Results 41 to 60 of 99
03-15-2012, 07:06 PM #41
We can talk more intelligently about this once you tell me how you would inspect a part without a print or inspection document, solely from the solid file.
You guys want to argue something you don't fully understand. That's not a good way to make a discussion on anything, it just shows a closed-mindedness towards new technologies.
03-15-2012, 07:12 PM #42
I simply stated that it was interesting technology Joe.
FYI, I program in G-Code and have been doing so for some years now. I don't push STEP-NC nor have I suggested our own shop move to STEP-NC. I have a number of clients from my consulting days of whom I'm still close to and never pushed for THEM to move to STEP-NC.
I'm the salesman that owns none of the product he's selling... at least that's how you see me here?
No benefit here... the "experts" have spoken right?
Wonderful forum for ideas and discussion.
03-15-2012, 07:18 PM #43
Unigraphics NX for both CAD and CAM.
I use cam software to get a general procedure of what I want to do and then dial it in on the machine by hand.
03-15-2012, 07:27 PM #44
As far as teh STEP concept, you're saying that the machine reads a condensed XML file to generate the motions.
Now, obviously that requires a translation from the XML to the motion commands, which is likely in a form of an intermediary set of commands.
Those commands - for all intents and purposes - may in fact be G-code.
Granted, STEP aware controls might be ( in the future ) smarter in that they have inherent additional capability, at the end of the day all they do is
control a tool in a very specific manner demanded by the part geometry, tolerances, material conditions and some or any other parameters that might be defined.
So that still leaves me with the question that why is this concept not possible for implementation today with existing G-code?
I understand that the control might need to be smarter for this technology, but why is the blame on G-code?
Perhaps there are limitations within the interface? I don't know but this sounds like a wonderful question for the folks at STEP-NC to answer, because it's probably a bit more technical than either of us could understand without a strong understanding of object-oriented programming.
All I'm saying is that G-code itself is not the obstacle in the implementation.
And for a single file containing all manufacturing data. To a certain extent, it already is contained in a single CAM file.
Not sure how other CAM packages work, but at least in FeatureCAM, there is a single .FM file which contains geometry ( solid or 2D ), features, operations and notes.
If they want to, I'm quite certain other parameters could also be added. If those additions were to include a CMM control file, and the controls on the CMM and the machinetool
were to communicate bidirectionally, then the concept could be implemented with once again no restrictions from G-code.
03-15-2012, 08:06 PM #45
As for STEP NC it seems the biggest benefit is the ability to run the same "code" on different types and controlled machines at the expense of optimization.
Other than that, to me, it doesn't offer that much earth shattering technology. So not a game changer and thats probably why they've been working on it for so many years and it still hasn't taken hold. This reminds me of how DMIS was supposed to revolutionize CMM measurements, same code, works on every machine from every manufacturer and controller. Don't think it took the world over yet either.
The next revolution will probably be additive manufacturing getting better.
03-15-2012, 08:28 PM #46
I don't honestly know enough about STEP to expand on its capabilities. Perhaps everyone here already knows more than I do about it, considering the information being given in opposition to it.
I'm pretty much indifferent to be honest. We do just fine with G-Code, whether at the front-end or behind the scenes on conversational controls. I've worked with code long enough that I'm comfortably numb. A change will come, but probably not in my life time and certainly not by people who are rooted in any old way of doing things. That's not, historically, how these changes come about.
You're right, DMIS was the next 'revolutionary' thing and that never really happened either. However, it may not be fair -judging the validity of any technology by its market implementation. The market makers hold too much influence in these matters and change, as with everything, can come slowly, or not at all dependent on those involved. The petrol fuel engine is testament enough to that.
03-15-2012, 08:29 PM #47
relevant G-code completed with either BobCAD or FC. It is rare that you can get dimensional data faster from a blueprint than to just cut/paste into CAM ( or Bobcad in some cases )
It takes me a whole lot longer to fill in a bunch of question fields for a drill cycle than to fingercam a simple G83 line.
Does ladder logic convert to G-Code for motion control? I don't think so but then, I'm not a 'controls' guy but I don't believe the g-code relays electrical information to the drives. One way or another, there is another conversion going on behind the scenes and it's probably one neither one of us would want to look at.
Can you drive your probe from data within the 3D part file now? Could it generate say, a point cloud and check an array of points within the cycle without any additional software like, Renishaw P-plus or CAPPS-NC?
That however isn't to say that FeatureCAM for example cannot do the same in their 2020 release. Or MasterCAM or OneCNC or any other for that matter.
