Well, well, well…
Long time since my last post uh… like you, I was starting to question if CAMZone was ever going to get a new post… This one came to my mind when I was at the bed already… then the light came and I said: Let´s do it before it´s gone…
It´s true that the inspiration for today´s post came from some events in the last two days, more precisely the new cool features Siemens PLM put on NX 8.5 – Let´s talk about them among other things… today´s post is more about the industry, with a few technical points… it´s about my perceptions about where the CAM tools are heading to…
Like I said earlier, from time to time the CAM user base demands a new milestone in CAM – It´s like those guys in the 80´s that used to say that verification was a luxury, just wireframe toolpath simulation was fine, and then some guys set a new milestone and the rest is history. Then in the middle of the 90´s machine simulation was eye candy stuff, and by the beginning of the 00´s new rules were stablished, and today, a vendor without machine simulation does not get a chair around the table anymore.
So to me, every time a CAM dealer says something cool and helpful is eye candy just because he does not offer it, I think “Time will say it…” – I was born in 1980, so as an adult, I missed the best part of the grow of the IT industry, but I´m continuosly trying to catch up – While many people worship .NET, I try to buy old used ASM books to understand the roots of the things… Sadly, my children´s generation does not know what a command prompt is… too bad…
Yesterday a colleague sent me some material about the what´s new in NX 8.5, and that really shook my boat because I was able to relate to the passion for developing CAM systems again… I could smell the sweat of the developers trying to code elegant and highly customer driven enhancements… that inspired me to sit here and write about these feelings…
A lot happened in the last year… merges, acquisitions, buyouts, consolidation, and finally, awakening…
I´m just an ordinary human being, so I can´t recall and mention every move that occurred in the past 13 months (Nor do I know them), but I´ll try to use the most important ones I can recall as examples of what I perceive as an awakening happening with CAM.
Well, the Germans are going after big dreams very seriously and steadily. Yavuz is the kind of genius that you can recognize just by his pair of glasses – And he is surrounding himself with a bunch of geniuses too. The result is that, what was born as a set of alternatives to Mastercam multiaxis customers in the end of the 90´s is now a solid company dreaming about and delivering a new scene in CAM.
In April 2012 they took the almighty Dan Micsa, a true scientist in the best sense of the word, to work for them. The guy alone can put Space Shuttle in Mars, but he is working with other bright minds like David Plater and high level developers, marketers and whoever more to take CAM to the next level.
In 2012 they released at least 3 builds of their components, and each one of them brought long time asked tools to the table. I can´t really list everything here, but I´ll mention the most significant ones for me, the ones that influenced my opinion in regards their dedication and hard work to be the best CAM components supplier out there.
Dan Micsa enters in Plater´s boat and while he was still talking to HR about his corporate credit card :-) they released important features like stock transfer, for mill-turn machines.
With Micsa settled down, they released new ways to import geometry for roughing strategies (IGES, STL), adaptive roughing, enhancements in 5 axis motions and feed control.
Important changes like true helical and arc interpolation without linearization took their engine to the next level in regards performance…
Saw and chainsaw simulation shows their commitment to wood machining.
They end up a very exciting year with very exciting enhancements: New SWARF algorithms, dedicated port toolpaths. More granularity and user control were put on 3 axis toolpaths, allowing the user to user distinct feedrates in linking moves. This is so cool and often overlooked!
As Vericut, they now support non-center cutting tools, saving your bacon and that new insert cutter that allows aggressive ramp angles…
And then, the best part of it: via Twitter they announced a move to a bigger office and the procurement of machine tools to do more testing and R&D. Wünderbar!
These guys were using customers and partners like Fraunhofer Institute to test their toolpaths, and did an awesome job so far. Imagine now what they are going to do having machine tools in-house.
What all this tells me: these guys are serious, they are building a robust commercial structure while they continue to excel and hone their technical solutions. Many companies in CAx struggle to survive because they don´t know to manage it, or they don´t put enough R&D on it, or they simply put too much R&D without the proper marketing and commercial backup.
