Welcome to Our Community

Some features disabled for guests. Register Today.

The Phox

Discussion in 'CNC Mills/Routers' started by Kevon Ritter, Nov 18, 2017.

  1. Kevon Ritter

    Kevon Ritter Veteran
    Builder

    Joined:
    Apr 30, 2015
    Messages:
    597
    Likes Received:
    293
    With these two negatives, I decided to come up with my own design.
    1) I despise belts, therefore I do not like the Ox.
    2) I built a Sphinx, but the serviceability makes me want to shoot my foot off.

    I have designed quite a few V-Slot based machines, lasers, printers, hotwire, and far more rigid moving gantry mills. But this particular design is something I'm doing for a friend. That's the reason why I'm starting this thread.

    The idea behind this was to build a light duty machine with precision, reliability, and complete serviceability. The end result seems to be a cross of both the Ox and the Sphinx, but there are only so many ways you can skin a cat. The particular model in the picture below would be the intended light duty small frame variant. This is just my opinion, but with just a slight modification, this thing may have the potential to outclass both machines. (I'll redesign later to emphasize that idea.)

    I started with a NEMA 17 belt version to get a good idea. This was intended to be a very cheap build option. At only 500mm, the downsides of a belt shouldn't show. This started with a 3/16" plate thickness.

    From there, I modified it to a NEMA 23 lead screw. The switch alone added somewhere around $100 extra to the build when compared to the belt, and that's before electronics. As pictured, it's quite a bit bulkier, but I already have ideas to change that. The plates where beefed up to 1/4" (not applied to Z axis yet), which is what the Sphinx is using.

    Both models are intended for 52mm spindles. I think that goes up to about 600W for brushless.

    I think I'll call it the Phox. :p I'm unsure of when or what this will be by then, but this will happen. Let me know what you think.

    Kev
     

    Attached Files:

    Grantman likes this.
  2. Kevon Ritter

    Kevon Ritter Veteran
    Builder

    Joined:
    Apr 30, 2015
    Messages:
    597
    Likes Received:
    293
    Here's the offset setup. The x offset can be achieved with a 150mm pitch length belt. The standard belt is 210mm.

    And of course I had to start playing with something bigger. The left is a 1000x1000. The right is a 1500x1000 with a double C-Beam gantry. I would have to change the plates for that one though.
     

    Attached Files:

  3. Kevon Ritter

    Kevon Ritter Veteran
    Builder

    Joined:
    Apr 30, 2015
    Messages:
    597
    Likes Received:
    293
    The Mini Phox is, for the most part, complete. The only missing parts are limits and cable carrier, but that won't be difficult to add. The last two pics are the bigger variant.
     

    Attached Files:

  4. Florian Bauereisen

    Florian Bauereisen Well-Known
    Builder

    Joined:
    Feb 24, 2014
    Messages:
    188
    Likes Received:
    60
    HI
    having your lead-screw on the backsideof the x-axis is, in general not the best idea... in your case it is very bad indeed as, in order to move your z-axis it transferes the movement only via the bolts and that is to be called flimsy..
    I often wonder, where is the point of running lead-screws as long one keeps on using wheels directly on the extrusions, opposed to real linear rails and wagons? Which , in the end doesn`t matter as your twin x-beams hanging free in the air between the end plates. Once your gantry moves these will flex and vibrate.
    Z- looks like it could be much shorter - introducing a lot less flex.

    While looking build strong these are major design flaws- to my eyes.

    You sure to get that short belts? Using bigger sprockets is fare more sensible as you can transmit far bigger loads (your torque) and have a lot less stress on the bearings. ( one of my biggest criticisms on the Omega-drives like the ox`s)

    Sorry for sounding so negative but i am just trying to save you from frustration.
    Building a machine (not a working thing -but one for specific goals) is more than using strong parts. Momentum, flex, levers, over- vs under -engineering...

    You are off to a good start so keep at it.

    greets

    flo
     
  5. Kevon Ritter

    Kevon Ritter Veteran
    Builder

    Joined:
    Apr 30, 2015
    Messages:
    597
    Likes Received:
    293
    having your lead-screw on the backsideof the x-axis is, in general not the best idea... in your case it is very bad indeed as
    I along with many others haven't had an issue with the lead in the rear. I used to think the same thing, but as long as the linear motion is solid, this isn't a concern. A good friend has a 4'x8' and a 5'x12', both by Thermwood. Their y axis are driven by a single ballscrew located to a single side. I can't even fathom the mass behind them, yet they can interpolate 1/4" diameters faster than I'm sure many of the much lighter builds here can do. I would like to put the lead screw to the front, but the compromise would be to push the Z axis further away. This creates a lever. The ideal situation would be to use a c-beam, but then there goes serviceability. The c-beam extrusion is great. The assembly is convenient and compact, but not worth having to completely tear a machine down just to access the lead nut.