What I'm failing to see is the reason why such burden be put onto the machinetool control for no reason whatsoever. There is always always always a PC ( or a MAC as to not offend anyone ) in the loop, and that very
loop can easily include those features by way of the PC ( or MAC ) as a central location, completely off line and not impeding the cycletimes.
03-15-2012, 08:45 PM #48
I can create toolpaths within a Hurco Ultimax that would probably surprise you. You can even loop, change retract parameters and the like easily, which has the advantage of making something like a 3D 2-axis arc a breeze, with very near continuous motion. I can code the same thing in G-code if I want to sit there doing it for hours.
Pockets with islands? Same story; pick a boundary, designate islands, pick the entry type and tool... all tool data comes in and you've got a toolpath.
I was VERY reluctant to conversational when it first came along... then I really dedicated my time to learning it after seeing so many wiz-kids banging out 3D programs with it. I've been a firm believer in it every since. It has its place!
Again, that is just a guess, but it is also a guess that no XML drives the motors either. And if that is the case, then G-code is once again not the bottleneck and it is at least viewable.
I could if the CAM software implemented it. If StepNC does, that is a good point for them and might be a worthy addition.
03-15-2012, 09:06 PM #49
I'm sick of 50 year old men saying changes suck, this ways sucks, I hate "insert new technology." (funny he says I hate this technology until he learns how it works and now it greatest things since sliced bread) You older guys are scared of some "young punk" telling you what to do. You think your ego is shrinking or some shit.
In reality you cant even compare, meaning you can only run one machine, cant run a cnc at all or know one cad/cam program.. Hell, you meant not know cam software at all. I know eight cad/cam programs and five machine controls. When I get bored, I look into some else to challenge me.
Ran and programmed:
5axis mill turn
various bar feeders
Your wondering how can this "kid" know this much? I spend hours and hours on the computer reading learning doing... machinist forum,books,day job, working on my own stuff, leaving machine shops because I was bored with the work, talking to software companines, talking to machine builders.
Another rant... Step machining is/will better that is a fact.. Just because your scared of learning something new doesn't make it worse
Example 1: CNC vs Manual... CNC is better face it. I can program and machine one part faster and complete in far less setups then you could have dream of. I will agree manual machines do have there places in a machine shop.
Example 2: CAM vs hard writing gcode.. CAM far better for programming parts;faster, less mathematical errors, verification. g code again does have it places for quick programs bolt circles, rectangles etc.
Examples 3: Probes vs good manual method of setting your tools and fixtures offsets. Do I really even have explain this? Go set your cnc lathe with no probe or go set your 40 tools on vmc/hmc with out a probe and you will understand.
STEP machining- Once they perfect step machining it will be the way of the future. (one neat program, no more messing with posing, cmm report, tolerances etc)
John Welden liked this post
03-15-2012, 09:08 PM #50
03-15-2012, 09:20 PM #51
GarlicDude liked this post
03-15-2012, 09:49 PM #52
StepNC is only taste of whats possible.
If the iPad hadn't been invented yet and I was trying to sell the idea, some of you guys would be going on and on about how it's totally stupid. ALREADY GOT A PERFECTLY GOOD COMPUTER! TOUCH SCREENS SUCK...
I love simple and elegant systems. CAM + G code + CNC controller isn't simple or elegant. Would Apple ever design as system like that? Apple and other modern tech companies have proven that ridiculously complex systems can be made easy with good software. CNC machining could be made way easier, no doubt about that. Look how easy modern CAD has become! It's fucking incredible! Should we have just stuck with autocad since it worked just fine?
I can't understand why some of your have zero imagination or vision to see what could be done in the future. You're really going to sit then and argue for the standard way of doing things? As if that's where it should stay?
03-15-2012, 10:00 PM #53
If not, is there ever a time when the CAM seats are in use and you need something programmed to keep the spindles turning?
Can you imagine a prototype or short-run shop of any genre with one seat of CAM, lots of work and no way to program at the machines? You'd have a lot of dead iron sitting around...
Again, it has its place, but I think we're in agreement on that. If you're buying CNCs and looking at control options, why not buy one that gives you the flexibility of programming right there at the machine so you're not so reliant on CAM...
I think that's the mind-set behind the push to buy machines with good shop-floor programming capabilities. I've seen it in action and it works extremely well with the right people.
03-15-2012, 10:04 PM #54
03-15-2012, 10:44 PM #55
I am - by and large - not a prototype shop, so for me having to write programs on the control is not in any way beneficial.
