valuable check process as well.

just my 2 cents...

dave
AW
AW 2009-01-16
Michael,

I will dis-agree with you to a point. We do reduced dimensiondrawings. Nominals are drawn from the math data. We do provide adwg with gdt, tolerances, critical dimension/notes, etc.

We do this everyday with a lot of success with our manufacturingpeople, both in house and out-sourced. There is very little that weare not capable of manufacturing in house for the products we selland contract manufacture/design.

Totally math model based with all of the PMI (Product ManufactuingInformation) info embedded in the model is a bit tougher. The mainreason we do not do that is we would have to invest in morecomputers to have on the shop floor.

We are not a company that scrimps on computer resources either. Wehave better IT resources then most very big companies. There is apoint though that we do not have the added gain by going 100% ModelDefinition Based. So we have come up with best practices for ourcompany that gives our manufacturing personal what they need andkeeps engineering from spending non-value added time 100%dimensioning drawings.

Cheers,
SF
SF 2009-01-16
I have been trying to convince people at my company thatAnna's way is the best way. and I can't describe it better than touse her definitions. Minimum content or Reduced dimension drawingscan cut design time up to 60% and provide manufacturing with theleast amount of re-input errors with highly accurate information.Drawings are needed. BUT don't assume you don't have the 3d modelat all. This would be ignoring an important aspect of ourincreasingly "digital" world.

Can I come and work there?)
SF
SF 2009-01-16
Dear Anna and Paul,

The need for keeping a digital base system is great. I'll give youboth that much.

Like you (Anna) our company has a great IT backbone, with computerstations at all CNC machines. We do give the machinist the model sothey may produce their tool paths. I ask you this... how often doesyour shop floor (machinist) come to you to explain a feature? Talkabout time management. You now have two highly paid employeesdiscussing what should have been. Who knows maybe a third or fourthparty has to join in. While this is going on the machine is sittingidol and not making money. The answer is given and back to themachine. Now has to reprogram the machine to make the changes.What's your cost?

My point to both of you is it just takes a little time to produce adrawing to communicate your needs. Simple drawings I would agree tohave 'General Requirements 9900000' take care of it. It does savecost, but opens up wide range of guess work for tight tolerancework.

A man walks into a Mexican restaurant orders beans, ground beef,lettuce, tomatoes, and cheese with a tortilla. Does he want a Taco,Burrito, Tostada, Enchilada etc.? Good thing the restaurant tookthe time to draw up a menu to avoid such complications.

Sometimes the simplest things need a drawing.
DL
DL 2009-01-16

Anna Wood wrote:

 

. . . We do reduced dimension drawings. Nominals are drawn fromthe math data. We do provide a dwg with gdt, tolerances, criticaldimension/notes, etc. . . .

Anna

That describes our approach. Our drawings also provide revisioncontrol for the model file, by specifying the model file in thedrawing note.

Our system lacks the use of models for inspection. Any dimension wewant measured, ever, has to be on the drawing. Our metrology peopleare not going to open models and measure the solids. I would liketo see the time where we scan parts and compare the scan to anominal model. I've seen the instruments amd programs that do that,but they do not yet have the precision we need. We'd need around atenth of mil, and I have not seen close to that yet.

I do wish, also, that our company had the same attitude as yoursabout our computer gear.
SF
SF 2009-01-16
Michael,

I have seen, much more often, where our machinists come up when thedrawings are fully detailed because loose tolerances and midtolerances cloud part function and design intent. I like puttingthe extremely functional and tight tolerances on the drawing andthen it is very clear as to what is tight and what is "fit towind." Especially on legacy hand drawn stuff. (We have been openfor 40+ years)

It is obvious that some here have not tried reduced dimension styledrawings. They are much clearer with intent and as for your Mexicanfriend. Imagine that they listed every chemical compund that was inthat burrito in alphabetical order instead of puttting "#5 burrito"you listed Garlic, NaCl, nitric acid, pinto beans, and the like. Wehave the menus with combination style recipes. These are theimportant parts. This is tapped locate to model. It is clear andconcise.

Dwight, Our system is all on one drive and drawing numbers matchmodel numbers and revs are on both latest rev rules. I also seeyour using Adept. I couldn't quite convince management to make theplunge on that. It is way better than PDM eh?