Yet about the marketing, they have built an interesting YouTube channel and are moving on day after day in letting people know what they are about. What’s the point in developing technology if you can´t make it accessible to ordinary people, like you and me? This is the kind of technical marketing that makes the difference. And ModuleWorks is using social networks modestly and convincently.
ModuleWorks has been balancing the whole thing smoothly, I think. I´m not a professional industry analyst, so I reserve myself the right to make mistakes. But like the Germans like to say, “Something in my stomach tells me I´m right”…
In 2012 they also started more important partnerships – To read all ModuleWorks press releases, visit this page: http://www.moduleworks.com/news/index.asp?NewsItemID=52
Well… HSMWorks was the big surprise in 2012 – Autodesk bought the Danes in last October, nobody saw that curve coming…
A lot of noise and suspicion about their move… looks like the company was struggling to move on alone with their vision of easy to use CAM. I know some of their developers, and I know how talented and capable they are. Out of nothing (Well, a bunch of code from a NC-program editor) they built an elegant suite of tools, with possibly the best multi-core engine out there. There´s a problem in developing software tools: the bigger and more monolithic the solution gets, the harder it is to make it portable/adaptable to new programing technologies and hardware.
HSMWorks do it well because they could start with a fresh approach to develop their CAM kernel, starting the code using possibly the best engineering workstations out there in mind, and with the backup of an important player. This, plus the inventivity and the passion for well written code, did the trick for them.
But HSMWorks did not gain their reputation only thanks to their C++ skills: They did what every smart CAM vendor should do these days: they removed the endless barriers between people who codes the software and people who use it. It was amazing, I told René Fonseca once that I thought they were taking their laptops to toilet and coding from there, because the response time to their customers was impressive, considering the industry standards.
CAM product managers and the industry are very conservative – Also because the big guys in this industry are around for +25 years… average… after witnessing so many startups and big names that closed their doors, these guys are very careful with everything.
It was necessary to have a bunch of young company directors and developers putting a new company together to have this change in customer care/feedback collection…
I would not fear to say that in their own way, they set milestones in the CAM world not only because of their technology and performance, but also the way they approached their customers while trying to appear in the CAM picture.
“A true genius admits that he/she knows nothing.”
― Albert Einstein
In their beginning, the only new differential they could have was their customer service, and that did the whole difference for them.
Time will be the judge – Let´s see what Autodesk will make with HSMWorks customers. I´ll keep my thoughts for myself on this one.
Mastercam, the master of the 2D seas… and a handsome boy in the 3D and multiaxis world.
Well, the truth is that CNC software took a long time to get the idea: It was necessary hundreds of threads bashing them, a lot of noise, some foam at the mouth blogs and customer departures, but they finally got it. And that´s the reason I´m writing this post – To highlight the awakes in recent CAM history.
Mastercam is unbeatable in 2D. The last time I touched it was 2005, but I still miss it sometimes. CNC Software is a very conservative company – They only walk the next step when they have a ton of graphs, surveys and reports telling them to do so… I think it´s a Connecticut thing :-) – Or maybe just in the Old Post road?
These guys took a long time to deliver a solution to MTM (Multi-Tasking Machining) – They promised MATSS (Multi Axis Turret Spindle Synchronization) 11 years ago and only now, in the next version, X7, Mastercam will be a decent solution for MTM. After all this time, we all hope they come up with a killer solution. They bought a Mazak Integrex with all bells and whistles some time ago, as well as a small swiss CAM company, so they won’t have to learn and code turret synchronization from scratch.
I think after all punches they got in their face in the past 4 years, they will make it right. They’re entering in this game a bit late… but if they take it seriously, the future will be good for them… Well, their MTM algorithms will be limited like everyone else in certain aspects, but we´ll talk about this in the end of this post… ;-)
CNC Software also added stock management in X6, and now they generate less air cut and track the stock changes. It´s tesselated data, but it´s what they can do for now. Tesselated data in stock updates is the first generation of stock management in CAM. It sucks, but it´s better than not having it. Users can´t generate a true B-Rep model from their stocks, nor can them detail it decently in a draft package. But for what Mastercam proposes as a CAx tool, it´s good enough.