The overwhelming portion of my jobs are repeat parts and are scheduled in the production line.
That means I always have the programs ready and done before the parts hit the floor.
Also, typically the qty's are high enough that I kinda have to dial out as much of the inefficiencies as possible. Sometimes I can do that is CAM, sometimes I do it by hand. When done by hand, the simple ones
are done on the control, if complex it gets modified in CAM.
What that means is that I absolutely need a flexible software to do as much as I can automatically. If that fails however, fingercam is the only option.
Now we both agree that at some point any and all CAM packages fall on their faces with customization and tweaking. Let that be SuperCAM, FeatureCAM, Hurco conversational or StepNC, at some point
they all just give you the finger. When that happens, a good knowledge of G-code is your answer.
So, in retrospect I should have said that conversational is absolutely useless in my operation, the time getting to know it is much better spent on getting good with my primary CAD and CAM package instead.
As for G-code, if you know it well enough there is nothing you cannot do with your machine, and no alternative can and will make your machine any more efficient or better.\
From what I saw in the vid. StepNC is nothing more than a smart CAM running on the machine control itself.
The exact same thing can be accomplished with bidirectional communication between the same StepNc running on your PC and the machine control, using the very same G-code you're already familiar with.
What's even more, in the video they've used a Faro arm to measure the part. The PC can already control or communicate with the Faro arm ( or a robot or anything else ), make the adjustments in the code and
re-send it to the control on the fly.
03-15-2012, 10:45 PM #56
03-15-2012, 10:59 PM #57
You're right, nothing more than a bit of a miscommunication but that happens often over the internet. It's one thing I dislike about text-talk but I've always been better with people eye to eye.
I know where you're coming from with that type of shop setup as well. Worked at Boeing for years machining parts that took days to complete, so we had lots of time to see what was coming down the pipe, tweak old programs and the like. Many of the long-term projects were revisited to incorporate new cutting technologies, make performance tweaks, etc. I'm a firm believer in doing this within the CAM system, then incorporating those tweaks in custom processes and the like. Again, some CAM is better geared towards this than others.
So yeah, totally different environment from the proto to short-run shop.
By the way, I did mention my "ideal" system was UG/NX, so I might just be the worst STEP salesman of all time here.
03-17-2012, 06:59 AM #58
Reminds me when I was very young. I carried shells from General Pershing to the French guns...didn't fit. I carried ammo to the riflemen....didn't fit.
Later, I worked for the US government as they became digital. IBM was always number one. Then came GE, Honeywell, Sperry Rand, Xerox, Burroughs and a bunch more. Any international conflict of any magnitude we would have been reduced to the telegraph. We came so far....the ammo fits the guns, the shells fit the cannons, we know where the products of conflict are stored....BUT IBM's Basic Assembly Language cannot communicate with Honeywell's Easycoder because each is propriatory to its own system. I had read that perhaps there were over a hundred conflicting systems within the federal government.
Despise Bill Gates and Apple if you wish, but at least we now are on two platforms for this digital age we now are in.
Someday there may be a 'standard' controller for machinery, standard blueprints with GD&T, and perhaps a standard way to CAM parts.
03-17-2012, 09:19 AM #59
Conflicting proprietary systems are part of the problem groups like the Smart Manufacturing Leadership Coalition (SMLC) have been trying to address, putting standards in place with the NIST. From everything I've read, XML is the cornerstone of that standard, hence the push---not from ME but the SMLC--to adopt STEP for manufacturing. The idea being that you could, on a machining level, access a model via the web, and drive toolpaths, probes, etc., from the features without ever taking 'ownership' of the model.
Interesting reading on this from the NIST/SMLC perspective.
Smart Manufacturing and Construction Systems
03-17-2012, 11:31 PM #60
ok....this only went a little OT but thats ok, as it was informative. The idea of being able to adjust programs on the fly while cutting sounds Fn awesome! Could you imagine adusting the scallop on a mold as your cutting it! Anywho....in regard to the original question, I was hoping after people gave their opinion on what software they'd buy that they would give a reason why. Lost of guys simply stating... id get this 'cad' or 'cam' cause, I just would....Id like some reasons why! Some people love certain CAMs for their programming automation, some for their better surface coding, some are dead easy with 2d. Id like to know WHY you want ug/nx or rhino cam.
And lets get back to CAM that actually exist today, Im with ya tho JW....from model to cutting edge, there has to be better ways than g code.