How NOT to Design a 5 Axis CNC
ฝัง
- เผยแพร่เมื่อ 22 พ.ค. 2024
- Use my DIY 5 axis CNC router as an example of what not to do when making your own. We'll cover the biggest problems that my machine had and how I would fix them.
My previous video: • I Built a Custom 5 Axi...
Instagram: / mach_super
Blog: www.machsuper.com/articles/ho...
After getting questions on my machine after the first video, I wanted to answer some of those questions and also talk about some design flaws that you should avoid in your own projects. Most of them relate to the gantry, such as gantry shape, orientation, and style. I also talk about Mach4 and the Warp9 Ethernet Smooth Stepper and the issues it had.
To everybody excited for the mill project, I am working on it whenever I can, I've been studying full-time since January
Music: Let it Go by ItsWatR on Pixabay
0:00 Introduction
0:52 Gantry Shape
2:31 Gantry Style
6:46 Gantry Orientation
8:06 Linear Motion Guides
9:24 Motion Controller - วิทยาศาสตร์และเทคโนโลยี
As an engineer who works in industrial control systems and automation (30+ years) this is a fair effort for a non-engineer and you have done something that few do - LEARN.
Good pick up on the linear bearings it was the first thing I noticed, but then I do know why people use circular shaft linear bearings($$$). The linear rails are way more accurate and more importantly handle larger loads for their size. HOWEVER those rails are heavy and you can get into a lot of trouble on vertical axis if you are not careful (see below).
Before we get to vertical axis we need to cover an issue that confounds fully qualified engineers - *INERTIA RATIOS.* This is incredibly important with motion control if you want to get any sort of accuracy. All mechanical engineers learn about it in their very first dynamics class. I know because I lectured and tutored that class as a postgrad. *Its also something all mechanical engineers immediately FORGET after that class and sooner or later bites them.*
To ACCELERATE & DECELERATE anything requires force and in the case of an electric motor (servo, stepper,... etc) that is most commonly a TORQUE force. Yes I know about linear servos, but the same thing with inertia still applies.
FIRST (and super important) we are *NOT* talking about static loads we are talking about *DYNAMIC LOADS* so it involves INERTIA of the system. There is the inertia of the motor and the inertia of what its driving. Under static conditions the torque required to hold a load or start moving it is simply the gear ratio *BUT under acceleration and deceleration its a function of the gear ratio and the INERTIA RATIO SQUARED. So if you double the mass of the load on a motion control system you require 4 times the motor torque to accelerate and decelerate the same.
THIS IS THE CRITICAL POINT FOR ALL YOU AMATEUR CNCers
KEEP YOUR INERTIA RATIOS BELOW 10:1 as in the inertia of the load must not be more than 10 times that of the motor.
Its even better if you can keep it below 5:1 and best around 3:1.
Consider that at 5:1 you are 4 times better of than 10:1 and 3:1 is more that 10 times better than 10:1.
Think of it this way. Imagine if you have a light but reasonably powerful car and you hook up a large heavy trailer. Sure the car has enough grunt to pull the trailer, but what happens if you try and brake hard at speed? The Inertia of the trailer just dominates the car and pushes the car past any limit it might have or just jackknifes it.
In general you will find the inertia for a motor in the catalogue or manual. For all the other items in the system you will also find that data in the catalogue or manual. Companies like Siemens, Schneider and others all used to produce inertia calculators with their catalogues already in them. If you use one of there calculators and it wont let you enter you motors particular stats just use the motor out of their catalogue that is closest.
On your vertical axis its incredibly important to get mass out of the moving parts. There's a terrible tendency for mechanical engineers to put the linear rails on the moving part. I'm pretty certain you did what I'd recommend in having the linear shafts on the fixed side. The problem with vertical axis is that you are fighting gravity as well as inertia.
The other thing to consider is filling the base with epoxy and granite to add dampening mass. There is a continual issue with all motion control systems. If they are very light they have less dampening and from just the motor vibrations you can get all sorts of accuracy issue because the whole thing is shaking. This is why milling machines, lathes and CNC machining centres are made from bag castings. Its not that the castings are cheap its simply mass damping.
There are videos here on YT of guys who have used epoxy and crushed granite to get that effect.
If you really want to see how far you can take certain ideas watch Dan Gelbert.
th-cam.com/video/sFrVdoOhu1Q/w-d-xo.html
If you want to watch guys do amazing stuff
Stefan Gotteswinter
Robin Renzetti
Joe Pie
amazing comment thanks for sharing your knowledge
@@sebastianjuara thanks.
thank you very much for introducing resources of wonderful precision machines.
@@denberg0584 No probs.
Just sharing like so many others do.
very Nice n useful info Tony !
I worked as a cnc machine designer for five years. The points you have about the structure are quite right
I love your comment! Thank you for the validation haha! I had some imposter syndrome making this video, thinking "who am I to tell people how to engineer CNC machines?" 😅
@@machsuper, In your defense, you told us how NOT to do it based on empirical evidence and how YOU would build the next one differently, based onyour experience. Not as an "armchair engineer."
I can only imagine the time it took to not only design this but make the animations. Looks great!