Their graphics engine still in the mid age, and looks like they don´t care about having a tool like HOOPS or Redway to help them out. They say machinists don´t need good graphical representation. Well, I disagree, but who knows they are right?
Circa 1996-1997 Missler Software realized they could not compete with PTCs, Dassaults, CNCs Software and UGSs out there offering the same tool sets with a facelift… so they silently engaged in MTM, and today, they have a decade of experience ahead of the major players. TopSolid V6 is a powerful, wide CAx tool, but it´s cluttered with a poor UI (These days – When it started it was likely a cool stuff) and a large legacy code that is hard to keep up and evolve to new hardware and programming technologies. Their V6 open API is weak and non-existent for CAM. They can do miracles with their own internal development tools and APIs, but this is not available to customers, as far as I know.
But while TopSolid V6 is the one when it comes to a complete tool, Missler probably realized again that in order to survive more 15 years they would have to re-design the car from the ground up. They knew they had a good car, a reliable one… but it could not break some barriers with the current chassis and engine, and they started about 7 years ago to port that code to .NET, optimize it, clean the ugly parts and make it their future. Their V6 PDM is powerful, but cumbersome, hard to use, hard to customize…
So V7 was born under the premise that PDM and CAx must be easy to use, while flexible and powerful. People don´t realize that it would be nice to version their changes in CAM files, documents, CAD models and such – Except for high-end tools like Creo, CATIA, NX that offers PDM/PLM (and a few others), most CAM systems use Windows Explorer to manage their files. If you save a document, you say goodbye to the last version. People don’t miss what they don’t know.
“A lot of times, people don’t know what they want until you show it to them.”
― Steve Jobs
So TopSolid V7.7 was released this year. The experience and power of MTM in V6 was partly included in V7, but of course re-engineered to offer more usability.
Missler is a significant contributor to the CAM world, but they are too secretive, slow and often incomplete in documenting their products, but with V7 they are being offered a chance to change this as well. They have now a .NET based CAD and CAM API, where you can use VB to drive your toolpaths by writing relations and formulas inside it.
But all this is useless if the end-user can´t get a decent documentation. And I´m sorry to say my stomach does not feel confidence about changes in this direction. Simply because it´s not in Missler’s nature. They like to code, not to write.
God bless I´m wrong about this. And that only changes with facts, not with marketing newsletters.
Well, I work on a company who licenses thousands of NX seats worldwide. I can´t tell you how hard it is to some people to even look in different directions in regards CAM. The thing about CAM is that there are no tool that is good for all products and environments. This is sometimes caused by human factors, obsolescence in machinery, local support, technical features, learning curve, and a million more others. But I have to say that NX is a tool that although I do not use in my site, I deeply respect and follow closely, specially because I work in a company where NX is a common tool.
NX is the Space Shuttle – It can do nearly everything, in some areas with unquestionable success and easiness, and in some others, with limited success and at the cost of a lot of hard work and time. I like high-end tools, like NX, Creo, CATIA, TEBIS… they are hard to use but they can do things that other tools can´t… they are like assembly language: they demand more skills from the programmer, they have a steep learning curve, but they do it. So this is how I see NX. It is one of the tools I´d have in my toolbox if I had money enough to fund a machining shop for my own amusement. I love low level control over the things. :-)
In the past 3 years, at every release, Siemens PLM did what UGS did not do in 15 years. Their 5 axis algorithms were streamlined to almost the perfection compared to what they used to be (Good ones by the way). They bought Tecnomatix and are integrating their technologies in all their simulation solutions. They are being serious in fixing old problems and making the foundation more robust.
With NX 8.5 they made so many cool and important enhancements that I have to say I´m sorry for everybody above, even though I admire what they did.
The problem with high-end tools like NX is that they are cumbersome. With power comes complexity. They have to support 30 years old code, also they cannot really rethink from the ground up their products because they think their customers are not gonna pay for it. So they keep the ball rolling, but someday they will have to do it.
