An updated list of working 5 axis CAD and CAM (Post) options for Pocket NC?
paua...@gmail.com
Sorry, this post ended up epic, but I wanted to be comprehensive, because if I cannot resolve this basic issue I will be sending my machine back (unused). I don't need another 3 axis mill.
I am completely new to Pocket NC, but should not need to state the fact that I purchased it for the 4 and 5 axis capabilities. However, upon following all obvious instructions and tutorials I have so far been unable to generate a 5 axis tool path using Fusion 360 and its currently listed - pocket nc.cps post.
(My Active Plan in 360 shows up automatically as 'Fusion 360 Ultimate, Trial' by the way, which by the look of other posts must be a recent development.)
I am working on Mac, in case that matters.
I see that others have had issues in this regard, and that there was/is a new BETA post processor (Post) for Ultimate.
So my questions are these -
1. Has the BETA in question now been added to the Ultimate Posts list by default? Or is it still needing to be uploaded manually into Fusion Drive?
I tried that, and "My Cloud Posts" did not show up (yet) in the Source dropdown in the Post dialogue. Can I take from this that the package mentioned is now active, and is therefore no longer in BETA? If this is the case then all posts and references could perhaps be updated on the Pocket NC website and within this forum.
====
UPDATE - Not included in the instructions for uploading Posts from the Cloud was that you need to go here in Fusion and turn on the 'show A360' option -
https://myhub.autodesk360.com/g/account/settings/
====
2. Has anyone compiled a current and updated (and maybe even ranked) list of working end to end software options for generating Pocket NC tool paths?
Nobody? Surely not.. or am I the first person to buy one?
I see a few people talking about various other CAD and Post solutions for multi-axis, some of them produced by the users themselves it seems, but with no clear indication of relative performance or workflow issues (or costs etc).
It all seems a little haphazard (wilfully neglected even) for a product that is aiming to win customers over.
Is the Pocket NC 5 axis only 'in theory' rather than 'in practice'?
Overall, it strikes me that any hardware (no matter its physical capabilities) is precisely and exactly as limited as the software available to run it - which in this instance appears to make Pocket NC's claim of 5 axis status somewhat bold. Whilst the machine is physically capable of simultaneous 5 axis movement, there appear to be no software solutions capable of generating such a path for this hardware. And what this means (in effect) is that it is not actually a 5 axis machine.
Given that this machine has been promoted so fully for its 4-5 axis capabilities, I guess it would be natural to presume that those functions (the whole 'point of difference' being sold in that case) would be especially well covered in the related software suite - but apparently not. Quite the opposite - the end user appears to be on their own when it comes to developing the software to run these higher functions.
It is almost as if PNC company take no responsibility beyond the delivery of 5 axis hardware alone, with no consideration of the (absolutely essential) 5 axis software to run it. Which is, of course, like selling a computer that nobody has developed software for - (such would probably be illegal etc, if the implied claim was that it actually did something of value, rather than merely functioning as a very expensive paperweight).
I have to say - I did not expect this. I would have considered this an obvious priority for the company, assuming of course that it wants to grow. I would have thought that there would be not only one working path, but many, and that they would all be laid out somewhere on the resources page of their website. After all, without at least one end to end work path, the machine CAN'T WORK.
So anyway, and reluctantly accepting that this is not actually the case, I have not yet reached the point where it will go back in the box and be shipped back to them. I will persist for the moment, and see what emerges based on what other end users have encountered.
I suppose what I have been looking for is a map of the current landscape so to speak, a possibility tree, with all the various remaining workflow paths laid out - so that will be my mission here.
Therefore, if anyone can take the time to list any known and working options below, then I will happily update this post to list each here, along with any notes and issues etc.
==
Also - at the bottom I will add a simple 5 axis demo file, one that currently fails to Post out of Fusion under the existing suggested Pocket NC setup (using either the Swarf or Multi-Axis Contour options).
I will briefly describe it here to save you the bother of opening it up -
It is basically the top half of a geometric sphere, and at present even this simple form cannot be cut using side tool (Swarf) to shape it, or end tool (Contour).
Essentially, if you imagine a single line of latitude on a Globe - we are seeking to either Swarf cut along that line with the side of a tool, or Contour cut that line with the end. Either/Or.
In 3 axis such a shape (sphere or egg shape etc) would be cut from the top using a Ball Endmill and a tight Morphed spiral. But a side cut (Swarf) or end cut (Multi-Axis Contour - tilted 90 deg) of the same path using a Flat Endmill will do this in much less time, with perfect finish, and only a tiny degree of tool wear comparatively.
This is why we purchased the PNC. These are the only two 5 axis functions, and either one will do, but neither of these work (apparently).
Therefore, this will be the simple function upon which I will happily judge (and test) any suggestions. I want to edge cut a convex body. Any convex body.
==
Importantly - If Pocket NC cannot do either of these simple operations (and I don't mean to be disparaging), then it is not actually a 5 axis machine, and should not claim such. That would be misleading to the point where I would then be seeking to send if back and get a refund. Indeed, it would not even be a 4 axis mill in that case, given the file offered below is concentric and oriented around the centre axis of the B turntable. Any 4 axis lathe could do this without a hitch.
Basically - If all Pocket NC can do is 3 axis on various planes, then I would rather have made a dynamic vice and purchased a 3 axis mill. It is misleading.
Or am I missing something obvious?
Can Pocket NC side or end cut a freeform shape from surface, or not?
Can do ANY simultaneous 5 axis operations? - please supply details.
Even one would do as a start.
I look forward to any reply whatsoever.. or to simply being told that I am doing something terribly obvious incorrectly, and to have that mistake or oversight corrected. After all, and I given what I just spent on this machine, I would dearly love to be wrong about this.
Hambo
=======
CURRENT LIST OF KNOWN (OR PROPOSED) SOFTWARE OPTIONS -
- FUSION 360 ULTIMATE + POCKET NC.cps (from standard default list)
This combo will simulate the below cuts perfectly, but in Post both will result in the following error -
https://drive.google.com/open?id=0B706W4L4-DdlUF9KdGVqWk9lRlk
Other Posts listed in Fusion (such as the one listed immediately below pocket) will generate these tool paths no problem, so there is nothing wrong with 360. But what use is that if you only have a PNC?