    I often wonder, where is the point of running lead-screws as long one keeps on using wheels directly on the extrusions, opposed to real linear rails and wagons
    I'm really not to keen on the v-wheel thing either, but it works for smaller machines such as this. I have hit limitations on my Sphinx from it though.

    your twin x-beams hanging free in the air between the end plates. Once your gantry moves these will flex and vibrate.
    That's an assumption. The beams should be bolted together throughout the length to create a single 4060. I know exactly what flex you are refering to. One beam goes up while the other goes down as a result of load along the y direction. This also twists the gantry.

    Z- looks like it could be much shorter - introducing a lot less flex.
    The Z extrusion is fixed. Only the spindle slides. Having a sliding Z axis halves your effective cut depth and places more weight and leverage on the already sensitive wheels. In a fixed setup, what matters (in a negative way) is excess below the gantry. Excess above the gantry will only make tool changes easier. That exact concept is one I apply regularly. The gantry sits as low as possible. The clearance between the lowest point of t he Z axis to the top of the 4040 rail that makes up the bottom frame is 75mm. Lighter materials can be reached with a 4" end mill. For more demanding materials, simply raise the bed closer to the gantry. With the 250mm length z extrusion, the spindle has a 130mm range of travel, but only 75mm of it is usable. That extra 55mm is just there or may allow for long length end mills if the need ever arises. It's unlikely, but you can remove the raised spoiler entirely and use all 5" worth of travel if cutting foam. But removing that extra 55mm will lower the hanging stepper, which may help a bit during direction changes.

    You sure to get that short belts?

    It's just a standard GT2 3mm pitch belt. SDP-SI has them all. In my case, the 210mm length (70 teeth) is offered through the store for this exact setup. The shorter is a 150mm pitch length (50 teeth). However, after looking into it deeper, the OB offset design is actually on the large end of what's needed. The center to center distance on the offset plat ranges from 69mm to 77mm. At 69mm, the pitch length is 198mm. The spacing is 76mm with the 210mm which is at the edge of adjustment range. A 204mm would give a more ideal spacing at 72mm.

    Using bigger sprockets is fare more sensible as you can transmit far bigger loads (your torque) and have a lot less stress on the bearings. ( one of my biggest criticisms on the Omega-drives like the ox`s)
    A larger sprocket will apply a gear reduction. So it will double the applied torque, but will also half the rotational speed. The side load (tension) on the bearings is unaffected by sprockets diameter. Small will experience the same side load as large. The belt and pulley system on the traditional machine (like the Ox) is a bit different. In my case, we have a full 180 degree engagement. Slip as well as stress on individual teeth wont' be a concern like it may be on the Ox.

    Sorry for sounding so negative but i am just trying to save you from frustration.
    Building a machine (not a working thing -but one for specific goals) is more than using strong parts. Momentum, flex, levers, over- vs under -engineering...

    There is no such thing as bad criticism. Levers are the number one aspect to avoid, then build it strong to avoid flex.

    The belt drive tipped this into the over-engineered side, but the reason for that was to make it compact. Not only is the machine more space friendly, but wiring will also be easier as you don't have to route around hanging steppers. Waterproofing would be easier as well if desired. With ~52mm body length stepper, about 3" of space is saved along the y axis and x direction. 52mm is on the short end for a NEMA 23. It also eliminates the protrusion.


    I really do appreciate any thoughts, experience, ideas, and feedback. That's how we grow.

    As for the specs:
    500x500 gives a 345x345 work area
    actual working depth without spoiler is 75mm, with deeper reach if milling very light materials.
    NEMA 23 127oz-in all around
    400-600W spindle
    24V
     

    Attached Files:

  6. Florian Bauereisen

    Florian Bauereisen Well-Known
    Builder

    Joined:
    Feb 24, 2014
    Messages:
    188
    Likes Received:
    60
    HI,
    maybe i have not been quite clear (sorry no native english speaker)
    Having the ball-screw to the back of a x-beam is not the best solution if we talk about highest possible precision. It is ok for any hobby cnc as long as the setup is build very solid. Usually this results in a (figuratively spoken) tube like very strong structure that will surround the x-beam. (viewed from the side)
    In your case however the back and front are conected via the bolts only that support the wheels. That is what i call flimsy in order to transfere the load . On other ox-like machines the setup may be the same in general but then the front-plate will get "driven" directly so it doesn`t matter what happens to the back side.
    Yes you get a lever turning this around but that is simply a matter of engineering.
    On my cnc for example the x beam does not sit sandwiched between the side plates, it sort of rests on a L shaped cutout. (firmly attached of course via several massive brackets...)
    Additionaly positive about this is that your sideplates do not need to be 100 % accuratly at 90 deg, as it does not change the x beam relative to the board if you get my meaning. Very easy to set up 90/90/90 deg i may ad.
    You can see it on YT, look for "diy cleapath servo cnc "

    On the sprolket...hmm i meant both gear "wheels" really ... i actually do not care if you keep ratio or go 2:1 or 1:2 - just go bigger.
    And yes bigger diameters will carry torque a lot better than smaller ones (the torque you want to transmit) ,as well, as the load on bearings will shrink. Belts carry a lot more torque on a bigger diam. than on a small one....
    There is even calculators for that on german websites if one looks for a suitable belt setup.
    www.maedler.de

    For the beams hanging unsupported ... well your machine is tiny, i realise, so it may prove to be a non issue. But no, it is no assumption.
    Buying from supplies like smp bosch etc you can get specs on their extrusions and you may realise that these will even flex under their own weight. You may be aware that so called Tube cnc build sometimes even flex upwards when the bit tries to aply pressure to the part.
    A strong and heavy base is very important to a cnc.
    Why do you think professional machines tend to weight tons?
    Again on your tiny machine it may proves to be a non issue.

    Try it and let us know. It definetly shows your thoughts that have gone into it.

    greets

    flo



    greets

    Flo
     
  7. Rick 2.0

    Rick 2.0 OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Dec 20, 2013
    Messages:
    2,855
    Likes Received:
    1,525
    The issue of flex in the bolts is easily resolved by cutting a piece of full thickness blocking that fits between the front and back plates on the carriage.
     
    Joe Santarsiero likes this.
  8. Florian Bauereisen

    Florian Bauereisen Well-Known
    Builder

    Joined:
    Feb 24, 2014
    Messages:
    188
    Likes Received:
    60
    Right you are.
    I merely comment on what is shown...
    I guess people want to do the design work for them self. Creating their own.

    greets
     
    Kevon Ritter likes this.
  9. Kevon Ritter

    Kevon Ritter Veteran
    Builder

    Joined:
    Apr 30, 2015
    Messages:
    597
    Likes Received:
    293
    Ah I understand what you mean with the bolts now.

    I do agree with the gear statement. I initially thought you were saying to change the ratio. The three easy to find sizes for gt2-3m belts are 16t, 20t, and 40t.

    At this size, the wheels give before the extrusion begins to flex. Before building a 500x500 Sphinx, I used a Shark. It was built with 20mm unsupported rails that are 32" long. It was also made from HDPE. I can take my finger and flex the entire gantry. Flex is something I am well aware of as a result of the painful experience of owning that piece of crap.

    I meant to put pictures of some other concepts I had for longer lengths, but completely forgot. We'll save that for another thread though.

    As for the heavy base, have you seen the machines made with granite? Those things are heavy!

    I don't quite remember what I did, but I did make some changes based off of your first post.

    As for connecting the two gantry plates, would using additional bolts and spacers (no wheels) be acceptable?
     
  10. Florian Bauereisen

    Florian Bauereisen Well-Known
    Builder

    Joined:
    Feb 24, 2014
    Messages:
    188
    Likes Received:
    60
    Hi,
    sorry about the language barrier. Are we talking about the Z-plates now? the ones riding left to right?
    Then no- have you ever tried to climb a very old and used ladder? How it sways left to right while climbing? "loose"/beaten corner connects do not transfer these kind of motion they do not prevent parallel twist. In order to stiffen two parallel rails/tubes/beams you need solid square parts in between for connecting . brackets, square plates come to mind. Same as your railroad rail really - a base, an upright (square) spacer and the rail it self, I beams in buildings, square tube wherever you look around ( just double I-beams).
    Try it yourself with some scrap pieces of wood. first a ladder made of "sticks" (no need to be accurate or nice..) then two beams connected via 2 or 3 "plates". you won´t get the last into a parallel.
    now transfer that principle to your z-plates viewed from top. Than i will hardly matter as to where the ball-screw will push and pull.

    greets

    flo
     
  11. Kevon Ritter

    Kevon Ritter Veteran
    Builder

    Joined:
    Apr 30, 2015
    Messages:
    597
    Likes Received:
    293
    This project is far from dead. I did make some mods from what has been mentioned. Though I can't do much for the original Mini that has been previously posted. I'll comeback to that topic a bit later.