A lot of time, but not too much for you people. XD
A column machine is the most rigid, a fixed gantry is the next best than a moving gantry. This is because a column only has one moving axis attached to the spindle, the fixed gantry has 2 and a gantry has all its axis’s moving the spindle. The con of a column of fixed gantry is not so much about moving the part but more to do with travels. A fixed gantry need to be twice as big as a moving gantry for the same travels. A column is even bigger. So I guess it comes down to what you want to make with your machine.
@@machsuper models are very detailed, welding and milled textures are awesome! What software do you use?
@machsuper A response like this get the subscribe. An admirable attitude/response for creators.
Concerning Gantry Style. I was an early CNC PCB Drill and Router Programmer/Operator in my first career (late 70s-early 80s). I programmed and ran several different styles of machines. I think the best compromise between rigidity and size is the Moving Table, where the machine is wider than it is deep, the table moves Front Rear, on the shorter axis, and the spindle(s) move Left Right, on the longer axis. All the fastest/most accurate multi-spindle machines used this format. It also provides the best operator access to the workpiece.
Honesty is so refreshing. U will succeed and the learning curve will get shorter and flatter. Good luck, I’ll be watching and waiting.
Thanks for the comment Chuck. You're not learning if you're not pushing yourself, huh? I'll be taking on pretty big challenges on this channel if things go how I dream them.
I love when people do share their experience with community. And you have already a lot of exp. That's so motivating.
Btw I like the colors in your vids. Keep them coming! ;)
Thank you! Glad to give something back, I love making videos and absolutely plan to keep going and making more challenging things - things no one has done on TH-cam.
Looking foreword to the new build. Interesting observations.
I've been developing 3-axis CAM software for hobby/DIY for signs/engravings/art over the last 5-6 years now with my eye on 5-axis toolpath generation just about the entire time as a potential future CAM software project. It never occurred to me that G-code won't necessarily be able to know ahead of time the tool offset (i.e. cutter length, or even just how far the cutter tip is sticking out from the axes of rotation). Granted, not all 5-axis machines are the same as the setup you came up with. The workpiece can be rotated on one, or both, axes. These are things I hadn't really thought about (I figured I'd figure everything out when the time came). Anyway, thanks for sharing, gave me a few more things to think about while I wrap up my 3-axis CAM software.
That's awesome! Are you working with BRep data structures or mesh modelling? I think I could manage to do something with mesh CAM, but BRep remains beyond my understanding for now, as much as I'd love to build my own. When making your 5 axis CAM (if you do), only the post processor will need to know the tool stickout relative to the axes of rotation, the CAM shouldn't need to concern itself with that. Or in the case of a trunnion style machine, the tool stickout is irrelevant and it's all about having either the post processor or the CNC controller itself know the location of the work origin relative to the axes of ration. Ideally the controller would do it on the fly, hence the TCP functionality mentioned.
Care to share the name of your 3 axis CAM software?
Definitely interesting. Intuitively I think the machine would need inverse kinematics as J6 robot arms do and it's the machine's job to know positions with multiple tool frame offsets that can stack. Experience wise the machines are dumb/smart however you look at it and 'should' go to whatever position you tell it to go to and the CAM side should just generate the toolpaths.
@@machsuper I went with an entirely novel approach for my CAM software, everything is heightmap based - you design your project by loading/creating multiple layers that you blend together to form your project's composition and then I generate toolpaths off of that, performing a sort of Minkowski sum of the heightmap and the cutter's geometry to produce the actual manifold that the center tip of the tool should move over. There are a dozen different toolpath types and only the 3D Contouring toolpaths operate that way, the 2D/2.5D toolpaths generate a series of contours for offset milling, trochoidal milling, profiling, etc... For 5-axis I am thinking the best way to go is to generate a sparse distance field of the 3D geometry to be cut, and toolpathing would operate on that distance field to determine the best way to approach the workpiece. Eventually I wanted to take it a step further and invent a process that lets you just slap any old chunk of material into a machine, it scans it, and then with a pre-defined set of tools it goes ahead and dynamically mills out a target model on-the-fly - sorta automate out the need for CNC expertise entirely. PS: my 3-axis sign/engraving CAM software is called PixelCNC.
@Charles Van Noland would you use a lidar attached to the cutting tool to make a 3d scan? I like your idea it would be really useful.
I answered some time ago allready on this, but somehow it was removed. (Don't understand why)
Search for rudy du preez 5-axis machine tools. Then a pdf will come in the top of the resulta. This paper handles the inverse kinematics of 5-axis machines including some code samples. That will probably answer your quest.
Cheers, catweazleMagic
This is the type of video that is missing on most peoples youtube project videos, where they go over what they would change if they were to do it over again.
Awesome work man!
some people have obviously said hurtful things to you. It only takes one to hurt you in the wrong direction. You mister are still going in the right direction. Dreaming is cool but doing well this is what puts you at the top most level. Keep going keep improving. And thank you for sharing input, ideas, explanations. Well done friend. Well done.
Thank you for that very considerate and heart-felt comment Daniel. I am a little confused though, do you mean people have left hurtful comments or people have said hurtful things to me in the past? Truth be told, there’s be very little hate in the comments of my videos so far. Almost none, in fact.
Based and the voice over it seemed you received criticism that was possibly overdone. I found that critics rarely “do” and doers rarely critique negatively.
This is great! Thanks for talking about your improvements. Very helpful.
you can also double the gantry, so instead of the Z axis being mounted to a single face, it can be mounted between 2 faces. This would double the guides and control system.. but it would also increase the the gantry's mounting surface.
