Understanding Smoothing

[al.whatmough]

Here is a GREAT video from CNC router parts!

[RandyKopf]

So sadly I see I may have killed this post??? Dang it I hate my life!!! If I did kill it I'm sorry 😕

-

I was hoping others would continue to add their experience to this excellent topic "Understanding Smoothing"...  as this is the first time I was included in a post on this forum topic I would like to see more!

BTW @Steinwerks in addition to others, I meant to engage you in my last post but missed crediting you ... I greatly appreciate all the contributions you make to this forum (and I like beer), while we've not interacted to any mention-able degree your expertise is greatly appreciated. I'm a relative noob to Fusion 360 CAM. And that appreciation also goes to @HughesTooling who has answered ****loads of questions and of course @daniel_lyall who is a constant advocate here. And @LibertyMachine always answering questions with solutions. And my Pocket NC fellow enthusiast @xander.luciano finally @al.whatmough who is a major driving force for our Fusion 360 CAM interest.

-

So I'm not stroking anyone here, just acknowledging what I see day after day on the CAM forum.

-

You are all very interesting peoples... Like each of you I have my share of stories that you can't just make up right?. I greatly value each of you investing yourselves on this forum. Please don't stop.

I would like to challenge any of you who would be open to a continuation of advanced topics? Like perhaps "Understanding Chip Load" from a 2D and 3D aspect? I've used many CAM systems over the years and it's only been in recent years that you can actually obtain real tangible feed and speed information in context to the volume of engagement.

-

Or what constitutes a proper 5 AXIS Machining and Validation Artifact?

That is a part that can be cut with the sides of a tool and also the tip and in 2X to 5X modes and everything math wise is accurate. But more importantly specific cuts are made with the side of the tool that cross validates adjacent cuts with the tip. And don't forget all the COAXIAL alignment aspects. Most importantly it is considered an artifact as you then measure it and define how it ended up. And you can compare it to the perfect target. And it shows you what your reality is. These are common with CMM's but with Fusion 360 5Axis simultaneous around the corner this is something we should seriously collaborate on. To define together and make it procedural with tool paths and how they all correlate into validating your machine.

A LITTLE HISTORY FROM MY PERSPECTIVE: I think of CAM knowledge as a journey where we move along through time... we encounter new things that are improvements. We are always striving for the optimal CAM tool path strategy. And as we discover new things the current meta changes. That is the preferred choices that people naturally use. So consider REST milling an example of a discovery. (I know the inventor) While it was a huge place in development to help us get to where we are, it's not the destination just a stop on the way on greater knowledge. We recently have added high speed machining  AND support for variable helix cutters and advanced machine tools.

-

TODAYS REALITY: Fast forward to today, we have machine makers like Pocket NC and Mori DMG. Then on the cutter side Harvey Tool and also SwiftCAR. Each represent a niche market place. And there is a myriad in between of legacy tired machines like Hurco's and Fadals etc. Fusion 360 is poised to help cover all of these needs. REACHING OUR POTENTIAL: We need accurate setups to drive our machine tools efficiently. And the material knowledge, machine knowledge, work holding knowledge, cutter knowledge, and cam knowledge to drive our efficiently drive machines.

-

So when your engaging someone new to machining and they want to machine Wax, Hard Plastics, Aluminum and Brass on a light machine what do you tell them?

WHERE AM I GOING WITH THIS... AND WHERE SHOULD YOU BE GOING?

Getting a job done ASAP is where it's at... Pretend you could afford a high end Mori Seiki Mill (I have 3 at www.rmbproducts.com) and suppose you were offered a job to make a mold out of steel for PFA plastic. You select to best steel and you are given a small window to get the job done. Not in weeks but in days. And your success in cost is based on delivery and breaking even+ That includes time to do the job and tooling.

Consider these two screenshots below. 

I'm describing the exact same machine, The exact same material, the exact same cutter, and I am describing even the same exact depth of cut. But look at the speeds and feeds. This is a volumetric assessment based on a change in radial load on the cutter. That is what we know as step over or optimal stepover.

Now the mind blowing idea is we accept a constant feed rate, and while we can also use feed reduction in corners we are not optimal. Wait we can buy an outrageously priced license of Vericut Optipath that a single license is enough to pay for Fusion 360 for 10 years no joke. So consider cut optimization not only as a 2D concept but as a function of total chip load as a function of cut depth WITH ideal stepover WITH  the CONTOUR of your part shape WITH your machines rigidity and WITH adequate workholding rigidity.

-

Well that is enough to pop a few brain cells.

DO YOU GUYS THINK ABOUT THIS STUFF??? HOW DO YOU ADDRESS THE PRESSURE TO PERFORM AND ACCEPT THE DEMAND?

I think about this stuff all the time...  Each one that replies is the winner of a chicken dinner, and all the beer you can drink tomorrow. you have to show up in Colorado. I suppose you can smoke freely here if your into all that. Ok on that last note stay righteous man and peace out.

[daniel_lyall]

For me it's a bit different with routers.

but the way stuff is done is the same you still have to have the correct chip load, the correct spindle speed and feed rate.

Having one of the 3 wrong if cutting wood, I would blow the cutter to bits or set it on fire, I have only blown a cutter to bits once, I have seen a fire on a router the dust extractor system went flying out the door and the dust bags went the same way.

the program used to do the cutting is a massive part of everything. useing fusion I can do 75% depth and 75% width with 1/4 inch and above cutters at 1250mm/min 50inch/min, this is down to useing smoothing and Feed optimization.

haveing all the elements correct the router fly's through the wood. haveing the smoothing set at 0.01 the computer is happy and keeps up on all toolpaths it just seems to work good.

haveing Feed optimization set to 75 degrees at 900 mm/min with 1/4 mm feed radius and 1/2 mm feed distances works good with Mach3 (it's only a 3rd order planner with a flat trajectory).

I have used smoothing at .1 and 0.001 the machine was not all that happy, but that is down to mach3.

Having the correct cutter makes a differences be it a up cut, down cut, high helix, high compression (or all 4 in one cutter), even just useing router bits, everything has it's places.

this is something you can only learn over time, every machine is different, like mark said morph spiral does not work well on his older machines, but on my router it is fine and that is with a program that is older than his machine, it's down to experiences, what is time and I am not a engineer or a machinist. all hobby.

the 4th and 5 axis stuff when it comes in it is going to be a very bissey time on the forum, unless they have the learning done for it.

for me it is not a biggie I have been useing 4th axis wrapping and 4th axis bull**** machining for a few years. 5th axis I would not have a clue.

god I ramble on 

[Tbaker25]

I just dug up this thread to find out what had been decided on.

Xander, great work on some numbers. That really helps us all out!

Randy, I was fascinated to hear some of your perspectives from being in the business. Thanks!

I dove into machining a few years back, to build prototypes for my company Steradian Laser Tag. Quickly got bitten by the machining bug, and now produce many of our products right on our own Tormach 1100.

I've recently ventured into injection mold design and manufacturing, because apparently regular machining was too easy. 🙂 I've got some great help along the way from my local injection molding shop, who I'm now friends with. (He used to do machining, but now just does the injection.)

I'm in total agreement with you on optimal stepovers. Using a 1/32" ball mill on a 0.5" long 1.5deg tapered shaft milling P20 Prehard doesn't give a lot of room for extra load. So having Adaptive clearing has made it possible to do that cut with that cutter, because there is a guaranteed maximum radial engagement. I'm happily doing very deep cuts that other shops would only do as EDM. But adaptive clearing has some significant limits. Calculation time goes WAY up when the pocket constraints are close to the tool size. The best guess I have is that the path generation starts out trying to do a straight line, and works it's way to a tight arc only as a last resort. But tight arcs are all the fit in a tight pocket. Maybe the coders could offer some insight?

I end up needing to make do with non-adaptive passes whenever I can, but if I make a mistake on a previous pass and leave a small scallop somewhere, that cutter is going away. I'd love to see Fusion expand it's focus on adaptive passes beyond large roughing cuts, but to constant load or just max load toolpaths for small feature machining, because frankly those tiny cutters need adaptive more than hogging roughers!

Tom

[RandyKopf]

@Tbaker25

Tom:

Hey thanks for that reply. Sounds like you are having a blast using machining. Sort of a pun on the laser guns. You certainly are getting an education doing all the things you're doing.

I took a quick peek at your company website. Looks like plenty of opportunity for sure. I'm really pressed on time at this moment trying to get a 10L box mold complete today. But you scratched on some aspects that are quite challenging. Coming from a Injection Mold making background and then doing Rotation Molds I've really had to push the limits on difference between the cutter's diameter and it's overall length. Typically we see 10x diameter in reach as common, but I get lots of jobs that require 20x / diameter. The reason I mention it is trying to use the largest tool possible does dictate things. I found you need some breathing room especially on contours regardless if they are 2D or 3D it's about a tool radius vs inside corner radius are going to have lots of issues unless you can understand cutter loads and correlating feeds.

Well I'd love to talk more but got to get back at things,

[Tbaker25]

Good luck on the mold!

Yes, we have a bast making the stuff. 

I started by doing aluminum inserts, but have quickly moved on to the simplicity of machining a MUD directly in P20 Prehard. Haven't had the volumes yet to need anything harder.

I find that G-Wizard gives me great data for proper loading of tools, but on really small cuts like this the Harvey Tools provides some good data on what te tool can survive. Adaptive passes let me more quickly machine out that material with given loads, since I can spec a max WOC. I really like where these toolpaths are headed. So much better than even 5 years ago, for a price that I could afford.

[Anonymous]

Great Video, So just curious, what Nema34 motors and drives does cncrouterparts (avidCNC) use for there systems.

[seth.madore]

That's probably a better question for AvidCNC 😉