    The way I wired my Sphinx was a bit interesting (to me at least), but there is a huge fundamental flaw. The OpenBuilds type of machines are somewhat modular due to the basic extrusions. The electronics side is purely dictated by the builder. On most builds, the wiring is kind of just one mass. When I built my Sphinx, my goal was to make it as easy to work on as possible. The entire electronic side, can be separated. BUT, it's restricted to that particular machine.

    The goal for this control box was to have a plug and play solution with any machine. For that to happen, it had to have a compact and easily portable format. So here it is.

    Size:
    • Dimensions: Width x Length x Height - 185mm x 264mm x 195mm - 7.3" x 10.4" x 7.7"
    • The idea was to keep it as compact as possible.
    Cooling:
    • Dual 80mm Fans
    • The primary objective was cooling since it's an enclosed box. The drivers are located directly in front of the exhaust fans. The primary air intake is positioned to the bottom front, with smaller front mounted side intakes. The PSU would be pulling its own air from the rear if needed, which means that there will be no dead spots. Dust shields can easily be installed on the intakes as well.
    Wired Outputs:
    • Instead of using panel mount style connectors, having a wired connector extending from the panel allows the machine to be positioned at any orientation. It also makes hooking everything up easier since you can hold both sides of the connector. You will also be able to place it closer to a flat surface, such as a wall.
    • Front USB Port (the cut out to the bottom)
    • The wire output "block" in the rear can be replaced to accommodate any wiring requirement. In my case, I would have 10 outputs in two sizes.
    Modular:
    • The controller mounting plate can hold a controller up to 80x140mm with hanging over. Wiring may start to become a slight hassle if you go wider. This will hold all of the common controllers, such as the Arduino Uno R3 and the xPro V3. I doubt if will fit the Phoenix in a flat orientation.
    • The driver mounting plate is currently designed to hold four 542 base drivers. To change this, you would only have to drill new holes.
    • Removing the four bolts for the front panel allows everything to slide out. Of course you can not take advantage of this if everything is hardwired.
    Simple:
    • It's based on flat panels, four 250mm extrusions, and bolts/nuts. Aside from the panels themselves, the most unique part is a spacer. In the pictures, 20mm spacers are used to mount the controller plate. You can raise or lower this if needed.
    • The side panels are identical. They can be used on either side.
    Ideas, Improvements, and Possible Shortcomings:
    • I'll move the on/off switch location to the top of the back panel.
    • Throw some LEDs in it. :D
    • I can either design a remote control box for E-stop and other buttons, OR design a front panel to hold them. The latter wouldn't be as pretty, but the option should be there.
    • This will not hold a large VFD. The smaller 542 style (600W) VFDs may fit. This box just isn't designed for it.
    • A Meanwell PSU is required.
    • There is a 20mm gap on top. As long as it is sealed, it could be turned into a tool holder. Or a door can be installed for documentation. It would only be a usable 140mm x 235mm space though.
     

    Attached Files:

    Giarc likes this.
  12. Kevon Ritter

    Kevon Ritter Veteran
    Builder

    Joined:
    Apr 30, 2015
    Messages:
    597
    Likes Received:
    293
    I forgot to post a pic int he above post. I don't really like editing complete posts so I just attached it here.



    After trying to come up with a way to build a 4x2, I just decided to drop it. I did all kind of math to find the deflection of various structures (box hollow and solid, rectangle hollow and solid, i-beams, etc...). But as the structure got stronger, weight goes up. Even though it would be designed for lighter cutting (no metal), t would absolutely need rails after a certain point. So I flipped it to a 2x4. The actual cutting area is 32" x 52" if I remember correctly. The largest sheet I will ever cut is a 2' x 4'. Anything larger can just go on my buddy's 4x8 or 5x12.

    With that said, the first machine design was the Mini Phox. Here is the Extended Phox, which is based on 1m and 1.5m extrusion. The Phox (not pictured) is a 1m square (2x2).

    I went back to the standard stepper mounting. It's designed for 1/2" lead screws. With an L bracket, spring loaded anti-backlash nuts can be used. And bets of all, the X axis is now attached directly to the Z axis.

    As shown, the extrusion is setup for a 2' x 4' spoiler. The goal for this one would be to include a vacuum table. This would allow much easier cutting with light materials like foam.

    Opinions? Ideas for improvement?
     

    Attached Files:

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice