powers of mastercam

[jrhysdavies01]

hi

 

having been given the task to implement a CAM system within an SME i am currently conducting research into potential systme if any one could answer a few questions on the mastercam system i would be very grateful.

 

the SME currently utilises Inventor for all CAD modelling the veiw for the project is to compile all drawings as CAM drawings and download these directly to the CNC machines with little or no programing. the designs are fairly complex solid models utilising 4 axis CNC machinery to produce them. all components are solid tooling, not moulds or frames. if anyone could shed some light on to these issues i would be very grateful and highlightin of the strengths and weaknesses of masterCAM and where it lends it's major powers to.

 

Regards

 

James

 

P.S. if anyone has any suggestions for points of referance or systems they feel would be of use please could you forward any details to me via email thankyou once again.

[gcode]

quote:

---

the veiw for the project is to compile all drawings as CAM drawings and download these directly to the CNC machines with little or no programing

---

This is the Holy Grail of Cad/Cam and at this time

no one is even close. At this time the closest

technology to this goal is called Feature Recongintion. Mastercam has primitive feature recongintion of holes in solid models.

Some other vendors claim to have feature recognition and in some cases it actually works with varying degrees of success.

 

CamWorks has associative feature recognition

and runs inside SolidWorks. It does work, but

requires lots of work to setup the "knowledge bases" that drive the feature recogniton.

I've yet to meet anyone who was successful with

CamWorks in a job shop environment.

 

Catia has fairly robust feature recogniton of prismatic features. It comes with a very high price and a very steep learning curve. It also has its own design package. Importing models

from Inventor would destroy the CAM associativity

with the CAD model

 

There are some companies out of Europe, that I not familiar with that might be worth checking out

(TopSolid, Hypermill) but at the current stage,

no one can import a model and turn it into a part

without the input of a skilled CNC programmer.

[Guest]

There was an article in Modern Machine shop about 6 mos ago that stated there was software that was doing this with the information imbedded within the model.

 

Supposedly it is VERY highend stuff.

 

quote:

---

Whereas STEP NC standardizes how information about CNC machining can be added to parts represented in the digital STEP product model, N-STEP is being developed so that machining and manufacturing information contained in non-digital sources can be captured in a STEP-compliant format. This initiative underscores both the urgency and the complexity of implementing data-exchangeability standards. It also indicates why marketplace resistance continues to impede the implementation of standards that would facilitate full automation of CNC programming.

---

MMOL Article

 

 

Step _NC Homepage

 

Step NC Info

 

The end of G codes???

[Guest]

Reading through some more this might be of interest

quote:

---

STEP-NC Technology

The demonstration showed the potential of STEP-NC, using technology based on an extension to GibbsCAM and Mastercam that was developed by STEP Tools. The demonstration was done as part of the National Institute of Standards and Technology (NIST) Advanced Technology Project (ATP) Model Driven Intelligent Control of Manufacturing (MDICM) project contract awarded to STEP Tools. Demonstrations of STEP-NC technology such as this one are vital to illustrate STEP-NC's viability, raise general awareness of STEP-NC and move STEP-NC from demonstration to realization. The STEP-NC demonstration, demonstrated how a STEP-NC file (AP238) could be used in a conventional machine shop today. An AP238 format file was read into GibbsCAm using STEP Tools ST-Machine plug-in, which converted the STEP-NC information into GibbsCAM's data structures and then used GibbCAM to generate the necessary toolpath and output CNC code. ST-Machine also made use of GibbsCAM Cut Part Rendering functionality to validate the machining process. STEP Tools also announced at this IRB meeting the availability of their ST-NC product set, which will allow manufacturers to begin investigating the benefits of STEP-NC for their own businesses. The ST-Machine plug-in for GibbsCAM and Mastercam is one of the ST-NC components being offered by STEP Tools.

---

Linky

 

 

Here is a listing of operations that were completed using various CAM tools

 

Linky2

 

 

An interesting and worrisome technical development for sure.

 

Years away because of machine support I would think. I can not see shops replacing machines just because but I can see a gradual growth into this kind of equipment.

[gcode]

In very basic terms, they are trying to remove the CNC programmer from the manufacturing loop.

This is partially driven by logical economic

factors, but it is also driven by the inherent

disdain design engneers have for CNC programmers.

 

CNC programmers review the blueprints and models describing a part and then use thier "knowledge base" to create a process that will fabricate the part.

 

NC Step technolgy tries to build the CNC programmers "knowledge base" into model at the design level. In other words, the engineers are

going to tell the machines how to build the part.

 

I can see NC Step becoming viable technology for

simple parts in the big aerospace plants in the next 10-15 years. In the job shop world , no way.

 

While the big boys are trying to build the

knowldge base into the part, some software companies are trying develop software that

creates its own knowledge base.

CamWorks, Catia, ProE, and several others are hanging thier hat on 'feature recognition' and

user (CNC programmer) built "knowledge bases".

 

I think this is where the real progress will be made. The users will be buildng the knowledge base customized for thier machines and manufacturing capablities.

 

In the near future, NC Step will not work for 2 reasons.

