From Fusion 360 to... NX? hyperMILL? What else?

[boosted]

i'm aware, but even then, they come in different radii, and depending on the angle/curvature of a surface, you'd need to get different tools for different surfaces. at least to be somewhat efficient, otherwise it would work maybe a tiny bit better than good ol ball mill.

It's like surfacing with a huge ass ball though. You can take way bigger stepovers. To your point though, they only work great for very gentle curves. I've probably only had half a dozen parts where they worked well in a three axis situation.

 

[Hardplates]

There is a stlyle that has a huge radius only on the tip. hyperMILL calls it a lens cutter. Works great for 3 axis surfacing.

Solidcam has the same. Barrel Shape, Oval Form, Lens Form and Barrel Taper. Never really looked at them in the tool library till I just saw you post lens cutter

 

[empower]

It's like surfacing with a huge ass ball though. You can take way bigger stepovers. To your point though, they only work great for very gentle curves. I've probably only had half a dozen parts where they worked well in a three axis situation.

right, that was my point, very limited in only specific scenarios

 

[mutiny]

right, that was my point, very limited in only specific scenarios

We're in one of those scenarios! And in this case, it's a huge win since it actually works for a family of products that we expect to produce for a long time.

 

[empower]

We're in one of those scenarios! And in this case, it's a huge win since it actually works for a family of products that we expect to produce for a long time.

thats awesome!

 

[hanermo]

Very, very, very good thread and great comments.
I learned a lot.

No-one mentioned CATIA ?
First 3D CAD I learned, about 35 years ago.
And CATIA is well rated, although the typical users are aerospace and automotive, and their particular myopia may have a big impact.

Food for thought..

some want 5ax parts that look extremely good and hold decent tolerances (0.01 mm, less, volumetric) on high-end machines.
Qty 20-ish.

some want a few very very high-tolerance features, on their parts, and traceability is a big deal.

some want the flexibility to make a large family of parts in various flavours and sizes, sometimes with extra whizbangs.

some want good easy integration into the IT part, for sales/docs/liability reasons.

some want high volumes of same or very similar parts with minor feature mods

some want good production via better toolpaths

All above use cases are possible "core requirements" easily worth 40k for a single seat of sw of whichever flavour, when/if the parts are the typical very high value contracts.

I suspect most of these high-end packages could do any of the 5-6 use cases, decent-to-well, sometimes very well.
In some cases it would be far too difficult, or laborious, to get that "best possible result" bar none -- but a nearly-optimum result is probably quite easy, and quite fast to do, by someone truly expert in it.

My opinion:
No slight.
Talking as an analyst and 3d cad expert of 40 years.

My "feel" is that almost no users of any package can truly optimise any of the 6 scenarios, on any other sw package mentioned, apart from the one they most work with.
And that probably the best result anyone gets is only perhaps 70-80%, often 60%, of the best result available if the sw was truly used as well as is possible.

Competition:
Best cad + cam, advanced:
I would very much suggest a common contest from all You guys.

Make 3-4 simplistic fantasy parts with your favourite good or bad or hard part feature in it.
Everyone gets to want or add features..
and everyone makes at least that one model part on their favourite system.

All should share the model universally, full-res native and iges.
How long did it take to model ?
What are the critical features ? .. and how did You do them ?

Critical, imo:
How accurate is the model, in 3D.
--Whats the resolution, vs accuracy ?
How large is the native file, and the IGES.

E.
Simplistic things like high-pressure 300 bar (1000 bar ??) couplings with a mate like a flat boss (o ring), could be an example.
5x5x10 cm.
Thread form, root-crest-taper, thread end, manufacturability, efficiency come to play.

Can the sw make the part in a 5ax mill turn via turning, rolling, whirling, grinding ?
What are the accuracy constraints of each process, if any ?
Can the sw gage the part, via go, nogo, probe, or something else ? and how easy and fast ?


How long does it take to setup for machining in sw package "x" ?
And to what level in terms of tooling and accuracy and production.

In whatever machine+tooling one has available.
--
It´s fantasy parts.
No-one is losing any work, and EVERYONE will get vastly more efficient if 4-5 top guys share actual data.
EVERYONE may get a big lever to get their sw VAR to give them extra features, training, and swag, because obviously their sw lags on aa, bb, cc ..
EVERYONE may get a realistic appreciation of how sw "xx" works vs what they are doing now.

It´s like one of my long-time challenges :CAD.
"Model a ballscrew, including ogive arch".
E 25 x 5 mm, 200 mm OAL, == 80 mm usable length.

??
How does the thread end,
is it manufacturable,
will a mill-turn make one in one go,
how big is the native/iges model,
whats the real accuracy and resolution,
is the ogive arch how good,
how long did it take to do,
whats the native and iges filesize.

A:
About 3-5 hours first time to do,
20 MB file,
1 micron real resolution,
manufacturable with tapered out thread,
using an invented ogive arch of 1 micron modelling accuracy.

In 10-20-30 hours I could make a family of parts, using programming and a labelled model and an excel sheet to feed screw data of length, rise, thickness, with thread taper being a big part of the work.
I did something similar, first time took 4-5 days, learning the programming language.

 

[mutiny]

Week 1 report on my NX trial:

- You really need some kind of training to use this program effectively. I've used i GET IT ($89/mo I believe, Online Training for Manufacturing, Engineering and Design - i GET IT) coupled with YouTube channels and it's done a lot to illuminate things. NX has been updating rapidly lately so the courses on i GET IT are for an old version with slightly different UI and dialogs/menus, but it's easy enough to figure out. This YouTube channel especially (that someone on this forum linked) has been excellent at showcasing newer features on the CAD side: Bizlearn - Siemens NX Blended Training - YouTube.

- I've probably allocated 30ish hours over the last week to NX the program and/or reading through the i GET IT courses. Much of that has been without being at the computer, just reading through courses on my phone before bed. It would be better to have the program in front of me, but my trial license is node-locked to the work computer that I leave in the office. Did my first "CAD project" a couple days ago and was very happy with the result (a model that can generate a variable family of parts based on boolean parameters and external references).

- Overall, NX CAD seems to be extremely nice, easy to grasp, and very powerful. I was worried, coming from Fusion 360, that I would be tripped up by the top-down model that Fusion prefers compared to the bottom-up model that NX seemed to prefer, but it hasn't been an issue.

- NX moved to a "continuous release" model after NX 12, and the current version is now NX 1926. As of this latest release, a lot of work seems to have been put into making things more "discoverable" on the CAM side. I believe one of the reasons that people considered NX to have a "vertical" learning curve was that shit you expected to see was just so hard to find... hidden in layers upon layers of dialogues that you had to click through. With 1926, CAM operation creation has switched to what they call the "Explorer" view, where much more has been "flattened out" and made available by clicking through tabs as opposed to clicking obscure buttons that opened entirely new dialogs. For instance, stuff you'd almost always want to look at, like stepover/depths/etc., was behind a button that took you into "cutting parameters". Now the obvious options are available on the "main" tab and the less common options are behind another tab.

Beyond playing around with individual operations (confirming barrel cutters work with a fixed axis!), I haven't tried fully programming a part yet, but I may find some time this weekend to give that a try (reprogramming a part I've already programmed in Fusion to see how the process compares).

 

[empower]

Week 1 report on my NX trial:

- You really need some kind of training to use this program effectively. I've used i GET IT ($89/mo I believe, Online Training for Manufacturing, Engineering and Design - i GET IT) coupled with YouTube channels and it's done a lot to illuminate things. NX has been updating rapidly lately so the courses on i GET IT are for an old version with slightly different UI and dialogs/menus, but it's easy enough to figure out. This YouTube channel especially (that someone on this forum linked) has been excellent at showcasing newer features on the CAD side: Bizlearn - Siemens NX Blended Training - YouTube.

- I've probably allocated 30ish hours over the last week to NX the program and/or reading through the i GET IT courses. Much of that has been without being at the computer, just reading through courses on my phone before bed. It would be better to have the program in front of me, but my trial license is node-locked to the work computer that I leave in the office. Did my first "CAD project" a couple days ago and was very happy with the result (a model that can generate a variable family of parts based on boolean parameters and external references).

- Overall, NX CAD seems to be extremely nice, easy to grasp, and very powerful. I was worried, coming from Fusion 360, that I would be tripped up by the top-down model that Fusion prefers compared to the bottom-up model that NX seemed to prefer, but it hasn't been an issue.

- NX moved to a "continuous release" model after NX 12, and the current version is now NX 1926. As of this latest release, a lot of work seems to have been put into making things more "discoverable" on the CAM side. I believe one of the reasons that people considered NX to have a "vertical" learning curve was that shit you expected to see was just so hard to find... hidden in layers upon layers of dialogues that you had to click through. With 1926, CAM operation creation has switched to what they call the "Explorer" view, where much more has been "flattened out" and made available by clicking through tabs as opposed to clicking obscure buttons that opened entirely new dialogs. For instance, stuff you'd almost always want to look at, like stepover/depths/etc., was behind a button that took you into "cutting parameters". Now the obvious options are available on the "main" tab and the less common options are behind another tab.

Beyond playing around with individual operations (confirming barrel cutters work with a fixed axis!), I haven't tried fully programming a part yet, but I may find some time this weekend to give that a try (reprogramming a part I've already programmed in Fusion to see how the process compares).

really cool! hoping to hear more about your experience as you get more familiar with it.

 

[plutoniumsalmon]

Hello.
Can I recommend that you learn about selection filters. You can do so much stuff if you select stuff properly. Watch this guy called biz learn on youtube he explains this very well. I really like feature curves.

 

[mutiny]

Hello.
Can I recommend that you learn about selection filters. You can do so much stuff if you select stuff properly. Watch this guy called biz learn on youtube he explains this very well. I really like feature curves.

Thanks for the recommendation! I actually have already been digging into this topic and watching the very videos you suggested. Selection filters are an insanely powerful part of NX CAD and long term feature stability.

By way of explanation for those who haven't used NX, selection filters allow you select things by referencing the most stable objects in your model. For instance, say you're applying a chamfer to the 4 edges of an extruded square. If you select the edges individually and then change the original sketch to extruded a hexagon instead of a square, your chamfer will break because there are two additional edges and ambiguity as to which 4 you originally selected. If you use the "face edges" selection filter instead and select the top face of the extrude, the chamfer will be applied correctly regardless of the shape, since the top face reuses the same internal object id regardless of the shape. Simplistic examples, but there are many filters that can be applied in tons of ways to create really robust models that can withstand design changes with very minimal or no breakage.

 

[Mike1974]

Week 1 report on my NX trial:

- You really need some kind of training to use this program effectively. I've used i GET IT ($89/mo I believe, Online Training for Manufacturing, Engineering and Design - i GET IT) coupled with YouTube channels and it's done a lot to illuminate things. NX has been updating rapidly lately so the courses on i GET IT are for an old version with slightly different UI and dialogs/menus, but it's easy enough to figure out. This YouTube channel especially (that someone on this forum linked) has been excellent at showcasing newer features on the CAD side: Bizlearn - Siemens NX Blended Training - YouTube.

- I've probably allocated 30ish hours over the last week to NX the program and/or reading through the i GET IT courses. Much of that has been without being at the computer, just reading through courses on my phone before bed. It would be better to have the program in front of me, but my trial license is node-locked to the work computer that I leave in the office. Did my first "CAD project" a couple days ago and was very happy with the result (a model that can generate a variable family of parts based on boolean parameters and external references).

- Overall, NX CAD seems to be extremely nice, easy to grasp, and very powerful. I was worried, coming from Fusion 360, that I would be tripped up by the top-down model that Fusion prefers compared to the bottom-up model that NX seemed to prefer, but it hasn't been an issue.

- NX moved to a "continuous release" model after NX 12, and the current version is now NX 1926. As of this latest release, a lot of work seems to have been put into making things more "discoverable" on the CAM side. I believe one of the reasons that people considered NX to have a "vertical" learning curve was that shit you expected to see was just so hard to find... hidden in layers upon layers of dialogues that you had to click through. With 1926, CAM operation creation has switched to what they call the "Explorer" view, where much more has been "flattened out" and made available by clicking through tabs as opposed to clicking obscure buttons that opened entirely new dialogs. For instance, stuff you'd almost always want to look at, like stepover/depths/etc., was behind a button that took you into "cutting parameters". Now the obvious options are available on the "main" tab and the less common options are behind another tab.

Beyond playing around with individual operations (confirming barrel cutters work with a fixed axis!), I haven't tried fully programming a part yet, but I may find some time this weekend to give that a try (reprogramming a part I've already programmed in Fusion to see how the process compares).

That's good to know! The version I used was as you said, stupid menu after menu after menu to see the obvious stuff like you stated - stepover/doc, feeds & speeds, etc.

 

[Mike1974]

Very, very, very good thread and great comments.
I learned a lot.

snip

E.
Simplistic things like high-pressure 300 bar (1000 bar ??) couplings with a mate like a flat boss (o ring), could be an example.
5x5x10 cm.
Thread form, root-crest-taper, thread end, manufacturability, efficiency come to play.

snip.

Just want to point out from a machining perspective, fully modeled threads are kind of pointless. If you need a J controlled root rad, it is done with a "special" insert, standard threads are done with a tap, threadmill, or insert/inserted tool. There is no reason (for machining I want to stress) to have a detailed thread for programming. At least I am having a hard time imagining a reason for it. And before it gets said, I'm talking threads, not a variable helix or spiral or something like that.

 

[mutiny]

(Abridged) Week 2 report on NX trial:

- Definitely got less done this week than in week 1 (my business partner reminded me that messing around with a trial version of NX didn't make us money), but spent what time I did have diving deeper into the CAD side of the application.

- NX CAD is really fuckin' good. Gkoenig has mentioned this quite a few times and it's good to see what he's talking about. If you're a "jack of all trades" that finds yourself both modeling and programming, you're going to be extremely happy with NX CAD. It's extremely capable and has none of the "vertical learning curve" that people talk about with regards to NX CAM. There are a few different workflows you can use depending on your particular arrangement (solo, small team, large org, etc.), but that mostly pertains to how parts will be assembled, and once you get past that, the actual modeling tools are rather intuitive and very powerful.

- I still think that some kind of paid guidance for modeling is well worth it, but less mandatory than the manufacturing application.

- I "attended" a virtual conference called "NX University" put on by the VAR I'm working with, Swoosh Tech, earlier this week. It was... fine. Good content (various presentations by their apps engineers on both the modeling and manufacturing side), but it felt over-marketed. It feels very old school to require registrations/etc. for virtual presentations when people are frequently sharing their wealth of knowledge on platforms like YouTube or live on places like Twitch. I'm probably not the average customer, but I feel like the "conference" would have had better success as a Twitch event. Honestly I'm even kinda interested in the upcoming event from Titans of CNC, and I normally can't stand that guy.

- I have another few weeks on my trial license. Even though it started around August 24, Siemens was gracious enough to extend it through the end of September. Hoping to actually have some parts programmed and running by the end of that to see whether theoretical gains translate into the real world. Might have to rely on some community members for working Speedio posts in order for that to happen. (Greg, if you're reading this... <3)

 

[JJos]

We got a quote for 3 axis NX Cad/Cam Package for 22,000! and 1 year maintenance at 4,400, last year at this time it was 15,400 and 3,400, are we being ripped off or they increased the pricing?

 

[XD341]

There are so many different options with NX that I'd double check that it is the same package. Some of the Mach packages have things that a typical job shop wouldn't use. Team center...etc.

Go by the actual package number, NX cad/cam 3 axis milling should be 12450 or something like that 12451 for node locked.

 

[JJos]

There are so many different options with NX that I'd double check that it is the same package. Some of the Mach packages have things that a typical job shop wouldn't use. Team center...etc.

Go by the actual package number, NX cad/cam 3 axis milling should be 12450 or something like that 12451 for node locked.

Yes I am talking about NX12450 $22,000, How much should I expect?

 

[XD341]

Yes I am talking about NX12450 $22,000, How much should I expect?

I'm not to that stage yet, just started a trial and my research, current GSA pricing is listed online, I'm not sure how much of a discount uncle sugar gets, if any. It's a good 25% less than your price.

 

[rickyt]

One that most don't mention or really know much about is Topsolid. I'll say the Cad side is top notch, the machining side is kinda complicated, but the software is Kinematic aware and really shows what the machine is doing. If you do look at TS my #1 recommendation is to use the Tool Library within the software first and add custom tooling as you get more formulare with how to set up new tool geometry settings. The graphics are really nice and the verification has some pretty slick features within. Toolpath calculation times is pretty fast and so is the posting of the code. The setup sheets are really nice and almost everything is customizable. It's a large software that takes time to master and I don't thing you could ever outgrow it.

 

[JJos]

There are so many different options with NX that I'd double check that it is the same package. Some of the Mach packages have things that a typical job shop wouldn't use. Team center...etc.

Go by the actual package number, NX cad/cam 3 axis milling should be 12450 or something like that 12451 for node locked.

You are right, it seems like I was quoted 12450 instead of 12451, and the government gets a 25% discount buying direct so this is probably the profit of the reseller.

 

[merouane]

If that works out, the cycle time savings of being able to use barrel mills in a 3ax setting essentially pays for the software.

Are you serious wow didn't expect that