It´s like TDM Tooling Management Software - Their product is so old and big, that it´s hard for them to make it more usable and with a better interface. Their competitors, that are much smaller, already ported their tools to new programming languages because they did not have so much old code to maintain (Neither they did have the same level of functionality to offer).
Want to put TDM on its knees in a demo? Ask them to replace or make obsolete a certain cutting insert or component that is used in several assemblies in the database. The will tell you to duplicate that tool assembly, change the insert and give the new assembly a new ID. What about the dozens/hundreds of NC programs that are referencing that tool ID? Why do I have to manually change all NC programs to reference a new tool ID when in fact only the cutting insert has changed? Why I can’t keep my current tool ID if I want, and version the tool assembly with a new revision? They will tell you their database can store a million tool assemblies, but looking from the lean six-sigma perspective, it’s not smart to store things just because you have storage capacity.
So here you are: all you want is to update the tool assembly drawings and shop docs to use/mention the latest grade from the tool vendor, and TDM salesman tell you will have to do it manually (Or via SQL) in dozens/hundreds of tool assemblies? How is it possible that a very expensive tool management software is unable to do such a trivial task, a thing that happens in every shop, every other week.
So this is the thing with big guys: they only change when they have no options. Like PTC with Creo. Nobody was buying the “Wildfire” promise anymore. And then PTC moved on.
Back to NX, Siemens tore down some walls and reduced the distance and gap between end users, product managers and developers. The incredible amount of work put in NX CAM since NX6 is remarkable. The software is so powerful and flexible that it is hard to use. But no one can say NX can´t do it, and truth to be told, Siemens is doing a lot to make NX more friendly.
Moreover, only God knows how hard it is for a big corporation to tear down walls and get closer to its customers. So kudos to Siemens for their ability in remaining huge while acting to be nimble. Many big CAx vendors did not get that, yet.
In NX 8.5, problems with tapered walls and floors during finish and roughing operations are gone. The software is smarter now in regards stock management, and generate less air cuts. Remember I said this was going to become a trend in CAM, in this article? It´s happening… all vendors above and below acted in this direction and they are still moving on. But unfortunately, only a few like TopSolid are in the 2nd generation of stock updates – They generate updated stock models as B-Rep geometry from 2 1/2 axis toolpaths.
NX continues to rely on their IPW (In Process Workpiece) models, that are nothing but tesselated models (Kind of STL). In NX 8.5 these models are more refined and small problems they had to update milling / turning toolpaths are gone. This tells me that Siemens is making NX grounds more robust for better mill-turn support. In mill-turn CAM, toolpaths need to be harmonized to understand stock changes made by earlier operations. This is one of the most critical portions of a good mill-turn CAM software. If the turning operation is unable to recognize the material removal performed by a milling operation, and vice versa, the result is air cut or collisions.
In NX 8.5, Siemens added smart trimming/extend options to turning operations, and added support for smarter roughing turning/milling operations to reduce air-cut. It´s too bad I can´t share their what´s new material, but I can say it´s remarkable and a milestone in their history. No new fancy tool names or algorithms – Just important and wide enhancements to allow more control and efficiency.
This is what NX offers: by far more control over things than most of its competitors, and that´s why NX it is not always the best tool for certain shops because sometimes you don´t need all these controls. All you want is a straight turning or milling toolpath with the basics…
With the partnership with Celeritive to bring Volumill, they showed awareness of their limitations, and that they can still be perceived as strong and creative even licensing 3rd party technology to please their customers in an acceptable timeframe. This was nice from them, in my humble opinion.
ISV, postbuilder and multiaxis also got several improvements. Siemens did very well with NX 8.5 and I think it´s CAMZone´s duty to shed some light on this.
NX is not in my humble opinion a tool for every situation, but if I had to use it myself, I think I´d not have any problems with that.