1. The engineers really don't know how to build the part. They make think they do, but its book knowledge that usually fails on the shop floor.

 

2. When they are designing the part, they have no idea who's going to build the part and what

thier capabilities are.

 

Imagine Joe Jobshop, four tiers from the final user gets a Boeing part to build. He reads the model into his NC Step software which prompmtly

spits out a $70K tool list and requires 50hp

30K rpm spindles to run the NC file.

 

The alternative in the CamWorks/Catia process.

The user defined knowledge base knows the tooling inventory and the machines. It recognises common features and interacts with the programmer to create a procces that Joe Jobshop can actually use. I think this is where things will go in the next 10-20 years.

 

The ultimate end to all this is parts without human intervention. The software examines the model, reviews the inventory of available equipment and tooling and genterates the code.

That will happen, but not in my lifetime.

[Guest CNC Apps Guy 1]

Vega Milling Software From Mori Seiki's DTL claims very good feature recognition. I have never used it. They claim their engine is better than GSSL's. May be worth giving a look at. But as was stated earlier, at this point in time, this type of "programming" is better lent ot "Prismatic" parts rather than parts with organic shapes and features.

 

The mind is still the most powerful computer. Except when it comes to chess. Just ask Gary Kasparoff.

 

JM2C

[Bullines]

I think this thread belongs in the General Mastercam forum instead of the Developers' forum. Hold on everyone!

[bellyup]

quote:

---

In the job shop world , no way.

---

+1!!!!

 

i can imagine recognition software being very powerful, but if you look at the enormous amount of variables in even a relatively simple mold,

you would still need someone (your designer who has never run a machine?? yeah,right.) to select and/or modify the options presented to you by the "knowledge base".

spend a few days looking at the questions posed on the forum from some very intelligent and educated users. variables and unique situations arise all the time. seems to me that no matter how encompassing the "knowledge base" would be, it would still need someone to steer it in the right direction as variations in input occur.

sounds like a programmer.

[john316]

.

 

I am with you all on that one.

 

There are too many variables in machining applications that cannot be determined at the design level.

 

When I first started using Unigraphics (1999) they were venturing into imbedded machining parameters that could be posted directly from the UG file. I don't know whae ever happened to that, but in later versions i didn't see any of it.

 

I would think that the best they could do at this point would be to set up some type of preliminary toolpath that a programmer could go into and enter the parameters for his specific application.

 

The current feature recognition software is a great concept, but at this time it only works as a starting point that has to be manipulated by the CNC Programmer.

 

.

[Roger Peterson]

Automated anything only really works on holes, period. I may machine something as simple as a slot 4 or 5 different ways depending on conditions; how long is my tool, what is the material, coolant flow at point of cut(this encompasses the machines capabilities with the layout of the part - the slot is 1" deep but is at the bottom of a pocket etc...), chip evacuation, how many am I doing (do I need to minimize cut time?), fixture rigidity, is it monday or friday, etc... Well layed out operation libraries can really make short work of repetative or semi repetative situations and allow much, much more programmer controll compared to pushing the go button (which absolutely doesn't work unless you only machine square blocks with a hole in the center).

 

now, thats just my opinion.

[Storkman]

quote:

---

In other words, the engineers are

going to tell the machines how to build the part.

---

HAHAHAHAHAHAHAHAHA!!!!!!!!!!!!!!!

[gcode]

quote:

---

In other words, the engineers are

going to tell the machines how to build the part

---

Actually, they think they're going to build the

NC program into the model file

[bellyup]

 

 

"the machines became self-aware at 17:23 hours on

march the first, two-thousand and six!!

 

 

[Mike Mattera - TFM]

gcode: Feature recognition will never replace an NC programmer. We both know that. The point of feature recognition is to create a beter integration between engineering & NC programming. Getting rid of the "us" vs "them" approach that most companies have.

 

The idea is to find the things on the model, that you have to create manually now with the system you have. It also passes the information like tap & drill sizes & depths. The time you spend creating curves to create a pocket boundry is wasted motion. Especially when you get the same part back with an engineering change. Right now you have to start over.

 

Quality feature recognition, recognizes that a cavity is a pocket. It lets you grab the feature and drop a toolpath on it.

 

Mike Mattera

[Gary]

THe best bet is to hire a engineer that was/is a programmer. The designs will be less costly to build and the programming could then be done by the engineers.

[Reece L Gibson]

I think the wits of each shop is what separates them from each other. I think some simple base processes can come about, but it's the experience and thought patterns that create inovative ways to machine work pieces.

[Greg VN]

The engineer that programs handrails for me designed a system to create a drawing and code off of a autocad drawing automatically in (conjnction with PC Apt). There was a series of questions (type of staircase, rise, run, degrees of sweep, type of rail, etc....)

The drawing would be created from the questions.

There would be a program to cut out the treads, strings, risers, and the handrails for that stair.

While it is possible to accomplish that there are limits to it as well. There are so many possiblities that exist that it is not practical to try to cover everything.

There will always be a need for programers to make somebodys vision become a actual part.