After following the trails in literally every milling build here I found answers and wanted a mulligan. This is a great cnc build forum but isn't a cnc forum thank god, cnczone left a horrible taste in my mouth. Openbuilds is exponentially better but the computer still says no, Aluminum cnc diy is still $4k (same cost of having molds cut.) My background is from 3D printing and I'm intimate with the $350,000-no-thinking necessary mechanical, that is resolution=$ and no in-house machines. So, because lead times in good shops round here are still 7-9 months that gives me time to do the CAD myself and clone a version of katran's or a heavy weight extrusion version of yours. I learned welding at art school ... what did flo call it, "stress", well precision don't do well with stress, so fasteners for me.

You got it right, no offense China but the well dried up about three years ago. I'll have to bite the bullet and find a US supplier for the ball screws and bearings but and I hope I'm wrong, the US retailers are just buying up the good international production and doubling the cost. Still if the part store gets them in I'll keep the Nema 23 as the limiting factor and drop my 'tax refund' on the electronics and commit to this fantasy and post.

Ah ... Rhino just finished downloading, I can finally see katran's CAD, hope there's a bom ... I'm off to make some coffee and monkey with his design. Thanks again for morphing your ox hybrid and giving me some optimism.

jack

 

To get both the X and Y level and square I bonded on aluminium strips to the box section then machined those (easier to machine / abrade than the steel).
To get them true I used oversize holes in the box section /aluminium strips and steel 'nut-strips' on the inside of the box section to secure the linear rails. The whole process took forever (best part of two days) to complete.

Tweakie.

 

We spend so much time fretting over shipping and vat when what we need is a new kickstarer for bulk screws, bearings and plate, co-op supplier in the part store? One Datron M10 running 24/7 and everyone has to work 8hrs a month in the shipping department, scratch that it'll be a new religion, tax exempt Church of the Cut! I'm in charge of collection baskets. There is a market for designs, kits and finished CNC but the only prophets come from the supply chain and shipping and handling is the only real evil. Deerhoof says "Don't be mean to see the green." Every supplier now gives away off the shelf parts cad, machine design can be reduced or merged to the OX caculator. Of course designing, proving that it is reliable and worth the build will be the miracle: moses the AsexualOx , my first build (... too far?)

Here is the Katran design I got from grabcad https://grabcad.com/library/cnc-router-frame-2 but I had issues importing/exporting into my other cads and viewers and had to download the Rhino 90 day trial.

Onshape is another young cad, browser based or on your smartphone.
https://www.onshape.com/

Neil - Do you need ebay for extrusion? I thought Misumi and 80/20 were international direct sales, custom cut like home depot, (extruded aluminum rail going from Indiana to Florida, then to New York then to Ontario, good luck staying true.)

jack

 

Ask and you shall receive, you are closer to my thinking. My thinking is that the gantry plates can carry the load well especially with gravity supported linear rails. With your welded steel there is no way that a nema 23 would torque anything but itself. That is my current fear, now that I have a rigid enough gantry support the weight will stall the stepper. There is no nice part management in my free cads like inventor or solidworks so I can't easily calc the weight. Any idea on what your spindle and gantry assembly will be? Someone suggested to me today that my application would benefit from a 3kw spindle, but adding 25lbs isn't going to help anyone.

 

There are some pictures of the construction of the router here:- http://hobbycncart.com/forum/62-216-1

 

Thanks for posting.

I'm knee deep in costing and more importantly understanding the engineering behind the infinite variables of screws, back when ebay was cheap the cost savings made the knowing irrelevant, at these prices I need to get my wits. I keep adding info to your "discussion" that you posted last month, I have to start using the stalking app here to get a better idea of what has already been answered ... sorry, just remind me to keep diggin.

hillbilly weights: 22lbs in Al and 38lbs in steel, if my cipherin is correct? I can live with those moments or the nema 23 should be able, but physics is still just a silly superstition to me.

We seem to be hunting in the cracks between the ox and something like tweakie's or this one http://www.embeddedtronics.com/cnc_gantry.html

 

Hi sw3dp,
if we are talking (still) leadscrews:
At 5 mm pitch, on a 16mm leadscrew, a stepper only needs to put out 1,65 Nm to generate 800 N of moving force at 1000mm/min.
Assuming 25 kg gantry, a 100% Safety margin and 0,1sec accelleration as well as the use of the square type of linear rails.

800 N force should (roughly) suffice to mill aluminum at aprox 6mm endmill / 1,0- 1,3 mm deep per pass.

Hope that helps

greets

flo

 

Hi,
problem with steppers beeing their rpm/torque curve.
With the 5mm pitch leadscrew in the example above the motor only turns 200 revs per min. At that rpm one could easylie mill some aluminum with a smallish motor like a kress 1050 . But at say 1000 rpm or 5m/min it hardly puts out any torque at all.
So it is a design-aim game.
While beeing strong enough to cut aluminum, this example would still cut balsa at fast speeds and say 2mm ply to 1500mm/min...2,5 D would require some time as the roughing at high speeds and feeds only will allow 5or 6mm millbits.
The finishing cut on the other hand will be as fast as anything, as the ballnose bit being smallish as well as its stepover. The nema 23 allows for high accelleration and the milling parameters allow for high cutting speeds..

If one desires to go faster on that 2,5d example one would need to have higher torque at higher revs...leading to a longer Nema23 as mentioned above.
raising your voltage suppy will give you higher torque at higher rpm as well.