Last but not least, NX documentation is just superb in my opinion. Being a TopSolid and Pro/E user, I can say both simply sucks compared to NX when it comes to tutorials and documentation. Pro/E is not so bad, but TopSolid docs, oh boy, it´s embarassing…
The Brits are always an incognita, but often surprising. As I said in my previous post, ModuleWorks as they major competitor was an excellent thing for the industry. They become more diversified, technical and customer driven. ModuleWorks is in my humble opinion superior in toolpath R&D, while MachineWorks is stronger in verification. MachineWorks started to offer toolpaths after witnessing ModuleWorks take their customers because they could offer more from the same source.
CAM companies are constantly looking for commodities… they no longer hesitate to license and “componentize” their products… many CAM vendors now get the full package from one single vendor: toolpath algorithms and verification. I talked about this and mentioned some examples here.
PTC unfortunately is still slow in bringing components into Creo. It´s a managerial problem, in the very top of the pyramid: they don´t like to license from others. They prefer to save money and acquire companies that can bring the technologies they need to move on. The problem is that this takes some time, and manufacturing shops can no longer afford to wait. Creo CAM is still behind, although they made a huge progress and quality improvements in their code in the past 3 years. Creo is now more stable and more tested before reaching the streets. But it leaves us the lesson that if you don’t pursue excellence and keep your product aligned with the latest technologies, even if you are a leader now, it’s going to be hard to catch up later. 15 years ago Pro/NC was within the top 5 systems.
Don’t ever commit the mistake to live from your current reputation and past achievements. The technology world is merciless with the status-quo.
Back to MachineWorks, in 2012 they closed a deal with FANUC to offer anti-collision systems to run in their controls. This is a milestone, because FANUC was one of the last big CNC makers to embrace real-time anti-collision, and they picked MachineWorks to do it. Another evidence they are the best-in-class verification engine out there.
As I mentioned here, MachineWorks is the lead in regards anti-collision and clash detection in CNC controls. The Germans from ModuleWorks are not reaching this niche yet, and here, the Brits are the kings.
In 2012 they closed important partnerships and improved their products further – To read all MachineWorks press releases, visit this page: http://machineworks.com/company-eventsnews.htm
I wish I could write more about them, but it´s hard to find material on the web about their doings…
There´s a lot of consolidation happening on CAM. Mill-Turn is the future, and the CAM companies already realized this. Most of them are ridiculously slow to embrace it, or are just catching up or replicating what their competitors already achieved. My fear is that if these guys don´t really start to solve the unanswered problems in MTM, in a few years all them will have the same toolboxes and none will solve it entirely. It happened with verification, then with machine simulation, and now is stock updates and MTM, to name a few.
In MTM, I think that dedicated toolpaths to use the best of modern tooling and machines, targeting ideal G-code (No, programmers don´t want your 4/5 axis ruled toolpath with a tilt/lead angle and million of lines) – They want to approach in Yxx, go to Y0, turn the C axis 360 degrees and go to the start Y again. Four blocks of code.
But yet, CAM vendors insist to push old and inefficient ways to do things just because they can workaround the problems with it. Stock updates needs to go to the 3rd generation: This is the generation of B-Rep stocks also in simultaneous 3/4/5 axis operations. This will allow the drafting, assembly, FEA, MBD, accurate programming of CMMs in inspection software using the current state of the workpiece, and re-usage of CAD stock models with small sizes that occupies a lot less memory compared to tesselated data. Endless possibilities. More performance and stability, as usual, is what the industry wants.
PTC already supports the 3rd generation of stock updates in Pro/NC since I was a boy. But it is unstable and prone to crash. Anyway, you can use Pro/E CAD to represent virtually any material removal in a NC step. And that is really cool.
Siemens PLM advertises that they own Sinumerik and that they can make more with it using NX. But the truth is that they are too shy: they have the most powerful off-the-shelf CNC controller out there and insists in pushing simple cycles like CYCL95 in new releases. C´mon guys, you have the best MTM control in the world and NX can´t program a mill-turn operation using FLIN and FGROUP? Oh please, don´t come to me with UDE´s and heavy post mods… this is not functionality… it´s workaround…
Ahh Ok, the point in having UDEs and macros is to allow the user to build functionality himself: Yes. And in many cases they do just well. But when such resources don’t give you the output you need (And in MTM the desired output is clean and smart G-code), then it’s just a poor workaround.
My opinion about MTM software is that those who knock at their customer door and ask them what´s the best possible way to do it and translate this as native tools in their UI and products, using special machine cycles and clean G-code, will then gain the crowd.
Even an idiot like me would know how to spec a set of algorithms for optimal MTM. Impossible to believe a bunch of geniuses in the CAM industry could not if they wanted. Stop looking only at the ratio (How many people are asking for this / How big your user base is) and make it before your customer needs to move on because you are obsolete. He won’t wait for you. Specially because you can’t solve his problem within 6 months. You’ve got to be ahead of his needs.
I think after this post, it´s important now to come to the surface and share some positive views too… and this is what this post is about: positive signals in the CAM industry. If I was a CAM company director, I´d be tearing down my walls, getting my R&D closer to my customers (Like HSMWorks did), developing unique solutions even for small number of users (Usually the ones in the cutting edge of the technology), because the future is in the next second, and that feature you put for a couple of customers is what tomorrow´s customer will need to stay competitive and stick with you.
Companies don’t like switching CAx systems. They only do when they run out of options and need to move on.
I think the future challenges in CAM will be, how to create technologies that can be customized to a large variety of situations while allowing the customers to get the best G-code. It means, how to drive dedicated machining cycles in a wide range of CNC controls and machines? It is about mapping the common factors and requirements – and start the coding – CAM systems would then generate generic metadata in their neutral files (NCI, APT, NCL, etc) that could be used in a wide range of cycles and controls, and the post-processors would handle this and create short and efficient G-code.
Those days of coding algorithms to please thousands are gone. The market needs more granularity and specialized tools to drive the machine tools we have now.
MTM is not always complex – In fact, most of times it’s the inverse
The good thing is that MTM algorithms are not necessarily more complex than a multiaxis or pocketing algorithm: What MTM requires now are basic things like how to approach a part in Y, turn the C axis 360 degrees and get the tool out of the part – Using 4 blocks of G-code. What’s the rocket science behind this, compared to a SWARF algorithm? Looks simple and doable, but most CAM systems today don’t do this in an easy manner, without user customization. And even with customization, many times you can’t backplot or verify it, or you have to fool your post-processor one way or another to get it right. So this is how bad the habit to create workarounds for everything got us frozen in CAM R&D for MTM.
People tend to argument that MTM is complex: it is, but not necessarily the algorithms. I’m a Steve Jobs fan and I agree with him that the complexity lies on how to make it easy to use. The real pain lies in making it simple to normal people.
This is the mindset we have to change: quit the belief that MTM programming is complex – The machines doing MTM are – The G-code driving them, not always. Most of times, just simple instructions spreaded across many channels working simultaneously.
This mindset is similar to that one that believes High-Speed-Machining is about high RPMs – It’s not – It’s about high cutting speeds. And these two things are not necessarily linked.
We need to have a new set of simple algorithms to do MTM with focus on efficient and short G-code: These algorithms are far more simpler than a 5 axis toolpath. What we need in CAM now is the brilliant concept PTC is putting on Creo with its apps: small and simpler tools to do specialized tasks. And you bet that some of the algorithms we need to do MTM are far simpler than most Creo apps.
A good analogy to this is how the hardware industry evolved in a different pace from the software industry. Although software is a “living” thing, it’s hard to break up with your past and start fresh. In this sense, the hardware industry was braver than the other. And that’s why we have so much legacy CAx code around that is unable to use the modern hardware at its fullest. The same is happening with CAM and the modern machine tools.
It was never my intent to cover all vendors here, and I selected a few ones I’m more familiar with… but the signs are out there for everybody to see… rusty signs sometimes, but signals anyway…
See you next time, and thanks to Bono who helped me to stay awake,