Yes this is how axyz machines are build. Having two faces also allows the gantry to be small, saving cost.
Awesome video 🙌 insane that someone who hasn’t studied can be this brilliant!!!!! What an inspiration, shows how valuable being passionate is!!
Wow you animation and very illlustrative and it makes everything very clear. I just build a 3018CNC and you are really illuminating the way forward for me !Thanks !!!
Thank you very much mate! Good luck with your machine moving forward!
thanks a lot for your thoughtful advice, i am still in my progress of designing my first 3 axis cnc mill and its pretty cool to hear this.
Thanks for the comment Mubarak, and you're welcome. Good luck with the machine build!
Aesthetic vid! Loving these mate, keep them coming! Do what you can, there’s never enough time for everything I want to do!
Thanks Chris! I'm doing my damnedest to keep the action happening! The second half of this year will be full-time working on Mach Super in a new shop.
kkkkklmkklmk
Great animations. Impressive!
great video! love the animations
i'm execited. Waiting for next video !!
I'll get it out asap! :)
*From a former mechanical engineer:* Round tube is only better at rotation _as a single member_ and box beams are still pretty good at that. I have shifted from engineering to working on industrial machinery for a living - most machines don't use round tubes. They are much more hassle to mount and align everything properly vs just slightly upsizing and using box/rect beams, and they aren't single member designs.
When trying to resist torque, depth is the 1st most important factor. A piece of paper can be bent/twisted easily. To some degree, widening the paper will help resist, but simply making the paper thicker does far more. And if you think of cardboard, it maximizes this effect for minimal weight by concentrating the material on the outsides, then linking them together with an approximately triangular grid.
The thin sheets by their selves aren't strong in that way. The corrugation layer isn't much stronger either (look at how corrugated roofs can collapse). But once they are tied together as a unit, they become extremely strong for the weight.
A box beam with internal triangulation will be very strong. Also, 3 box beams in an approximately triangular arrangement plus a bunch of gussets will also be very strong, usually much more twist and flex resistant than just 2 heavier box beams.
I notice you didn't mention racking. That is another reason why extending bearing spans is important.
As far as wide vs narrow spanning, you are correct, though it may make for a much more compact machine. When you consider that the most effective control on flex is wide bearing spans + rigidity of members, you already want as wide a bearing span for the gantry as possible (really for any of the axies). Being slightly longer depth is preferable in most shop circumstances vs making an already long machine even longer.
& When you think about the practicality of doing that, going for as heavy rails and screws as you can afford, cost for longer screws and rails shoots up exponentially. But then all the gantry itself gets more expensive and heavier if it's longer and may need bigger motors, etc. Decide where to spend your money.
For a home garage, I really like the idea of a wide spanned machine that offers the option to feed sheets through. If you imagine cutting a 5x8 sheet (or a rectangular shape like a custom door), it is more steps, but *you could use a much smaller machine if it's spanned the wide way and cut & move in several stages. A short span machine would require the _short direction_ to be at least 5ft wide + the rails etc. A wide spanned machine would only require the wide direction to be 5ft+rails etc. That means walking space around both ends of the machine, except when cutting long sheets. That's a win for a home shop, imo, but if you are cutting large sheets all the time, then just make a bigger machine and avoid the hassle of cutting and moving stages.
Back to rigidity concepts: when mounting the axis on the beam, you want the rails as far apart as possible, so they resist tip flex perpendicular to the beam.
There is the possibility of going with rack and pinion for a big cost savings. If you want less backlash, helical gears are an option.
If you do rack and pinion, I highly recommend mounting them facing down, so chips & stuff fall out of them rather than in to them.
If you have vibration problems, fixing rigidity is the best solution.
After that the common cheap way to deal with that is filling sections of the gantry with cement (at the cost of weight). Cast material is also much better at dealing with vibration (scatter & absorb effect), but not practical for a diy system.
Those renderings and animations are great!
Cheers Ben! I put a lot of thought and effort into them. 😊
Awesome bro. Good work
Great information, I wish you were around when I got started nearly 22 years ago. Parts for CNC machines was crazy expensive. Drivers cost $800 plus each because they were industrial grade not that cheap crap from China. Huge learning curve none the less. Awesome video keep up the great work.
Super, thx. I also do not have lots of money. That is why I make my own linear rail, i just use ptfe sheet as dry bearing. You could use heat to shape it as needed. It will need some kind of backing plate. Thx mate👍
nice work man! very cool project
Thank you Daniel!
Thank you for sharing your knowledge.
Thank you for watching ☺️
Great observations! I came to the same conclusions after I built my first cnc from an aliexpress kit. My latest one includes what you talked about like linear rails, square tunes for torsion stiffness etc. I would also recommend tooling plate aluminium for the bed as it can be used as a good reference for tramming.
I went with premade 'z axis' for all axes, with the y axis turned on its side, so 2 rails on each side, 4 total and 8 bearing blocks to make it really rigid.
I'm interested in adding more axes to the current 3 so I'm looking forward to how you progress your build!
Wow sounds like a cool build! Thank you for the comment mate. I’m a little slow, but I’m working on more videos.
Great video!
Another great video man.
Production quality is right up there, I'm guessing the 3d animation took some time?
Thanks again man for sharing 👍
Thank you Drew! Yeah, the animation is usually the most time consuming part, but it's invaluable to explaining things clearly and aesthetically.
Thank you. Looking forward to watch more on this. I’ve been looking at improving my 60”x49” CNC I purchased in 2018.
I am looking at linear rails in place of the rounds like you have and increase my Z height.
This is a very solid and heavy all steel construction CNC. Hoping to get more info and ideas from you. 🤙🏼🇨🇦
Thanks mate! There's a secret router project that's come into my life lately, so you may end up seeing more router content from me.
@@machsuper this video was quite informative and helped to cement the knowledge that cheap round rails are garbage compared to linear rails.
My CNC has some great build quality mixed with some odd ideas that I want to address.
Great video, I designed my own CNC router for sale the entire thing is built around ball section with dual linear rails on the X, Y, and Z - It makes all the difference for accuracy if you get the box section ground beforehand
Nice! Yeah that would be really good to get all the mounting surfaces ground flat, as long as you can keep them flat after welding.
@@machsuper I've completely avoided welding in my design and relied on fasteners but if you can find a shop that can grind a welded assembly it's definitely the way to go
Im have to build my stiff gantry mill up to August and i had already done 90% of 3d project. great vid!
What is your mill made from? Steel?
@@machsuper epoxygranite base, steel moving beam and aluminium Z plate. Im going to cut aluminium, steel and every metal which this beauty could handle. i had finished MES simulations for Y beam ( filled with epoxy granite ), Z plate and working on base to simplify the MDF molds Im going to use for this base manufaturing.
These were all great tips. Thanks.
You're welcome, glad you enjoyed. :)
That's some impressive work you've done. I'm happy to see someone taking interest in the mechanics for a CNC machine, my self being a senior mechanical designer. I am designing a 3-axis router or rather portal milling machine and i have come up with a design somewhat similar to your concept. If you want to I can share some of my solutions. My focus has been to make a very rigid machine for milling steel and hopefully also hardened steel. Another aim I have is to make the machines outer dimensions small in relation to the work area.
I believe that the solution that you call "moving bridge" is the most promising one. Choice of bearings and placement of ball screws is essential and using twin ball screws is the way to go for both Y and Z-axis. The thrust bearings and the homing sensors are also of big importance if you want to achieve high precision.
I was also impressed by your approach to the project planning, working the last 30 years with product development I have taken an interest in conceptual work and how to manage a project. Please let me know if you are interested in exchanging some ideas, I think I could benefit from it too since I'm not so good at software and electronics.
Hi Leif, what gets me particularly interested to know more about your design is the fact that you’re aiming to achieve hardened steel, and also the fact that you mention twin ball screws on the Z axis. If you’d like, send me a message from my website contact page and we can have a bit of a chat: www.machsuper.com/contact
Whoa, another input. I might explore the idea of hardened steel. I am quite familiar with heat tempering, why not I have thought about it. Thx!
Esto está a otro nivel, nivel Dios.Me encanta ,tratare de emular su grandioso trabajo.
Gracias amigo! Lo siento, mi respuesto es muy tarde. Suerte! :)
Lo siento, estoy aprendiendo espanol.
@@machsuper yo le enseño Español
I wonder if the ESS/Mach 4 problems are due to the fact that Mach 4 is designed in the US where the axes are all about moving the part, not the tool, i.e., 5th-axis trunnions.
I can tell you as someone who has designed and built a fixed-gantry CNC machine and used moving gantry machines, you are partly right about the ease of making the table. The challenge is in squaring the table to the gantry since there is so much space between. Not impossible, just more tedious and/or it requires tools that are less common. (Who owns a 1m square?) It also increases the rigidity of the gantry, and typically allows for faster movements.
Very cool that you identified the problems in your design. Even more so that you shared it to help others avoid the same mistakes!
Can you pls do a video more in depth about how you built that 4th and 5th axis? it looks really interesting and well made.
this is very high quality content for a channel with only 4.36 K subscribers
Cheers Kyle! It's growing pretty steadily and quickly, I'm enjoying the journey.
Hey! Our family business had a great experience with the Okuma Genos M560 VMC-s and thus I highly recommend You study the frame of that machine as they made a perfection of a fixed column gantry construction. I loved the video, You are point on with your observations.
Cheers for the suggestion Ferenc! I like that style of machine a lot. I think moving parts would be very light, good for acceleration, yes?
Wow very good video sir.. It is like talking to me :)
Wish i have those days back, hopefully will do that very soon.
Thanks buddy. I like to make the video feel comfortable and easy to watch.
I thought long and hard about the design of my CNC before I built it, and addressed all of the issues you spoke about. But it wasn't without a lot of time in SketchUp and significant discussions with a friend about the pros and cons of each design. I was looking for something different to the well worn path, and I hadn't seen anything like what I come up with (8 years ago). The only significant difference between yours and mine is that I boxed the the X axis gantry in to make it more rigid. It trades view for stability but it is one I am willing to make. Looks like you're just down the road from me.
It’s very easy to end up spending many many hours thinking about the design, isn’t it? I block out the essentials and even some smaller details in Blender these days because it’s much faster and simpler to modify any part of the model without breaking other details. Then I go into a parametric modeller (Fusion360) to get every detail I need.
Did your gantry hold up like you wanted?
@@machsuper hell yeah, and even more hours drawing them up to make sure they are going to work as intended. I use SketchUp predominantly for my drawings and reluctantly starting using components and groups to stop parts sticking together. I looked at learning Fusion 360 but knew they were eventually going to stiff people and start charging or remove functionality. Sometimes you hate being right.
For the most part yes, but using two NEMA23's on the Y axis it can cause racking issues if I have problems with wiring or software glitches, a couple of years back I started on modifying it to be driven by a single NEMA 34. But from a rigidity stand point it was a good call, even thought it does complicate seeing things occasionally. My main goal was to get the spindle in the middle of the X axis gantry frame to make it as lever free as possible (was originally thinking 5 axis and supporting both sides).
Bearing load was my primary concern of the design and I have 340mm between bearing end to bearing end, and the updated version of my end plates is 360mm. Like you I made an error in not allowing myself enough length in the bed frame to allow for a decent width gantry. I am very close to getting 1200mm (once finished) of Y travel but I am just a fraction shy. I used 75mm box for my main frame, and some 65 x 35 for the bed. I wish I had made everything out of larger sectioned box.
The 3D images on the video are SICK
You say you are not an engineer but you could teach a thing or two to some ;-) really good discussion and very usefull !!
Wow, thanks! Glad you enjoyed it. :)
Would you consider putting up a brief tutorial on how to make these graphics in blender? These are awesome and would really help me convey ideas!
Thank you for sharing.
Thank you for watching Mr. Flowers. 😊
I like it. "Intuitive eye" 👁
Really nice Video!!
Thank you Maximilian!
those animations are sick
Cheers mate! I put heaps of work into that aspect of my videos. Many days and nights leaving my computer to render.
What tool are you using to animate those gantry type 3D views in the video? ^^
Nice video, wish you success
Thank you :)
Can't wait to see the mill from what you've done with this one. Will it have Tx/Tz movement as well? (just been looking at Peter's latest videos on Edge Precision).
Thanks man. I suppose Tx/Tz/Ty jog features will depend on whether Dynomotion supports it or if it can be programmed in. I really don't know if it's already got that ability. It will be nice if it does have it.
Great vid. 🎉🎉
Thanks Dog! 😄
Excellent illustrations for the different types of gantries and movements of the various axes. One downside from my experience selling gantry style machines is the ability to move heavy workpieces using forklifts onto the bed of the machines. If the machine has structure above the bed, a forklift will not be able to lower a heavy workpiece onto the bed, but that types of machine can be more easily design to be more rigid, so there are a few tradeoffs.
A single ballscrew for the table style can cause racking, so depending on the width, you still may need two ballscrews at the extreme sides fo that table.
I'm sure your mentioned this in a previous video, but what CAM program are you using? Fusion360?
Your overall mechanical and strudtrual solutions for your machine is great and I can't wait to see your development. Keep up the great work.
Thank you very much!
Both good points. I realised later on that I should have mentioned that the moving bridge is excellent for access. All pros and cons.
Yeah you're right about the racking on the table, just depends on the aspect ratio of the rectangle, right?
Indeed, I've only ever owned Fusion360. I used NX CAM at my previous job, but it was an old version and it truly sucked to use.
Thank you for the nice comment! Can't wait to get more videos out.
Yes, the best gantry is a double gantry. I made one, but of course you then are limited more, but there is no torsion at all since the spindle hangs between two bars. For 4th and 5th axes I use separate systems, so the Z spindle is always vertical. (also there you have your torsion problem)
I've got a feeling you're going to see some channel growth real soon.
Nice job.
Thank you mate! 😉
Thanks for all the videos! I'm enjoying watching them!!!
I think that you never mentioned how to use limit switches in your B & C axis. If you don't have any limit switches, how do you homing those axis?
Can't wait for your future series!
Thanks for watching them!
There were no limit switches on the machine, actually. I tried to set them up on the linear axes, but I was getting a bunch of interference issues. Eventually I just gave up and operated without them. The travel was large enough that it never bothered me much, I just made sure to have soft limits setup. The rotary axes also had no switches, I just put a straight round bar in the spindle and checked it against an indicator and kept soft limits on. Something to home the rotary axes would have been nice though.
@@machsuper In your next machine, try to use shielded wire for the motors. Unshielded wire can cause EMI resulting in false positive alerts on the switches.
@@stav3322 I was using shielded wire, still had issues. I managed to fix major interference by fixing the grounding to the 5V power supply, but never fixed those limit switches.
@@machsuperI think I could help you with this, i know a bit about electronics. I could even send you free of charge if you want, shipping would be rather cheap as I am close by in Indonesia.
Great JOB !!! more content bro =)
Thanks man, I'm goin' on it! 😃
Awesome project mate! Did you make all those animations? They're a seriously nice addition.
Cheers Sam! Yup, all made by me in Blender.
Thank you for making this content. It's highly educational. What controller do you use?
On that machine, I used Mach 4 with the Warp9 Ethernet Smooth Stepper, but as mentioned, it was a poor choice. I probably will never use Mach again (ironically). From now on it's likely to be Dynomotion or LinuxCNC.
Awensome Bro.
Your frame will always have some flex. Even 500,000 machines show this to some extent. If you want to get that out completely, it won't be easy, especially if you want to machine materials like aluminum or metal.
You have to find the right balance between the stability of the machine and the right programming. You can get a lot out of programming.
Have you measured the clearance of the milling spindle?