Verdict? - this option (the standard PNC setup) doesn't work for generating 5 axis tool paths.
What's up with that!
(Feel free to correct me by offering repeatable working examples.)
- FUSION 360 ULTIMATE + ULTIMATE POST (BETA) (from personal posts folder)
https://groups.google.com/forum/#!category-topic/pocket-nc/software/_d8qu43GOBg
Has this Post come out of BETA now, and been uploaded to Fusion as the new default? As mentioned, the instructions http://www.pocketnc.com/fusion-360-ultimate-post did not result in the new Post becoming available in the Source dropdown for Post even after one hour, and having restarted 360 twice.
See here -
https://drive.google.com/open?id=0B706W4L4-DdlWkJwdGpJMWtLcGM
And here -
https://drive.google.com/open?id=0B706W4L4-DdlLUlUTW5lSTN2dGc
Am I doing something wrong? Or are the instructions out of date?
====
UPDATE - So I added it to my personal folder using instructions found here -
https://forums.autodesk.com/t5/fusion-360-computer-aided/pocket-nc-fusion-360-ultimate-post-and-instructions-to-install/td-p/6771631?lightbox-message-images-6772404=306009i230B0E95349C5464 and then accessed it via 'personal posts' in the Post dropdown. And that didn't work either.
Then I found this -
Eventually I discovered that there were two folders of the same name -
https://drive.google.com/open?id=0B706W4L4-DdlYldOZ3czd2pjWDg and that I had added a 'Posts' folder to the wrong one.
It finally worked.
The two Posts I installed are here -
https://drive.google.com/open?id=0B706W4L4-DdlajVWTFMzclRXeTA (the same failure as before - see here https://drive.google.com/open?id=0B706W4L4-DdlemJWdUgwamZMdVE)
https://drive.google.com/open?id=0B706W4L4-DdlNDlETXQwQ1VTRDA (and this one worked - see here https://drive.google.com/open?id=0B706W4L4-DdlMlg5cmE0c095dW8)
The Post that worked was named pocketnc_m5x.cps and was found here - http://cam.autodesk.com/community/partner/pocket-nc/
Interestingly - this sole functioning output file is listed nowhere on the Pocket NC website. WTF.
I have yet to test the resulting path in practice, but at least it is generating G-code out of Multi-Axis tool paths!
FYI - if accessing the file 'from personal post' had not worked, then this was possibly of interest -
https://forums.autodesk.com/t5/fusion-360-design-validate/where-to-save-personal-post-processing-files/td-p/5606382 which led me to this -
https://drive.google.com/open?id=0B706W4L4-DdlaEkwVDViSHFpTHc
With the result being that finally the My Cloud Posts appeared in the Sources, but still no Posts were listed, even though I had uploaded to A360 Drive several hours prior. So that was another mystery solved also.
====
- 5XMONKEY
https://groups.google.com/forum/#!category-topic/pocket-nc/software/u1fBzxKp2nU
http://www.5axismaker.com/
http://www.5axismaker.com/software/
Has anyone had any luck with this in practice using PNCs hardware configuration?
Any issues, links, comments, workarounds?
- SIEMENS NX CAM
https://groups.google.com/forum/#!category-topic/pocket-nc/tips-tricks-and-tutorials/cuRmuBXUjTg
Has anyone had any luck with this in practice using PNCs hardware configuration?
Any issues, links, comments, workarounds?
-That's the list for now.. PLEASE SUGGEST MORE!
We only need ONE that works.. :-/
(I will update you on the test for the option - pocketnc_m5x.cps accessed via 'personal posts' in the dropdown - outlined in the update above, once I run it.. we may have a winner I guess. But after all this, I am not holding my breath.)
=======
I don't really care what it costs, all I want is one end to end workflow path that generates 5 axis G-code for PNC. Anything will do.
Here is the test file for any further options you know about - http://a360.co/2v9qGFt
Can ANYONE get either of these tool paths through to G-code for PNC in a form that will actually run on a PNC?
If not from the standard configuration, or using files found at pocketnc.com then - WHY NOT?
Hopefully this post offers a glimpse of what anyone might be in for if they purchase a Pocket NC in order to run simultaneous 5 axis milling. It's not pretty.
If the solution becomes apparent here before long, then it needs to be posted on the front page of the Pocket NC website in order to spare people all the grief of running this epic gauntlet.
I have better things to do with my time.
These guys need to do some urgent work on this critically important part of their business.
Tokyo Pav
Tokyo Pav
Pocket NC
paua...@gmail.com
Thanks heaps, that's a big help. I see the same suggestion below also, from the PNC team.
And I do apologise for the ludicrous length of my postings here, I am merely seeking to be thorough. I do not expect anyone to take the time to assist me, but would certainly be very deeply grateful.
To skip all the details, feel free to jump to the bottom to see my basic questions. :-)
My hope is that by taking the time here I will eventually help others who come up against the same issues, and also that the end result will be the development of a working solution.
I also support Pocket NC. As a product it is right up my alley. And this is just me trying to overcome what seems to me to be a very serious issue that could, in the fullness of time, be very bad for their business. Or, put the other way around, it is an issue that - if resolved, and demonstrated - open up huge potential markets for them. After all, for this machine to end up being useful, rather than merely interesting, could have a big impact on their bottom line.
Anyway, I will continue my quest here...
As you say, my error there was in transposing the axis orientation from the operation to the setup also, whereas they obviously need to be different.
So the solution to my mistake was to orient the Setup this way -
https://drive.google.com/open?id=0B706W4L4-DdlRHBoLWo5NUt2WDg
And the operation (Multi-Axis Contour - set to 90 deg tilt angle) as follows -
https://drive.google.com/open?id=0B706W4L4-DdlTHdYTTI5RXRhOXc
This was a basic error on my part that prevented the standard default Post from working, and fixing that means that I am now able to get G-code out of the Multi-Axis operations within Fusion. The machine will even run them, but in jerky and hesitant fashion you might say, and not at all confidently. It certainly stops for bumps and slows down for corners if I try any more complex geometry.
