MPCNC Part Advice -

[en97]

Hello!

 

I’ve been researching and testing for some time now and finally think I’ve found a way to make what thanks to MPCNC! MPCNC achieved a 3 axis CNC for <USD $500 with < 1mm precision milling aluminum (Cutting Aluminum on a Mostly Printed CNC (MPCNC) - YouTube)!! AMAZING!!

 

My goal: 5-axis CNC / laser for cutting / welding aluminum + computer vision feedback + 3D printer for under $1000. My initial research is documented here: https://www.cnczone.com/forums/diy-cnc-router-table-machines/364310-cnc-cad.html

 

I’m going to start off with a 3-axis CNC with 2’ x 2’ x 1.5’

 

Frame

I got 8′ 1″ solid steel round from the scrap yard for about $5.

From the thread https://www.v1engineering.com/forum/topic/concrete-filled-tubing/, it seems it may be too heavy / too much inertia for the gantry. Will it be too heavy for the outer frame? I don’t think so, it doesnt appear to sag at 2’ - I’ll have to measure.

Parts of it are rusty, ideas on how to clean it up will help. I’m thinking some sort of drill “lathe” + sand paper… but I’m still trying to figure out how to hold the opposite end.

The local maker space has a lathe I can use so I can cut it into 24" pieces and mill it uniform there.

Alternatively, I was thinking of getting a self centering metal lathe chuck https://www.ebay.com/itm/Metal-3-Jaw-Lathe-Chuck-M12-1-45mm-Self-Centering-Hardened-2pcs-Lock-Rods/192395274680 (~USD $20) and milling it myself

For the gantry I'll use :

3/4" SCH 40 (1.05 OD X .113 wall) A-500 ERW Structural Steel Pipe which is USD $13 for 4’.

1 OD x .065 wall A513 HREW Round Steel Tube USD $10 for 6’

If I’m milling the frame, I’ll mill this tube to the same OD of 1" / 25.4mm or maybe take everything to 25mm

 

Motors

I’m planning on using NEMA 17 59Ncm (84oz.in) 2A 42x48mm Model # 17HS19-2004S1 , a 5-pack is ~USD $40.

 

Motor drivers

I’ve been reading a lot of good things about the Trinamics drivers. The Quiet feature is nice because it reduces noise and my gut says this will also smooth motion resulting in less jitter, improving cutting and printing precision. I don’t see any current boards with the 2.0+ A Trinamics drivers so I’m still searching at the moment. The closest board I found is this: Duet 2 Maestro Electronic Board – M3D (USD $120) but only 1.6A per motor

 

3D Printed Parts

The local makerspace offers an ABS printed parts service using a Stratasys Fortus 250MC. Getting a quote or will be happy to support and buy from V1 Engineering!

 

CNC Spindle

The official recommendation is the Dewalt DWP611 (~ USD $150) but its manual control which is the biggest con for me. I need to digitally control the speed to improve performance with Vision feedback. I’m also curious about water cooled spindle and this 2.2kW Water cooled variable speed spindle is the same price 2.2KW 220V Water Cooled Spindle Motor ER20 80MM CNC Router Milling Grinding US | eBay

I’d also like to be able to mill 0.4mm fins out of a copper block for DIY water cooling blocks using https://smile.amazon.com/gp/product/B00QCL3RVA/ or similar . I suspect this will require a smaller framed machine with smaller gears for high accuracy. Another benefit of MPCNC is that its easily adaptable!

 

Laser

Using the theory from this: Making the cut | MIT News | Massachusetts Institute of Technology and US6327292B1 - External cavity laser source using spectral beam combining in two dimensions - Google Patents ; I think its possible to use CD/DVD lasers to make a low cost high power diode laser. I’m going to order 10 diodes (~USD $30) and see if it works as a 5W laser. I’ll use a 2.5W laser as a control.

 

Ultrasonic Sensor

see this thread https://www.v1engineering.com/forum/topic/ultrasonic-sensors/

 

Computer Vision

Use camera(s) to get feedback and improve control. This is mostly software trickery.

 

Thats all I can think of for now. More to come!

[en97]

I was also wondering: Why the design choice to use 50+ bearings in a triangular pattern vs. linear bearings?

1" linear bearings are about the same price for 10. Is it availability or cost at the time?

Would linear bearings provide smoother motion and better accuracy+precision?

[vicious1 (Ryan)]

@en97

Would linear bearings provide smoother motion and better accuracy+precision?

Linear bearings need to be matched with an equally high quality rail…at this size it is extremely expensive, and extremely hard to cut or drill, and leads to all sorts of other design compromises, and is not easily sources worldwide. The goal of this project is not the best CNC money can buy, it is the best machine we can easily source worldwide so everyone can have one that wants one.

You have a lot of questions in there. You might want to look through the FAQ’s and see if you can cut that down a bit. We would all love to see you make such a crazy machine but you are kind of asking us to make all your choices for you.

A 1.5’ Z axis is almost totally unusable.

[en97]

@vicious1

We would all love to see you make such a crazy machine but you are kind of asking us to make all your choices for you.

My intention was to document my thought process / build log and get feedback from people who more experienced in machining and building machines. I gotta a brain and I’d like to exercise it

@vicious1

en97 wrote:

Would linear bearings provide smoother motion and better accuracy+precision?

Linear bearings need to be matched with an equally high quality rail….at this size it is extremely expensive, and extremely hard to cut or drill, and leads to all sorts of other design compromises, and is not easily sources worldwide. The goal of this project is not the best CNC money can buy, it is the best machine we can easily source worldwide so everyone can have one that wants one.

I completely agree. The MPCNC design was released a few years ago. Now in 2019, 3D printers / CNC are much more popular and parts readily available. I can’t imagine 5 stepper motors shipped for $40 even a year ago!

This field is still new to me but the supply chain ecosystem looks more bountiful. I imagine more complex parts like the linear bearings are within tolerances but I won’t know for sure until I get them in hand.

 

@vicious1

You have a lot of questions in there. You might want to look through the FAQ’s and see if you can cut that down a bit.

Thanks for pointing to the FAQ - I missed the post on material deflection: https://www.v1engineering.com/forum/topic/stainless-steel-quick-and-dirty-flex-test/

Also found good material engineering books with the info I was looking for and have something to compare real world vs theoretical rigidity and deflection !

@vicious1

A 1.5′ Z axis is almost totally unusable.

Regarding the Z depth, having an extended cutting arm will multiply the forces on it. To stabilize it will be complex. 3-axis designs choose to make the cutting arm as rigid as possible and move the part/platform. This also simplifies the design. Still trying to figure out a way to go from 3 to 5 axis – portable designs also share this separation. I’m thinking of keeping the X-Y and cutting arm on the gantry and moving the Z-axis to the legs or a platform with 4 independent ball screws.

 

Due to the number of motors, at least 8, this will require a simpler control method than an all in one board. Maybe an Arduino + off the shelf driver for each motor and then a brain to tie it all together, perhaps the brain can be the control PC or RPi or Beaglebone.

[en97]

Here are the 5-th axis designs I’ve seen:

[attachment file="_6008134.jpg"]

V2-10 Machine & Accessories

[attachment file=86018]
[attachment file=86019]

CNC Router Machine Rotary Table 4th &amp; 5th Rotational Axis with Chuck &amp;...

 

These make sense and alignment can be simplified with computer vision + alignment sensor. See https://www.machinedesign.com/sensors/proximity-sensors-compared-inductive-capacitive-photoelectric-and-ultrasonic for different types. I’m leaning towards ultrasonic