Your frame will eventually be stable enough then it will be the spindle that is next.
But very cool project keep up the good work. You have more intusiasm than some mechanical engineers I know.
Certainly every machine has flex, just a matter of eliminating it enough to suit your needs, right?
What do you mean when you say ‘spindle clearance’?
Thanks for the comment, man.
@@machsuper Exactly. If you work mainly with wood, a little flex won't matter and you'll save a lot of money. You probably won't even be able to measure the deviations. If you want to work metal at full throttle, it will be really expensive.
With spindle clearance I mean the concentricity of your spindle.
When I move the spindle on a $1000 pillar drill at my local hardware store, it has a lot of play and the runout is bad. You can see that the drill is not running true.
Your spindle doesn't look like it, but with cheap spindles there can be flex. It is not always the fault of the frame. Especially after a crash you should check this. In woodworking it is mostly unimportant in metal important if you want to keep tolerances.
You can also compensate it with Milling Same or Counter Direction.
The tool can also make a difference.
Tools for metalworking will never run as well in wood as extra cutters for wood. These usually have a different angle and different coatings. In the beginning, carbide tools are perfectly adequate. If you want to use the machine to its full potential, it will not be enough.
High performance wood cutters are therefore equipped with diamond. Unfortunately unaffordable for private customers.
Kflop looks like many are using it today. Not so many when i started my project. A friend of mine was featured in Linux for his 5 axis a couple years back with twin tool changers. Im already invested into linux. I even have a centroid acorn but 2 limited to 4 axis.
Wouldn’t surprise me to change my mind by the tome i get back to my project since i may have out dated hardware.
I have a 5i25 and 7i77 if i recall correctly without digging it out of storage.
Gi Glenn, how long ago did you start this project? It seems to me like if you already have a handle on LinuxCNC, there’s not a lot of strong reasons to jump ship unless you’re just curious to try out Dynomotion.
just brilliant! subbed clicked the bell etc etc v v v cool man
Thank you Julian.
8:05 Can you elaborate on what's wrong with the round SBR rails? I thought they were the superior design aside from being heavier and taking more space. The ability to adjust preload on the bearings with set screws is a brilliant way to get high precision while keeping the cost down. MGN rails have to be perfect from the get-go or else they're useless. The ones I bought were all over the place. By mixing and matching the rails and slider blocks I was able to get them all usable, but if you weren't buying a lot all at once it would be a total crap shoot, whereas SBR rails can always be adjusted. And adjusted again as they wear. I've only used SBR for the Z axis on a 3D printer, but they feel solid enough for a router.
Two major negatives with round rail.... One is that the rail support is a tee so that means the bearing is 'open" which allows a lot of flex. Another is that the ball bearings are round and the rail is round so you get high point contact loading. Profile rails are, well, profiled. The bearing surface is curved to 'fit' the round bearing. This allows for more preload and therefore much more rigidity. They are also 4-way equal load rated which a 'c' shaped open round bearing can never be. There's really no comparison. If you had to mix and match rail and carriages then you bought knockoffs or just low quality. Or the mounting conditons weren't adequate.
Jevon, you said it better than I could have. Thank you.
I have the same rail and it's a bit curved in height: on a lenght of 3 m I have a height difference of about 2 mm.
It's interesting hearing those infos about a CNC gantry while in the 3d printing space people try to make them as light as possible to minimize inertia for higher speed without loss of accuracy, going as far as making them out of carbon fiber. Also we know that "moving bed" design as "bed-slinger", which is now considered more or less obsolete. Especially now as modern printers modify motor movements in real time to reduce vibrations ("input shaping"), and it turns out having the resonance frequency change as the printed object gets heavier on the bed is not ideal compared to moving a much lighter gantry.
All the immense effort that gets put into making these super heavy, stiff structures for CNC mills all comes down to the simple fact that the tool experiences forces while working. It's the key difference isn't it? It's the key difference between why a 3D printer should be carbon fibre and a mill cast iron, steel; granite or some heavy, stiff composite. I have wondered about building a machine out of carbon fibre though. If it's so strong, and light, could it make for a great high speed mill if you can work out the resonance?
Interesting point about the changing resonance in the 3DP bed as the part gets heavier. You clearly have looked into the details of 3D printing more than me.
Great video! Excited to see progress updates on the 5 axis mill. I'm interested to hear about your experience with Dynomotion and the KFLOP board. Are you planning on adding any direct linear or rotary encoders for your machine axes?
Thank you mate! I can't wait to dive into the Dynomotion stuff. I probably won't put linear encoders on, I've got closed-loop steppers, that's enough for my budget. The rotary axes... I haven't thought that through yet. I may well put encoders on them, first I need to figure out how to interface them with Dynomotion.
Studying, working and building a 5 axis CNC. Someone's quite the high achiever, hey 😆 looking forward to the build videos. Nice graphics, what tools do you use for the animation and video editing ?
Evidently doing all those things pretty slowly latley. XD
Thank you mate. The 3D graphics are done with Blender. Video editing and 2D graphics usually done with the appropriate Adobe software.
@@machsuper Blender is awesome 😍
Masso is also an amazing option for 5 axis machine control!
I can't seem to find any mention of RTCP functionality on the Masso controller, otherwise I'd be more interested to give it a go.