If one desires to simply aply more brute force to aluminum ( 3kw spindle and say a 20mm bit) one would need more torque at low rpm leading to a Nema 34 setup. The "ocasional" 2,5D mould or plug still beeing possible but the inertia of the big diam rotor will slow the process down ... especially at the finishing cut using small bits and stepovers.

So as anything in live it is a tradeoff...

Servos beeing the ultimative solution, i have the clearpath ones on order for my new mill.

greets

flo

 

Hey flo,

Yes and thank you. At this point I'm going to pretend I understand everything you said, there's that 3kw beast again and clearpath as in $300 nema 23s?

I was playing with the idea of the 2.2kw or slightly smaller spindle as the Chinese ones in my price-point are rumored to have exaggerated specs. On the x with two 5mm pitch 16 mm x 1000mm lead screws and 20mm flat linear rails, 76mm nema 23s that supposedly pull 3.5A, allowing for "max" holding torque of 3.0Nm(425oz.in) but I would rather drop down to the 1.8A, 2.4Nm(340oz.in) for lower voltage/heat and backlash? Higher spindle speeds with 4-12mm end mills and .75 mm depth per pass (calculated solely on acceptable machine time per mold as I have no real experience with realistic feeds and speeds) for 750mm x 300mm x 15mm 6061 mold parts. 8+hrs machine time/4days per week. Running in a enclosure at low enough decibels to listen for that missing cat. That was just the initial hurdle. If the improbability drive was capable, I would have a machine for the cost of the molds ... and still better education than community college working on a 5 axis machine.

Those were the economic and arbitrary assumptions I had going into the minimum specs for the build design after throwing the Chinese fortune sticks. Like I said too little experience and way too many variables but I will build something and find out the hard way for the fun.

Steve, sorry for the hijacking of your discussion, I swear to migrate these opinions and specualations to my build once the extrusion and stepper motors arrive (I, me, mine.)

jack

 

Sorry Steve to, again answer to Sw3Dp within your build:

The Servos had only been mentioned as these offer the best solution if you are willing to pay....alot...

As i wrote before it is the best to use a as high as possible voltage supply, in order to get the most out of your steppers ( or servos)

There is no gain at all to run them at lower voltage. Quite contrary, having some overhead will prevent missing stepps and stalls.
I merely tried to point out the difference of a nema 23 to nema34 drivetrain..

Backlash has nothing to do with the power output of your steppers... it is a mechanical thing...

greets

flo

 

flo,
thanks, got it, backlash is on the screw not the holding torque of the stepper motor.

Steve,
In your early iteration you had the stepper motors off set with spacers on the y gantry I assumed to gain some routing space inside? I've been thinking about that today was the change a stability thing?

jack

 

Hello friend,

It's very nice job, but we need some what the max pay load ?

 

HI Steve, and wow has your design evolved! There's probably a business somewhere in there if you can compete with China-Incorporated (who are mainly putting out pretty but low performance cheap machines). Where you at lately, I've seen so many different renderings here, which one or ones have you actually built at this point?

My hat's off to you!

Neil

 

Fist of I would like to congratulate you to the great design! I'm very new to the CNC world but the design caught my eye.

I've been trying to find the size of your machine and was not able to. How big will you create it?
At the moment my experience with CNC is at the level of building a Rostock Max V2, printing finished designs to get used to the printer and the CAD. Some small own mode designs mostly for the household or my workshop.
A CNC mill like the one you are working on would be perfect for my intention to design my own models. The mill to make negative forms which can then be used to make full carbon fiber airplanes instead of making the forms from an "accurate" hand made markup.
Next to that I love building machines like this one so I'll be checking once in a while on the status of your build. I guess the aluminium extrusion version would suite my purpose and build skills as I cant weld.

Anyway.. Thank you for sharing and looking forward to maybe build one on my own.

Phil

 

Hey man.

Your designs look amazing.
I am looking at upgrading my existing OX now and I have studied your designs in detail.

One thing I noticed with your design evolution is that you moved from to 2 SBR round rails mounted on each side of the Y Axis to the 1 Bosch rails per side mounted on the top.
This seems much better, from what i can see but how will you keep the crap from going into those slides ?

Obviously most of the rail / ball screw kits that come from china only come with one Y Axis rail, so this is a problem for me.
Also i am struggling to find any Bosch or Hiwin rail kits that A) Are the right size for me, or B) Are not to expensive.


Hope you don't mind me asking you a few questions...
I think you could save me a lot of time and money..

1. Are the Bosch bearings that much better than the SBR Rail ?
1a. Do you think your design would work if you mounted a single SBR rail on the top of the Y Axis instead of the Bosch rails?

2. What was the quality of your Chinese ball screws / SBR rails like. Did you use a particular ebay supplier?

3. How are you planning on cutting your plates and what material do you spec ?

4. What steppers and drivers have you spec'd. Will the Nema 23's still cut it do you think ?

5. What router head unit do you plan on using ?

Thanks man
Best of luck
Martin

 

PS. What made you decide to move away from rack and pinion ?

 

HI,
its a little xls. sheet. Unfortunately i do not know how to post the results..
well anyway that is just a quick calc... of course that assumes that friction and all other factors really are
what you type in and of course it assumes a real ballscrew with 80% efficiency.
If i change that to trapezoidial screw than figures will change dramatically. Your 110 N sound to me like that.
screwlenth.. motor inertia friction ....onandon....
so if you change a few factors you will see dramatically different figures.


greets

Flo