FYI - Here (http://a360.co/2ws9Yj0) is one of my actual working files, rather than the dumbed down test globe I offered there. If you run that you will see what I mean about the machine having trouble (I would suggest taking your vice off and having no too loaded!).
Overall, and as my post above probably indicates, I have been a little dismayed with regards to the software that accompanies the hardware in this instance. It would appear that most people purchase this item as a 'toy', in the mode of tinkering and learning about coding etc, and are quite comfortable seeing the machine itself as a end unto itself in that sense. But a CNC is a means, not an end, and the entire focus of a good package should allow a user to focus purely on the important stuff - that being the design and fabrication of the actual end product, with the software-hardware package acting as a means, and being as seamless and invisible as possible.
My main concern was that the machine (using any of the three official PNC Posts I have so far discovered) was analogous to a 'colour' printer that only comes with a 'black and white' software driver supplied by the manufacturer. Or, even worse - that the end user was somehow responsible for generating their own driver in order to 'print in colour'.
Which is to say - a 5 axis capable machine, with no packaged software capable of generating a true 5 axis tool path.
And I have to say - nothing I have seen so far contradicts this view. All tutorials and demos indicate nothing beyond what might be termed 3+2 axis, not ever 5 axis. They are merely running 3 axis paths from various different orientations, never 'live' 5 axis movement.
For instance, I have found zero published videos of a PNC working in simultaneous 5 axis from a path generated out of PNCs supported software suite (Fusion Ultimate + pocket nc.cps Post).
Not one. Can you contradict this?
I suppose some may even consider my approach here a little heavy handed, as if I am taking this all a bit seriously etc - after all “this is only a hobby machine”.
But really - IS IT?
This was not stated - and is not stated - anywhere. Not that I have seen.
This is precisely what I am seeking to establish here I suppose, if only because we are considering the purchase of more than a dozen of these machines, with a view to have them all up and running before Christmas.
In my part of the world, and with import duties etc, that is a purchase approaching $100,000.
You will hopefully agree that changes things a little, and that some considerable investigation might be warranted on my part.
Which is to say - I am not a hobbyist, and this machine needs to be able actually do what is claimed or it will be of no use to me. And at present it will not do what we require, so I am merely seeking to explore why, where the problem actually lies, and whether or not this issue can be solved.
My quest here is basically just to find out whether or not this is an issue with the code generator, or the machine itself.
My point, then, is not to be negative. I will be overjoyed if we end up with an outcome here that achieves our relatively simple requirements, and I will certainly post all details for anyone else to be able to follow. But I will not be giving up, and if it turns out that Pocket NC cannot actually do what it claims, then this will also become plainly apparent here. I will not simply be placing it on a shelf and assuming that I was at fault, not if there is actually something wrong with it.
So I have a very real and quite practical agenda here, and that is simply to find a reliable work path by which we might both generate G-code for the shaping we require, as well as achieve the execution of that code in practice, using the PNC. It is not a lot to ask, given that I just purchased a machine that on the face of it claims much more than this.
But along the way, and perhaps unavoidable in this context (if I choose to go through this exploration openly, as I am doing here), this is going to result in a kind of commentary in relation to the product and its associated software package. This is not my direct intention, I am really quite busy and not at all into doing product reviews - it is merely that the requirements I have (in terms of both generation and execution) function to set a simple end to end standard by which this package might be judged. And I see no reason not to do this exploration openly, if only so that others may benefit from anything discovered along the way. Be that positive, or negative.
Anyway - So far, and for our purposes at least, there appears to be a rather significant deficit in the 4 and 5 axis software side of things here, which does not immediately seem particularly aligned with the entire ‘point of difference’ under which the Pocket NC is being sold. It is starting to look unfit for purpose, and certainly a little like false advertising. And once again, I do not mean to be unfair here, this is simply an observation, and this appears to be an accurate assessment.
Once again, I am offering the entire PNC community the opportunity to prove me wrong here.
For instance, in order to generate tool paths capable of side milling a freeform shape, the only Multi-Axis options available within the only software package supported by default (Fusion Ultimate) are Multi-Axis Contour and Swarf. And neither of these are capable, under this software, of generating a 5 axis tool path from a body surface. Please correct me if I have missed something.
What this means in practice is that any convex form must be 'sliced up' in various ways in order to produce a series of edges that these functions can use in order to generate the appropriate tool paths.
Think perhaps of a boiled egg cut in half, and then laid down. How can you generate a cut path to shape such an object using Fusion, using the side of a spiral mill?
Fyi - our application is wood, and our tool is the side of a 1/4" spiral flute flat end mill.
Using any 'real’ software package, and on any 'real’ CNC, from 'built model' to 'milled form' is a five minute operation.
Indeed, we have copy routers that cost less than $500 to hand-build, and that do this no problem. https://drive.google.com/open?id=0B706W4L4-DdlOUR3d2RYYWN5eDQ
It's not rocket science. See here https://drive.google.com/open?id=0B706W4L4-DdlTTBUMGVJdXhlNFU for the finished item. We don't need all the detail you see in the image by the way - just the basic wooden shape, one side at a time. Pretty simple.. :-)
These machines are what we are looking to replace with PNCs, if only because they only cut one shape (as set by the copy rail), and we would like to be able to adjust the shapes on the fly, with just the push of a button so to speak. We are looking to replace every one, and this is why we are exploring PNC.
So I bought one, just to try it out.. and the hardware looks great. Wow. In practice, however - using the existing PNC suite, with its support of only one low end software package - what this looks like is a few minutes building a model, and then hours upon hours of slicing it up in various ways in order to test out cut paths in one of only two available cut modes.
So even before we begin, we need to go from this, which might take ten minutes to build -
https://drive.google.com/open?id=0B706W4L4-DdlVkV5Uk9oX3hVN0k
To this sort of thing, which can take hours by the time they are tested and refined (unfinished) -
https://drive.google.com/open?id=0B706W4L4-DdlV2s3ODYySkt0VmM
Or even this -
https://drive.google.com/open?id=0B706W4L4-DdlTUlXZHRoSmR0aWM
For us, this defeats the point entirely. And surely this is true for anyone else also. Even at its best - it doesn't even work. There is simply not enough dynamic adjustment to get a path that this machine will execute in any realistic way.
Compounding this are the associated limitations of these two options (perhaps pointing to the fact they are both being used here for things they were not intended) -
In the case of the Swarf operation, which requires two edge contours in order to generate a toolpath, the outcome is not only that the 'egg’ model will need to be sliced very finely into hundreds of edge configurations, but also that any resulting cut will by default always mill down below the surface of the defined convex shape. This is because it is not treating the object as a curved surface, but as a faceted one - think of a peeled apple. It cuts flat between the defined edges of the curved surface etc, thereby cutting into the ‘original egg’, and does this more in one part than another, meaning there is not even the possibility of allowing for this using an offset without distortion.
In the Multi-Axis Contour operation, which is designed to use the end of a tool rather than the side, and which also requires a series of linked and continuous edge contours to follow (which cannot be manually selected as the Swarf ones can - so good luck with that), the result is that you will need to tilt the tool sideways at 90° in order to get a side cut. But if you do, the resulting toolpath can tend to go a bit haywire, and the machine will hesitate constantly as it seeks to interpret such complex and jagged telemetry.
Remember - this is a simple and very pure form we are talking about, one that has no sharp corners or bumpy disruptions. In cutting this form the machine should sound like pure silk, in a continuous rise and fall of soft siren sound.
Needless to say - this is not the current outcome. Even if you bother to jump through all the hoops here to get a bit of G-code out of Fusion, the resulting cuts are executed very reluctantly, haltingly, and not exactly accurately. For instance a 90 deg tilt will not end up parallel to the final cut surface.
Therefore, I am now at the point of looking beyond the supported software package towards something a little more suitable for generating 5 axis toolpaths, so that will be my focus going forward. And I have some questions for you in that regard Paul, if you have a moment. And as I say, just skip to the bottom if you want to get to those.
But before I do that, all of this does leave me with certain deeper questions that maybe need to be asked in this instance.
It is a general question for anyone from the PNC community or company, not asked with any malice, only confusion, and growing bewilderment -
Why does PNC support a software package that does not in turn support the only distinctive strengths of its hardware package?
I mean, can the suggested (officially supported) package actually do 5 axis at all, in any meaningful or useful sense? I have seen no published evidence so far, at least none that demonstrates a repeatable workflow path.
So far, in seeking to orient myself in this new arena, I have gradually encountered a confusing mass of scattered information - an entire web of random and disconnected, sometimes contradictory, and often out of date information - and navigating all of that has been 'interesting' you might say. And, as my original post above might suggest, not particularly hopeful.
None of these offerings (http://www.pocketnc.com/s/Untitleddocument.pdf, http://www.pocketnc.com/impeller-training-series, http://www.pocketnc.com/s/WaxRingTutorial-1-kyes.pdf) are 5 axis, that is the point. And in light of all my initial failed attempts at generating Multi-Axis paths, this is what led to my original post above, and the questions (challenges) it contained.
So we should be clear - even though that challenge was well met, and with that rookie error thereby corrected, with G-code generating, there is still no working path here. Not one of any practical value.
So I ask again -
Why does PNC support a software package that does not in turn support the only distinctive strengths of its hardware package?
Why is 5 axis not 'fully’ supported, rather than 'only just’, or 'sort of'?
I mean, you see things like this - https://youtu.be/kbrrF_oNaFs
And the first (apparently valid) question asked in the comments is -
>How did you generate the code? Fusion only does 3 + 2 axis, right?
>Reply - Mastercam
But no mention of a specific Post used, or where to get it.
If there really are suitable PNC Posts available for Mastercam, then why are they not listed on the PNC website under resources?
And the same with your own reply here Paul - Siemens NX CAM + a homebuilt PNC Post.
So why officially support Fusion? What's up with that..
Are Autodesk buying them beer???
Basically - Why on Earth don't PNC produce and support at least one Post for a CAM software package that actually supports full 5 axis, given that 5 axis is the entire point of their hardware???
And this leads me unavoidably to a more disturbing question -
Is this a way of hiding the fact that the hardware will not actually execute a decent 5 axis toolpath?
It is at least veeeery suspicious...
My own hope is that it will indeed execute a decent toolpath, and that all of this will merely be discovered to be an oversight on the part of PNC in relation to appropriate software support for their product. But even if that turns out to be the case, and if the hardware passes the test, we will still be lift with an obvious 'suggestion' (which, by the time we are finished here will surely be looking more like an open demand) -
Why do PNC not just pick at least one serious top end CAD-CAM package (a REAL ONE like Open Mind HyperMILL, SolidCAM, Mastercam etc - not a toy one like Fusion) and develop at least one Post for that, if only to satisfy end users who want to actually race this little car, rather than just have it sit and look pretty in the driveway?
Essentially, I would like to see any tutorial (or even a single demonstration) of a PNC milling of any freeform shape using a simultaneous 5 axis tool path, along with the full workflow that led to such, using only the supported default package that PNC suggests.
If not - WHY NOT?
I was convinced that was what I was purchasing. Am I really the only one?
Honestly, has nobody else come up against this, and managed to solve it yet? Really??? Do people just accept that it won't do what it overtly claims to do?
I guess posting that demo will be my own task now, given that nobody else seem to have done this, and that I have been silly enough to ask the question!
Anyway, another rant.. sorry, I type very fast.
Back to the issue at hand..
It seems I made an elementary error in the CAM setup as I tested these Multi-Axis tool path options out of Fusion, and that it only took a couple of minutes for various adepts to correct my basic mistake. :-/
So we have progress. But that did not yet offer is anything of much value, as explained. I remain at an impasse.
In the end I would only like to finish up here with at least one (or even several) end to end workflow paths for any new user of PNC to follow, in order to be up and milling using simultaneous 5 axis tool paths, and in a way that is not profoundly onerous. Hence, I am gathering a series of steps that anyone might take as they first make their way into this, right from taking the machine out of the box, all the way through to milling the first 5 axis job.
Obvious Note: Defining this first path should not be my job!!! It should not be the task of any end user.
But at present, this road remains blocked. The default and supported PNC package will not currently function to achieve the workflow path our simple application requires, not without days on end of modifying and testing shapes in Fusion, and even then the outcome is not really suitable. And our shapes are VERY simple.
For instance, a tight 3 axis morphed spiral using a round tool (see demo in the the file linked above) will significantly outpace either of these so called Multi-Axis functions in both file prep, run time, and finish. By miles.
This is simply not right. That's like a Mini beating a Ferrari in a three stage drag race.
The whole point of having a 5 axis is to be able to do ALL of these things better than a 3 axis. And if it can't, then it would seem somebody put a plastic Ferrari badge on a bicycle.
Therefore, I will move on now, and reluctantly go off piste as it were, as others (including yourself) seem to have resorted to doing, and try my luck there instead. And sooner or later we will have a fully dynamic bit of 5 axis “egg cutting” code here, one that will then be used to test the real physical capabilities of this machine. It will then have no excuses, and I will be sure to post the results, either way.
My obvious hope is that the hardware team at PNC will come out roses. The machine seems physically capable, although I still have doubts about the command processing speed based on how it has handled my initial tests. But I live in hope.
It seems inevitable, however, that the software team at PNC is going to end up a bit red faced here, in either outcome. From where I stand, having to deviate from the supplied package in order to realistically and easily do what the machine claims to do is really not good enough. Especially given how easy it would be for them to develop Posts for one (or even many) software suites that actually support full 5 axis functionality and the generation of dynamic toolpaths from object bodies rather than merely following manually defined surface edges.
Why haven't they done this? Why not do it for ALL available packages? How hard could it be?
Importantly - Why should this be my job, or yours, and not theirs?
From where I currently stand these are all unavoidable questions.
So is this just software neglect on their part? (best case)
Or is this 'passive software barrier' hiding something important about their hardware? (worst case)
Given that I have no intention of backing away from getting to the bottom of this, I guess we’re going to find out..
Therefore, I have a few more questions for you Paul. Hopefully very simple ones. Thank you for bearing with me..
At the moment it would certainly appear that any example of true 5 axis PNC milling you see anywhere (including yours) has not actually been the result of a $4k investment - but much closer to $30k, when you factor in the software used to generate the necessary dynamic toolpath. And what that means is that for the $4k you might spend on PNC, you do not actually get a 5 axis mill. Not in effect.
So this is where we stand now basically -
If it can actually do this, and if we can prove it (and prove me wrong), then everyone should know about it, and everyone should have one, because that would be amazing.
But if it cannot, then likewise, everyone should know about that, because this would certainly seem to border of false advertising given the PNC 5 axis marketing spin.
So I really want this to work, if only because they seem like very good people who have put together what looks (on the surface of it) like a fantastic product.
However, this path appears to be entirely missing here, and that is a glaring omission on their part. So if they won't go out of their way offer this accessible path to their own customers, then I guess it's up to us..
So back to you Paul, all of that basically leaves me here -
I would like to at least replicate the setup you have ended up adopting in order to be able to use a Pocket NC in a fully simultaneous 5 axis mode.
First of all, however, and to make sure I am not wasting your time even in answering, I should ask this -
- In NX, are you able to generate toolpaths dynamically based on a surface body geometry only, rather than having to follow manually defined surface edges? Can you freestyle skate on that icy surface, or do you still have to hold onto the rails like a child?
You can see in the grey images above the sort of hoops that need to be jumped through in Fusion to get anything at all, and even then the results are not really worth much.
- Can I assume that NX CAM is able to bypass all of that rubbish and offer a far more intuitive series of toolpath options?
If so, then we are in business!
So could you please offer the configuration you have ended up with, that goes with your home built NX-PNC Post?
- Your OS and version? (I will assume you are running some version of windows?)
- Your NX version?
I have signed up for a 1 month trial of NX, and this will be enough time for me to test all of this out. But I assume this will automatically download the latest version.
- Do you think your listed Post will be compatible with a trial version of NX?
You probably have no way of knowing, but even a guess may assist. I will try out it anyway.. desperate times and desperate measures etc!
BTW - I also saw this in my travels, which hopefully makes sense of my $30k comment above -
NX41000 (Advantage) Node locked = $5,900 + $1,600/ year
NX41000 (Advantage) Floating -> not an option
NX11100 (Mach 1 Design) Node locked = $11,300 + $2,400/ year
NX11100 (Mach 1 Design) Floating = $14,200.00 + $3,000/year
NX12100 (Mach 2 Product Design) Node Locked = $17,800 +$3,400 / year
NX13100 (Mach 3 Product Design) Node Locked = $24,000 +$4,600 / year
NX13300 (Mach 3 Industrial Design) Node Locked = $26,300 +$4,900 / year
For the machine shop:
NX12450 (CAD/CAM 3 Axis Milling) Node Locked = $18,400 +$3,700 / year
Wow. :-/
This could be the real reason PNC hasn't bothered - any ‘real’ 5 axis software may cost many times what the machine does.
I will certainly be looking to get as much out of that month as I can!
Either that, or pay somebody who has one of these heavyweight CAM systems to generate to dynamic toolpaths, once the design part is done using Fusion.. ;-)
But even then, we need available Posts for the particular CAM package.
One final question, then, to maybe the only person on the planet who would know for sure -
- How difficult is the production of a working PNC Post for a serious bit of software like NX? Can we assume that this is a job for someone with serious skills, or it this actually quite accessible?
Being quite accessible would once again cast something of a shadow on PNC of course, so I'm unsure which one to hope for here. If it is reeeeeally difficult, then that might also explain a few things.
Anyway, and once again, thanks in advance for any feedback you may choose to offer, I will certainly test out any suggestions and post the results in due course.
In the end, then - if anyone else had hit this wall, then perhaps watch the space below.. And if nothing else emerges here, if positive proof is not forthcoming, from somewhere, or anywhere, then I guess everyone can assume that no viable and accessible path to simultaneous 5 axis was found here.
My aim will be to make sure that doesn't happen, and that I get to eat every disparaging word.
Believe me, that would be great. Bring it on..
Cheers again,
HR
paua...@gmail.com
Tokyo Pav
Pocket NC
hamis...@gmail.com
Yes the vid shows simultaneous movement - I too have been able to generate that no problem withing the Fusion environment. However, it does not show or explain the work path used. We have to presume, from the geometry followed, that it was a Multi-Axis Contour.
As explained (I believe correctly), this is only possible using either of two functions in Fusion 360, each of which end up being so clunky and limited and onerous as to be pretty much worthless in any practical terms, at least for body shaping etc.
I myself could even demonstrate the production of something useful in simultaneous 5 axis out of Fusion, and post a vid, as I have already achieved that much. But if I posted a vid of the actual workflow required to generate such then it might exceed the YouTube time limit constraints. I could literally carve the object by hand faster, many many times over. Not to mention that the resulting operation would not exactly be ideal compared to what a decent piece of software will produce. Even the best outcome so far out of Fusion is not really workable, for our purposes, compared to machines I can build myself for $500.
And I'm sure I sound extremely negative in all of this, but I have to say - this outcome surprises me. Not coming boxed with support for at least the generation of toolpaths capable of Swarf cutting a simple convex body was a bit of a shock. It may just be that we are discussing the difference between an outdated paradigm (where hardware manufacturers have zero responsibility for compatible software) and the emergence of a more holistic business model (where full software support is implied). I have a $7k printer sitting here on my desk, and guess what - it came with a software driver produced by the MANUFACTURER (one that will work on any OS you happen to throw at it).
At this level, this is normal, and totally implied.
So the idea that a machine like this is still operating in the heavy commercial environment (and only for those with deep experience) may be morally convenient, but in effect what that says is that this machine is really only for people already well established in 5 axis milling - but in which case they probably don't need one. What that says is - here is a tiny 5 axis mill for only those who already have a big one.
It strikes me that the entire point of the business model here was to break that mould and allow realistic access to 5 axis milling by end users who are precisely NOT already heavily invested in such hardware and its associated software - with the only limitation being: very small.
Go Pocket NC!
If they are not already thinking this way - then they should be. There would be lots of money to be made..
I very much doubt that many of their Kickstarter supporters had heavyweight CAD subscriptions, for instance. I believe the expectation was squarely within the new paradigm rather than the old. Without that it was like a revolution that merely reinstalled a MiniMe version of the old despot king.
It's interesting actually, maybe someone could develop a software proposal based on their hardware, and do a Kickstarter campaign!
Sorry, couldn't resist.. :-)
The point being, it would probably attract the same people again (an entirely new and MASSIVE sector by the way), and actually deliver.
You can't cry "cheap food for the masses!" ...and then wisper "sorry, tuxedo only" as they all approach with coin in hand. So if they can bridge this final gap then they (or whoever does it) will be sitting on a goldmine, instead of inciting a mob.
Anyway, and speaking of ripping a tuxedo of someone's back, there appear to be two working top-end posts (Mastercam and NX CAM) available here, including your own generous offering, and I will see what I can work out using those.
Thanks again for your response, and for explaining about the axis limits of the NX trial. I won't bother with that, in that case.
Cheers HR
hamis...@gmail.com
Thank you for taking the time to pitch in here.
First of all, I stand entirely corrected in relation to my claim that no MasterCAM Post is available within the resources section of your website.
Indeed, there appear to be several - http://www.pocketnc.com/mastercam-information
Thank you, I will pursue those avenues shortly.
I may even be completely wrong about all of this, and will certainly say so if that works out. But I will at least take a moment to address the emerging perception of claims not matching facts here, and your statements about for whom this machine is intended, and what their basic background might be presumed to be.
You display this package at 'makers' fairs. See here for a well exposed example - https://youtu.be/13872p-UynI
This is not a commercial environment.
Notice also that the machine is shown running a full 5 axis operation under the hood there.
Please correct me if this is wrong, but that path is not possible under the supported package, and is NOT what you get for the $4k listed on the sign next to the machine.
Actually, to run the machine in the mode shown, in practice, end to end, would cost a user $4k for the machine, and a minimum of $19,500 for entry level 5 axis using MasterCAM + your supported MasterCAM Post. The path via NX CAM would be closer to $40k for dynamic 5 axis.
Please correct this if there are known cheaper options.
This is my point. I am not saying that the 5 axis claim is false, I am just saying that in effect such a claim is like selling somebody a beautiful 10 ache ranch - on Mars. It may be beautiful, and it may only cost $4k, but unless they also happen to have a $30k rocket, they can't get there from where they are standing. In which case, at the very least, a brief statement of disclosure might be warranted?
And I am not even saying this is so, I am merely asking. And who better to ask?
So I am asking -
- Can your supplied package - without the addition of another $20-30k investment in software - be set up to edge finish a (any) surface body using the contour face only as reference, rather than following along defined surface edges?
- Or can it not?
Which is to say, can it actually do (for $4k) what it was doing next to the $4k sign in the video above?
Bearing in mind that this is a YES or NO question, and of course that a 'yes' will require a repeatable end to end workflow as an example.
I will happily redefine this question if it is in any way unclear, or if you are not sure what I mean. You might call this a 5 axis side cut, see here for clarity (if you can stand the music) https://youtu.be/TpfDIBT6m4k?list=PLIKB9JsUVS9YbAhNjBGO57GNeZFiWpxQr
Perhaps I will add a simple test file below.
It is - milling the surface of a shape, with the side of a tool, not constrained to following defined edges. At present Fusion can only cut a globe using lines of lattitude and longitude etc, rather that following the surface contours freely or as defined by the user. This is basically worthless.
Needless to say, I obviously remain completely open to any avenues of exploration by which the answer to this simple question might be firmly established.
Importantly, however - Anything but a yes (with accompanying evidence) will set the record straight here once and for all. That will be a 'no'.
And in that case, any sign next to a machine running such an 'expensive' file at a fair might need to be say - 'disclaimer: in practice what you SEE here will cost you at least $24k to achieve'.
Anyway, I am not actually seeking to be a pain here. Only in the presence of a 'yes' emerging here will these questions be deemed even mildly unreasonable. And in that case, you will win the table.