@@machsuper After nosing around on the Masso forum i can confirm it does not currently support RTCP. Though they do seem interested in incorporating it. I have a masso G3 touch for my 3 axis machine with a future rotational axis planned. And the Masso is a breeze to use, wire and 'program'. I actually have build the machine with the gantry over the long axis as you suggested not to do in your video. But this does allow me to put full sized sheets in my otherwise only 1/3 sheet machine.
@@DikkieKlijn ah thanks for confirming that about the RTCP.
As for your gantry, there's definitely reasons to span the long distance sometimes. It's just horses for courses, right? Generally you just want to improve rigidity wherever you can.
How do you figure out what’s each component also would love to learn how to design and build an articulating robot
what is the software that you use for this animations, that so clean
i need some 3d animation for my project in my class
looks great
Thanks for the comment mate! I use Blender for all 3D animations. Good luck with your project. ☺️👍
thanks for the video man just a taught would you ever consider a masso g3 controller ?
Probably only if the machine is 3 axis. Anything above that and you can’t customise the kinematics well, from my understanding. Thanks for the comment Steven 👍
I'm in the process of building a LinuxCNC with Mesa card 5 axis machine. Can talk a little about the CAM workflow side of things? I'll probably go the Autodesk Powermill or Fusion 360 with machining extension route, but I'm uncertain if the gcode will play nice with my sender. I'm curious what issues you may have had in this regard. Cheers!
I haven’t actually done any 5 axis work on it in 6+ years. But if you haven’t already watched it, the video linked in the description does a lot of talking about the CAM side of things. The most important thing is to make sure you can get the RTCP setup and working. LinuxCNC has good support for that, so I suppose you’ll just have to make sure you have a post processor that suits LinuxCNC and your particular needs. What type of machine are you making? VMC style with trunnion? Portal router like mine?
A how to or a learning / tutorial series on CNC controllers ..in thos case the dynamotion One ..would be fantastic
I have been thinking about making a course for beginner CNC and whatever other areas I think I can offer unique value. I might get the time for that later this year.
@@machsuper sweet
How would you fix box sections into the gantry beam - obviously hard to access. Maybe bolt them in? Plug welds? Or would that risk warping?
Yeah I’ve always assumed I would drill holes in the box section that line up with the gussets and plug weld them.
@@machsuper cheers, I'm making a gantry right now and been wondering about this. I could easily plug weld in some plates and skim the whole thing in the mill after, as long as I could get it aligned straight enough in the mill.
@@Joe_Bandit I can imagine that's how I'd go about it. If you have the ability to stress relieve it, do that after the welding. But that's a pretty serious step that many people can get away without doing, depending on their needs.
thanks for the perfect explanation! nice job also with the animations. I was wondering how did you do the transitions such as the one in t=119 did you use blender? I would love a video about you explaining how did you do those animations. Is very nice to demonstrate modifications. I also think it would be interesting showing some 5 axis toolpath planning in fusion 360, it would help the channel to grow. there's plenty of it around but is a more generic subject, so more people would be fished to the channel and be able to enjoy the high quality content. I'm also making my channel later I pass you the link when I create it. For now I'm trying to evolve on video making.. feel my videos are not good enough to share yet. I suck at everything related to content creation 😂
Thank you for that William! I’ve been asked a lot about the animations. I’m not interested in making 3D animation tutorials, but I’m happy to offer tips when people ask how I achieved something. I do all my animations in Blender, I’m lazy and tend to do as much of the texturing as I can with procedural shaders as well. The trick to the animation around t=119 was Shape Keys. There was a base model (original design), then a shape key of the thicker, updated design. Then you can animate between the two.
I definitely do want to make a lot more videos around creating and running 5 axis toolpaths. I’ll be getting into that in the future as it comes time.
If you want to make videos, make sure to get your first video out. It will probably suck. Mine did, but you just power through it.
great vid, commenting for the algorithm :^)
Hahaha! Cheers Stimpy. Appreciate it. 😉
Funny I thought that a moving bridge would be more useful/rigid because of the part never really lying flat but rather in a vice or on a higher surface if you have a big part etc... Keep in mind I'm more interested in a "big mill" that a "2D router"
Well it's all a matter of what you need from the machine. If you need really good access to load large parts on the machine, a moving bridge is a good option. In all cases, you can build the rigidity you need by designing it to suit.
What do you use for the animations?
3D animations all done in Blender. :)
Your 3D animations were awesome! What software did you use to make them? Is this inside your cad package or did you have to use something like blender?
Thanks Tyler! All 3D animations are done in Blender.
what software u use to make the machine animation?
I use Blender for all 3D animations.
My 8 axis motion control crane uses a kflop for motion. It also has a L head with harmonic drives, just holds a camera instead of a router.
Is that it in the cool pan/tilt head video?
Have you seen the channel routercnc on here? He's got an awesome build series on a flipped gantry (gantry spans the short width) and a dual gantry beam setup. If you watch some of his follow-up videos where he's using it - it absolutely shreds ali with a 1kw (I think) spindle and not a hint of chatter. He also has Fusion files available that are really thorough. Probably the strongest gantry style machine I've seen well documented on youtube
I just watched the video where he poured the levelling epoxy for the gantry. The machine looks really nicely made. One thing that concerns me with the dual gantry is that the two beams spanning across aren’t tied to each other. I see it as an I beam with the centre bit removed. Clearly it’s done the job for his needs and it looks great. I just don’t see it being a better solution than one large square shaped gantry. Do you understand what I mean?