SF
SF 2009-01-16
Let's face it. Drawings involve a lot of interpretation andthey also assume your audience knows how to read what you put downper the ANSI/ASME standard. I have seen mill/lathe operatorsmisinterpret the drawing and mirror machine a parts bolt circle orother features. So drawings are not perfect they involve a measureof intelligence in themselves and whenever you have to rely onintelligence you are taking an extra chance that that intelligencemakes the correct decision. This is what machine to model does foryou. It makes sure that the machine operator is working off of anominal, machine interpreted geometry set.
SF
SF 2009-01-16

 

assume your audience knows how to read what you put down per theANSI/ASME standard. I have seen mill/lathe operators misinterpretthe drawing and mirror machine a parts bolt circle or otherfeatures.

You need to get better machinist. Send your parts to us.

 

So drawings are not perfect they involve a measure ofintelligence in themselves and whenever you have to rely onintelligence you are taking an extra chance that that intelligencemakes the correct decision.

Let us do your detail work for you too.

 

(We have been open for 40+ years)

How did 40 years go by without you?
SF
SF 2009-01-16
Nice Michael, good arguments. Very logical. I realize thatpeople make mistakes. Maybe I made one just now in thinking I wasposting to someone that could respond logically. Thanks for votingme down too.
SF
SF 2009-01-16
I see your awake now Paul.
Those were you quotes not mine.

Bottom line is if one can't take the time to do the job right thefirst time, get outa dodge. This includes proper modeling as well.I have seen models come to us from other sources with no knowledgeof machining practices. A machinist may very well say they can'tproduce a feature. Designers think it looks good on paper or inthis case your computer. Guess what, back to the drawing board. Somodels are not perfect ether.
DL
DL 2009-01-16

 

Originally posted by:Paul LemkeOur system is all on one drive and drawingnumbers match model numbers and revs are on both latest revrules.

I don't think that would get by an FDA audit. Our designhistory and approvals have to be very clear and documented.

 

Originally posted by:Paul LemkeI also see your using Adept. I couldn't quiteconvince management to make the plunge on that. It is way betterthan PDM eh?

I find some thing about Adept inconvenient. It is moredifficult to open a component model if the higher assembly isalready open. You cannot save items to the library without closingthem in SW and checking them in. Should you want to check in a copyand continue working, you still must close, check in, check out,reopen. It is easy to do the wrong thing and lose a relationship.Some things are handy, but overall I wish we were using PDMEnterprise.
SF
SF 2009-01-16
I am NOT saying that models are perfect. I did say that theyare better for communicating a nominal shape. I also say that"tolerances must be communicated on a drawing" and yes a knowledgeof how it is manufactured is always a must.

What gives you right to insult my companies machinists because Ihave seen them make a few mistakes? Our machinists are very good.They know how to "machine to model." Is your companies machinistsperfect? Do you expect them to interpret everything perfect? Areyou perfect at laying things out? I think my approach IS right. Itfocuses the machinist on the tightest tolerances and featuresversus relying on him to make the judgment or expecting him to findthe few dimensions on a drawing with hundreds of dimensions. Maybeyou just do simple stuff and don't have 20 page D size drawingsthat could be done on 4 pages with machining to model?

Of course I am trying to appeal logically again.
SF
SF 2009-01-16
Dwight,

Really? I could have swore the library card approach was muchbetter than the presentation SW VAR gave on PDM. How long have youbeen on Adept? and What is your experience with PDM enterprise?

We do not do a whole lot of medical equipment but we do have ISO9000/9001 cert and it does pas that.
SF
SF 2009-01-16
Paul calm down,

Discussion started with your question on the pros and cons. Do youdo full detail drawings or do you manufacture to models? This is aclear communication gap.

As far as your company... I apologize if I came on strong.

As far as my company... everything has to be perfect. We work on amultimillion dollar program where one mistakes could cost hundredsof thousands. All I's dotted T's crossed. So the devil is in thedetails.
SF
SF 2009-01-16
Designers think it looks good on paper or in this case yourcomputer.<--- This was you not me and I am calm.

We also work on multi million dollar equipment. We do turbopumpsfer Rockets, and turbo driven powerplants that run in the 70,000 to100,000 rpm. The devil is in the details and I am telling you thatthe machinist can see those details better if he doesn't have to gosearching through long complex unimportant dimensions for thetightest tolerances. He will find the tightest tolerances easy ifthose are the only ones on the drawings.
SF
SF 2009-01-16
ANSI/ISO?ASQ Q9001-2000 >> 7.3 Design and Development.
ASME Y14.100-2004
ASME Y14.5M-1994
ASME Y14.3-2003
ASME T14.4M-1989

Should I go on...

Standards are in place if one chooses to use them.
SF
SF 2009-01-16
Yeah, Just one more....
ASME Y14.41-2003, Digital Product Definition Data Practices
SF
SF 2009-01-16

Paul,

One more acronym:

MID -MinimumInformationDrawing orManufacturingInformationDrawing

SF
SF 2009-01-16
Good one Ed,

Our personal fav is. . .
MIL-FD41
DL
DL 2009-01-16
Paul

My experience with PDM Enterprise is like a hour of playing with itat an SW World2008 booth. So maybe there's a gotcha somewhere thata person with real PDM Enterprise experience would know, but it didall the basic stuff much more easily, in my opinion. I left feelingenvious.

- We've been with Adept about 4 years.
SF
SF 2009-01-16
Thanks Dwight. We evaluated it a couple of years ago so maybepdm works has changed a lot since then. I know SW has changedsignificantly in that time.
AW
AW 2009-01-16

Michael Maestas wrote:

 

I ask you this... how often does your shop floor (machinist)come to you to explain a feature?

Sometimes the simplest things need a drawing.

Michael,

Not very often. I rarely hear from the guys in the shop withquestions.

Cheers,

SF
SF 2009-01-16
Anna,

Solute to you and yours.
KQ
KQ 2009-01-16
Here's my take on this. I work as a product designer for anumber of different companies, and I get involved in setting updesign systems. I think what you have in place very much depends onhow much work you sub contract out and the nature of the subcontract. Let me give you an example.

I have one customer who does most things in house - family run,niche market, wide range of manufacturing capability (pressing,forming, sheet metal, injection moulding, assembly etc). When Ifirst started working for them the owner would sketch out a designand take it to the toolroom and they would make it - more craftthan science. Tools were made like this as well. The products arenot that complex but even so that was the process.

I started to work with them and used 3D modelling (back in 1994)for design and visualisation, and CAD generated 2D productiondrawings. One product we did was a hook device consisting of ablanked stainless steel hook shape, insert moulded with a couple ofrivetted levers. This took 10 A3 sheets to fully define using 2DCAD. The dxfs were used for the wire cutting the tools.

I put SolidWorks into their company in 1997, along with CNC Mill,Lathe and laser cutter, driven by AlphaCAM. Within 12 months thedesigners were sending packages out to the toolroom that consistedof fully dimensioned and toleranced drawings, dxf blanks for lasercutting, BOMs linked to models, models with the SolidWorks Viewer.We put in a quality process whereby a designer and toolmaker wouldsit down and check the issued drawings. It was a disaster. Theproblem was that the toolroom guys could not interpret the designas they saw fit so there was an endless stream of machinists askingfor clarification of details and dimensions and tolerances. Theother (bigger) problem was that the designers did not have themanufacturing experience to place critical tolerances or even tofully comprehend the process they were specifying. There was a highstaff turnover!

Going forward to now, the design team consists of a manager (who Iemployed in 1997), supported by 3 designers, all of whom are timeserved in the toolroom. The toolroom consists of a manager (thesame one) backed up by 4 cell leaders, who in turn controlproduction. A lot of the tools are now out sourced.

My point is that it has taken this company 10 years to move fromthe "fag packet" to the 3D model. Detailed drawings are stillproduced but now they are done primarily for the purposes of outsourcing tools or sub contract components. The company knows whatthey want and what processes they need to achieve it so they arevery clever about out sourcing now.

The other example is a completely different situation. A new startmedical company, where we have gone through clinical trials, andmoved into volume manufacture and licensing. Due to the nature ofdevelopment, there were hardly any drawings done at all up untilproduction tooling tenders stage. The reason being that we workedclosely with a toolmaker for prototype tooling and this was choppedand changed with each trial. The process was model, GA assemblydrawing for approval, model issue with critical dimensions marked(and I'm talking maybe 2 or 3 a sheet max). The toolmakers knewwhat needed to be done and they knew what had to fit where so weleft the tolerancing up to them - it was the end result thatmattered.

Move onto production tooling (£400k investment) and we issued3D model data, GA assembly drawings and component drawings with the2 or 3 critical dimensions, and assembly procedures as pdfs, and,most importantly, the prototype mouldings that worked.My experienceis when dealing with China for tooling and sub contract manufactureis don't get too precious with tolerancing and how you think itshould be made. 9 times out of 10 it won't be done like thatanyway, and to be honest as a designer, provided the function isright, I'll always bow to the experience of a manufacturer whomakes parts like these day in day out.
SF
SF 2009-01-17
Kevin,
Pure poetry.
SF
SF 2009-01-17
Clearly there has been some improvements to both machiningprograms(like Mastercam) and machines like 5 axis mills as well asmachinists becoming a little more familiar with 3d machining. Ihave worked out of house with other manufacturers, besides ourinternal shop, and they have had little problems with my reducedcontent drawings.

So just to get a long story shorter. Sometimes it works andsometimes it doesn't? I have to believe there were some reasons towhy it didn't work as well as why your last two worked.

What the heck is a "fag packet"? A packet of drawings? Unfamiliarwith lingo.

Certainly designers who do not have manufacturing experience ortooling experience should try to get some.

Michael your comments are really unneeded. They are of little or novalue. What is pure poetry? Are you delighting that Kevin hadproblems and couldn't solve them in the 1990s?
KQ
KQ 2009-01-17

Paul Lemke wrote:

 

Clearly there has been some improvements to both machiningprograms(like Mastercam) and machines like 5 axis mills as well asmachinists becoming a little more familiar with 3d machining. Ihave worked out of house with other manufacturers, besides ourinternal shop, and they have had little problems with my reducedcontent drawings.

Absolutely. These days CAM systems are a lot more robust thanthey used to be. The reality is that a decent CAM system (like sayPowermill from Delcam) operated by someone who knows the processand the software can get results that we could only dream about 10years ago. Everything gets better - the CAD software, the CAMsoftware, the CNC controllers, the cutting tools, the CNC machinesthemselves, and most importantly the toolmakers now have a breadthof experience in cutting from 3D models as opposed to manual onmachine programming.



 

So just to get a long story shorter. Sometimes it works andsometimes it doesn't? I have to believe there were some reasons towhy it didn't work as well as why your last two worked.

The main reasons why the process didn't work was acombination of human inexperience, immature technologies and toomuch control. Back in 1997 this was still fairly new for small UKcompanies. Sure it was established in the big companies but notsmaller ones. Nobody in the process really knew what we were doing- we were learning on the job and we tried to put systems in placethat we thought would work. In actual fact the processes settleddown after a few years, but it was really only when the staffingissues were resolved that it all worked out. That and theacceptance by the company owners that they couldn't compete bydoing everything in house.


 


What the heck is a "fag packet"? A packet of drawings? Unfamiliarwith lingo.

That would be cigarette packet PaulAs in sketch on the back of. It is a figure of speech here - but Ihave actually seen a tool drawing on a napkin


 


Certainly designers who do not have manufacturing experience ortooling experience should try to get some.

This is the missing link in design education as far as I'mconcerned. It needs to be more integrated into the curriculum.Having said that I learned more in 6 months at that small companythan probably in the 10 years before then. Sometimes designing canbe a join the dots affair when it comes to manufacturability andyou actually need to do things to fully understand the implicationsof what you are proposing. I think that the wide availability ofrapid prototyping by additive machines has actually made this worsein recent years. I know plenty designers who have no experience ofmoulding yet happily fire off crazy models for prototyping thenpresent these to customers for approval. Use with care!


 


Michael your comments are really unneeded. They are of little or novalue. What is pure poetry? Are you delighting that Kevin hadproblems and couldn't solve them in the 1990s?

Paul maybe I'm seeing the light side but I'm taking it thatMichael sees some resonance with my recollections about the issuesfaced in getting from the design to the finished product?

SF
SF 2009-01-17
Thanks for the input Kevin, it lines right up with mine. Iagree with your appraisal of what is possible working directly witha manufacturer or more directly with the guy on the floor who isproducing the part. Knowing about woodruff cuters, shaped tools,and what is possible with a five axis mill really helps with DFM.However, we recently hired a couple of younger but really talentedguys who have sort of created a renaissance in our shop. Thingsthat were previously considered "impossible" are now easy. Both ofthem have different angles of coming at specific problems but theyhave really created a competition among the "mill" guys to"manufacture whatever is on the model. I have also seen that evenif I ask a set of questions twice to the same machinist he will notanswer the same given a short amount of time between.

Some of it could be that they are constantly learning new tricksand they have certain tooling around but I think it is morefundamental than that. I think you really have to see how it couldbe manufactured which is not an easy job and can be quite tough. Itmeans that you have to keep up on even the capabilities in the shopof specific people and processes. But if you do this and thenout-source you are stuck with tailoring the design. Do you panderto the most incapable person or to the median or highest skilllevel? But there are a lot of processes out there as well. The ASMEcodes tend to focus on function of the part and not manufacturemethodology. So I think this becomes a problem while using it. Ithink it does depend on the job and the mode of manufacture.

This is the primary reason I started this thread and I think Ireceived my answer loud and clear. Thanks for the input people.

P.S. His resonance does nothing for him, he knows if he does ordoes not resonate, does it do anything for others to just know hedoes? I am guessing you wouldn't be asking a question if you knewwhat he meant directly? Which is why I stated that it doesn't doanything for anyone if he does not state what he thought was poetryit is useless.