Indeed, I do need to clarify one thing, just in case I am in any way misunderstood - my references to 'false advertising' or anything being potentially 'misleading' are all questions, not statements, and the jury is still out here in that regard. You now have every opportunity to set the record straight. After all, it remains a distinct possibility that your package will end up being spectacularly vindicated here at the end of the day. And we should also be clear - that would suit me right down to the ground.
Because if it can actually do this, then it will be me buying the beer...
Cheers again,
HR
paua...@gmail.com
If you can cut the top of this cylinder to produce a simple convex shape like that, using the side of a flat tool (or even the end if you really must) then you will have proved me entirely wrong, and I will take it all back.
This - the dynamic shaping of bodies with the (fast) side of tools - is, I believe, what what most people would recognise as 5 axis milling. And, generally, you would be able to set the path orientation from sweeping, to circular, to spiral etc. Just as you can in 3 axis.
In my view, and that of pretty much everyone I have talked to about this, at least this much is what it needs to be able to do if you make the claim '5 axis mill'. And if that mill happens to be cheaper than the software required to drive it in this manner, then the software needs to be included, with the price raised (and indicated) accordingly. I believe this is what people expect.
Therefore, that is the question -
As it comes, can it actually do this?
As I say - it's a simple yes or no question that sets the record straight.
The file is there for anyone to try, you don't even need a PNC in order to see for yourself.
A Fusion 360 trial is free, and already has the pocket Post listed.
HR
PS - this simple challenge might even be a ticking time bomb I guess, in that the requested functionality may yet emerge in Fusion at some point, which would be a very big deal for PNC going forward. So we will wait and watch..
Tokyo Pav
Tokyo Pav
Fred Walter
hamis...@gmail.com
Thank you for your input.
Yes that's the sort of thing I'm looking for, although even then I might be seeking something more like this, if it were possible -
https://drive.google.com/file/d/0B706W4L4-DdlV1RlQTBCby1rdnc/view?usp=drivesdk
FYI - I have already been able to easily achieve something similar to your spiral example above merely by adding many more contours to the face and using the Contour function out of Fusion. I can even get side to side out of Swarf. But even then, for our purposes, the result are not really workable in practice due to the way the PNC will run such a file if the geometry is not symmetrical around the B axis (other than round). It certainly slows down (or stops) for corners etc, which is why I have tended to adopt the path shown in the above link instead.
Can we assume that something similar is possible using other types of projected geometry also? That would be a great start.
Yes I understand that the Swarf function is not really suited to what I am seeking to achieve, but you work with what you are given. Whilst this function does require two contours as reference for each tool pass, which may take a lot of work to model up, at least those contours are able to be manually selected.
The problem with the Contour option is that it will not generally recognise most contours in the same way, but it looks like you have manage to achieve that there. That's great. I appreciate you taking the time to demonstrate that even this is possible, I will give it a try and see how I go.
Would it be possible for you to share that Fusion file, so that I can see what options you selected? That would be supremely awesome.
Regarding your second example there, yes that's even better, but clearly not available here within the supplied package. But at least we have clarity regarding the terms now, thank you.
I am perhaps looking for 'flank cutting’ based on software that supports 'surface drive’.
But actually any shaping of a body would suffice, be that geometrically defined, or freeform mesh. And in this regard my more general points regarding the 5 axis capabilities of this package appear to stand.
But I suppose I should perhaps be clear about what I mean by 5 axis milling. What I mean is what most people think I mean. I mean the carving from solid media of freeform shapes, at pretty much the push of a button. As Fred explains below.
See here - https://youtu.be/TuFcx4IttnE
Or here for much simpler - https://youtu.be/MJQ60XwVS3M
Or this etc - https://youtu.be/b95zanOYb-U
And yes it could well be that what we are really talking about here is not the current state of PNC, but rather the state of CAM software, and its associated costs, which at this point in time apparently happen to make the development of very accessible hardware almost moot. That's like offering a brass mount for a 5ct diamond. What's the point? It's not a deal unless the diamond is included.
Which is to say, if what you are aiming for is accessible 5 axis milling, as most people would understand that, then you can only claim a breakthrough if the software associated comes bundled in the package. Or, if your machine happens to be very many times cheaper than the software required to run it (in the way it is being displayed in front of you) then this needs to be openly stated. Because I do not believe it is naturally implied that any typical end user for a product at this price range will be aware of these limitations, or what it might take to surmount them. In which case, your business model is basically screwed, because you are forced between raising the price of the package, or leading people up the garden path, so to speak.
So here is the problematic polarity here, in its essence -
1 - What most people mean when they say 5 axis milling is: The carving from solid of freeform shapes, at pretty much the push of a button. This is precisely why PNC run such files at their shows - it is because nothing short of this would be recognised as fulfilling the claim.
2 - What most people would expect from any piece of hardware that is claiming to offer such, is that the software necessary to achieve such will naturally be included (or be freely available).
This is, in my view, what you need to be able to deliver if you make the claim “5 axis mill”, outside of a strictly commercial environment anyway. Anything else would need to be accompanied by a very strong disclaimer.
And let's be clear - YouTube is not a commercial environment.
5 axis is what you can't do with 3 or 4 axis. So when the claim is made, then this is what people expect.
With what compromise, then? How so cheap? What allowed the breakthrough?
Small. It is accessible because it is small, this is the presumption, and the price you pay. It is not because it is limited to less than the basic claim implies.
So anyway, my point that whenever anyone uploads a vid of a PNC running full 5 axis geometry, they really need to be saying - “this is what the $25k PocketNC package can do”. Because that would at least be the truth.
- For $4k what you get is actually a (tiny) 3+2 mill that will not cut 5 axis functions in the way most people would imagine such.
- Only for around five times that (plus annual maintenance costs that may be more than the costs of a PNC) will you actually get a package that can, in practice, run simultaneous 5 axis in the way most people imagine it, and in the way you see this machine running at shows etc, or in demonstrations on the web.
It is simply not a 5 axis mill, not for that price, not in practice, and they should therefore -
a) only be displaying functions that are available within the capabilities of the package that comes for the displayed price, or
b) loudly disclosing the difference.
All else would be (dare I say is) disingenuous.