@@machsuper yeah gantry flattening is a whole other topic that a few people on here have come up with some really cool takes on! I think I understand your thoughts on the I-beam - kind of like having two plates floating on bearings and they would sort of slide relative to each other when the spindle torques the gantry?
@@drivenba torque on the gantry like your saying could be another mild weakness. But I'm mostly concerned about horizontal flex in the direction the gantry travels in. Imagine pushing horizontally on the gantry, those beams will be weakest in that direction because they're the thinnest that way. They're not connected to each other except at the ends, so you're missing out on a bunch of additional rigidity.
@@machsuper yeah that makes perfect sense - especially with how you had talked about boxing the main beam as a design improvement. I could see a box-welded, larger, square single beam being much stronger than hollow rectangular dual beams. It'll be really interesting to analyze these different configurations with FEA - that's on my list for the future. Another cool channel to check out regarding beam improvements - Bryan Howard. He's also doing a really interesting flattening technique with a webcam and some software he wrote!
What do you think about rack and pinion instead of ball screw on X and Y axis...
If they are low enough in backlash for your needs, then I think they're great. I've never used them though, so my opinion is purely based on speculation. Perhaps the best solution for long travel lengths is a ballscrew with a static screw and rotating nut.
Nah look at what the big bois are doing ..you wont see rack and pinion only big ballscrews
What software was used to make these animations?
Blender for all 3D animations. :)
i wonder if it's possible to make a moving bed design, but effectively the opposite from the depicted model, where it's combined with the moving beam, so you handle the x and y with the beam, allowing the mounting on the components to be very rigid to facilitate rapid motion of the cutting head, but handle the z axis with the bed, raising and lowering the whole workpiece, that way the mass of the piece would have less of an effect on the mechanics of the system, and you could use a set of robust worm gears at the corners, a series of hydraulic pistons, or some other method that hasn't occurred to me at the moment, as the means of lifting the piece, whichever permits the highest degree of precision.
Unfortunately that’s not a good approach. Imagine machining the inside of a deep bowl shape about 300mm deep. The gantry will collide with the rim of the bowl unless you make the spindle permanently hang very low down from the gantry, in which case, you may as well just put the Z axis on the gantry. Also, putting four drive mechanisms on the bed is a lot of extra cost and a lot more opportunities for failure.
It’s great that you’re thinking outside the box though, keep doing that. 😊
makes sense, i was thinking about things a little backwards, the form factor i suggested would work better with additive manufacturing than subtractive.
Nice video! What do u use for the animations damn?😮
I use Blender for all 3D animation. :)
@@machsuper thanks, they look awesome
i made a moving table design, and i really have no complaints... i get the most stifnes out of that type of design...
What did you make the chassis with?
@@machsuper A combination of steel profiles , steel bended metal and cast aluminum whereever the rails go to reduce vibration with material combinations.
how did you make the 3D
may i ask what software used for animation lol. looks great THX
Thanks flower little, I use Blender for all 3D animations. :)
What about doing what i did on mine, fill the gantry with Epoxy Granite
It does not add TOO mutch weight, but man it does wonder on both flex and resonance
I have no doubt it would improve it, especially with the resonance. But I would still, sooner rebuild the gantry with a large square section as shown.
Are these cad renderings/animations done in blender?
Indeed they are!
From building milling machines professionally: If you can - by any means - make it more rigid, make it more rigid! Make it more rigid until there is no crane that can lift it, no truck that can carry it, no servo that can move its axis, no customer that can pay for it and then back off about 15%.
Hahaha yeah, well said. Rigidity is king in machining.
I always thought it would be more rigid to have two connected lightweight gantries, one on either side of the Z. That way you’d cancel out any flex, in exchange for a bit of X range.
I’ve discussed with someone on another comment on this video about why I think it might not be that great. I think, if you’re doubling doubling the same size gantry that you were already using, then of course that will improve rigidity. But if you’re taking up that much space with two gantries, you may as well take up the same space with one very big gantry and stand it up right. The two small beams are losing out on a lot of potential horizontal stiffness by not being tied together for their entire length.
It would be awesome to see a tutorial covering the animation side of the project. (Or if you saved any screen capture a timelapse maybe?) ☺
It’s really not in the scope if the channel, but it has be an extreme point of interest from people, so I’ll have a think about it. Thanks for the comment Burhan. 😊
@@machsuper I've been working with Blender on and off for some time now and it is always fascinating to see new approaches on product design and realistic machine simulations. (Thx for the reply)
You are an absolutely AWESOME guy. Please create a Patreon or something so we can at least buy you coffees !! please keep it going.
I've been thinking about ways to make a bit more money from this to help fuel it along. Thank you so much for your nice comments, mate. New video soon.
any plan to release the cad?
Not at this point, not for this machine.
Hey check out Teknic Clearpath Servos. I haven't used them myself but I will if I get the time to build my own mini CNC.
Oh yeah, I’m well aware if them. I was SO close to using them for this build, but The budget was already being stretched with the closed loop steppers that I bought with the conversion kit.
cooooooool
Try masso cnc for the controller its a much better option for what you are trying to do
A few people have suggested that. It can't do what I need. It doesn't have RTCP as mentioned in the video.
@@machsuper the multi axis variant does I'm using it as we speak