And in the end, the implied claim here that this failure has something to do with the end user’s level of experience is not supported. The supplied package simply won't do it, as you have demonstrated - not in any natural or workable sense, or as most people would understand such, even on the simplest form. Not so far anyway, unless someone else has any ideas.
Actually, and given how little noise there is around this, I can only assume a good many PNC users are (very quietly) handing around cracked copies of Mastercam. Wink wink.. ;-)
Anyway, this is all I really wanted to know, so that I can stop banging my head against a brick wall here. We have got to the bottom of it. If there is 5 axis gold at the end of this little rainbow, then it is not in the form most people might naturally expect, and would certainly take a lot of digging.
I may yet play with it a little further, but I am very busy, and basically won't be bothering trying to get anything workable out of Fusion for 5 axis. No point flogging a dead horse, as they say. But I will at least try your suggestion just to see.
Even if we didn't get the racehorse we were hoping for, a 3 axis workhorse may still suffice. As you say, at the end of the day the PNC really isn't that bad value, even when working only as a 3 axis package in practice, provided you make very small things of course - which we happen to do. I rest my case about claiming to be more.
We will see how it goes. At least this outcome simplifies things somewhat, and we can always hope for new pathways to open up here in future. After all, that would be worth millions to PNC, given that their hardware is so excellent, and how much this must be holding them back.
So one final question I guess, in that regard, just case you may know -
Will any of the 3 axis functions within Fusion allow for the tool to be tilted over a little?
To run a round ended tool on it's very tip can be somewhat wearing, slow, and very soon quite costly. I can't see an option for that in the setup. But if we could achieve at least this much we might be in business here!
Anyway, thanks again for taking the time.. I realise I have taken extreme liberties here, and that I may even no longer be particularly welcome. But I only care about the omelette, and obviously don't mind cracking a few eggs along the way. Or wasting a few words..
So it is very much appreciated.
HR
Ps - I really would love to see the working Fusion file for that spiral, if you still have it :-)
Tokyo Pav
paua...@gmail.com
Unfortunately it looks like the CAM setup for the spiral may have dropped off in the download, with the only added setup looking like this - https://drive.google.com/open?id=0B706W4L4-DdlSV9GNE5sWlVkVUk
So I'm not sure what you did there to get the operation to recognise the projected contour.
Never mind however, I have now moved on from Fusion 360. But I appreciate all your help.
We run lots of gear here, and after poking around it became clear that we had current both Rhino 5.0 (which via Grasshopper and the X5Monkey script make one path possible) and RhinoCAM 2012, and Licom AlphaCam Ultimate (which is apparently up there with the big boys, and can edit posts to suit particular hardware).
Ultimately, I would like to have seen a list, or map, as outlined in the title of this forum thread, with all current and available paths laid out by which you might travel here - from taking the PNC out of its box, through to running your first simple 5 axis shaping task.
I didn't realise when I asked that question that it might be so awkward, or uncomfortable.
And yes I took the long way home here at every stage. Brevity was not the point.
It strikes me that there may be lots of people stumbling around here, and doing so in the dark. Not dark in the sense that they cannot see, but in the sense that nobody else can see them stumbling about, which might lead them to imagine they are alone.
So the point of all the wordiness in this instance was merely to turn the lights on, so that anyone could follow along as I embarked on this brief journey, regardless of whatever might be found. Maximum exposure, so to speak. Nothing malicious. After all, I could just as easily have discovered wonders and delights.
At the very least, that way, others might not fall in the same potholes I have. Or, perhaps, to spare them the trouble of looking for what does not actually exist.
Sorry to have been so much trouble. Given the overt claims being made, what I was asking for really didn't seem that unreasonable. Little did I know that there is no such map, if only because there are no such paths here - not if what you are looking for is 3D carving from surfaces or meshes. Not one. Not in any practical or workable sense.
I will be sure to do a final update here once we get to see all of this in action, using optimised 5 axis toolpaths generated for our very simple shaping tasks, one way or another.
The hardware looks great, so on that score I have high hopes.
Cheers HR
Tokyo Pav
paua...@gmail.com
And the necessity of 5 axis, for such incredibly simple shapes, comes down to the medium - wood.
The issue is the cutter vs the grain angle, and the speed possible vs wear incurred, which demands a certain style of cutting. We use the side of a spiral cutter and maintain a cut path along the grain. Easy in 5 axis, you would presume. Actually what we use now is 2+1 rotary copy milling, so only 3 axis in fact, and it works fine.
But we have so far been unable to replicate what we do on a $500 homebuilt machine on this $4k machine ($7k landed due to us being on the other side of the world).
Was I expecting too much??? Believe me, I was desperately trying to avoid the final answer turning out to be - caveat emptor.
Anyway, here (on the right, vs our original on the left) is the best that 3 axis on the PNC will do, in a time a little over what our existing gear will match - https://drive.google.com/open?id=0B706W4L4-DdlSzVUYl9QLW44TTA
Note the ridges, which we can live with. But also note the tearing in the upper left and lower right quadrants. This we cannot repair or undo.
So in 3 axis we have no realistic path here. And yes we could I suppose operate a 4 quadrant operation, tipping the axis of the offending two over to sideways etc, and maybe we will. 3+2 etc.
But hey - we just bought a little "5 axis mill" - so why should we need to do that?
If it doesn't work in practice once we have these file paths output, due to the machine being just too slow in processing the code for simultaneous movement, then we may end up going down this path - https://youtu.be/Ovc9bfeXS1A - and seeing about tipping the tool over to a 90 deg tilt. They seem to have a philosophy that puts both hardware and software in the same basket, which is the future here, of course. After all, nobody but a palaeontologist is looking for the smallest dinosaur. (I'm sure my sense of humour isn't appreciated there any more than it is here!)
We went small in order to get accuracy. It remains to be seen just how much else we sacrificed along the way.
I will let you know how it goes..
Cheers HR
PS - imagine what these software titans stand to lose overnight with the emergence of AI..
Tokyo Pav
paua...@gmail.com
Someone is going to make a ton of they can just push through the final yards. :-)
Mea culpa for imagining that had already happened I guess.
HR