Opinions on the Pocket NC 5-axis CNC mill?

[626Pilot]

[youtube]http://www.youtube.com/watch?v=UQutUdOg3gI[/youtube]

It costs $4,000 and is currently in the post-Kickstarter phase. (They pulled in $355,000 during their campaign.) It can take straight G-code, but was developed with and uses Autodesk Fusion360 for the toolchain - you get a year's subscription to that with your purchase. A BeagleBone Black + steppers are used for motion control - not sure whether they use encoders, but I'm guessing "probably not." The work envelope is 4.5 x 4.8 x 3.3 inches. The spindle runs at up to 10,000 RPM, and has an advertised runout of 0.000078".

I'm looking for feedback from people who have experience working with CNC mills, in terms of build quality, choice of drivetrain components, backlash, accuracy, and how well it's likely to stand up to machining forces. (The deflection numbers look "okay" to me.)

I have a few projects I'd like to use this on, but mostly I'd be renting time on the machine to others in order to defray the cost.

Specs:

X Axis
Max Speed: 40 ipm (inches per minute)
Resolution: 0.000125in
Backlash: 0.0005 in at tool base
Deflection at 5 pound load : 0.001 in at tool base
Max Travel: 4.5 in
Homing Repeatability: +/-0.0005 in
Repeatability: +/-0.002 in at 0% load

Y Axis
Max Speed: 40 ipm (inches per minute)
Resolution: 0.000125 in
Backlash : 0.0005 in at tool base
Deflection at 8 pound load : 0.001 in at tool base
Max Travel: 4.8 in
Homing Repeatability: +/-0.0005 in
Repeatability: +/-0.002 in at 0% load

Z Axis
Max Speed: 40 ipm (inches per minute)
Resolution: 0.000125in
Backlash at 18 pound load : 0.0005 in
Deflection at 18 pound load : 0.001 in
Max Travel: 3.3 in
Homing Repeatability: +/-0.0005 in
Repeatability: +/-0.002 in at 0% load

A Axis
Max Speed: 20 degrees/second
Resolution: 0.025 degrees
Backlash: 0.05 degrees
Deflection at 2 pound load : 0.001 in at table center
Max Travel: 100 degrees (90 degrees +/- 5 degrees)
Homing Repeatability: +/-0.05 degrees
Repeatability: +/-0.05 degrees at 0% load

B Axis
Max Speed: 40 degrees/second
Resolution: 0.025 degrees
Backlash: 0.05 degrees
Deflection at 2 pound load : 0.001 in at table center
Max Travel: continuous rotation
Homing Repeatability: +/-0.05 degrees
Repeatability: +/-0.05 degrees at 0% load

Spindle
Spindle Speed: 65-10,000 rpm
Power Output: 104 Watts
Spindle Motor: BLDC 3 Phase with Hall Feedback
Spindle Runout: 0.002in or better
Tool Change: Lever Type (no wrenches needed)
Tooling: 1/8in shank (collet for 2mm, 3mm, or 4mm can be purchased)

Machine Control
BeagleBone Black running Machinekit/
Features
Spindle Override
Feed Override
Stop, Start, Pause
Jog while pause
5 axis simultaneous movement
Ethernet, USB, mini USB, and HDMI ports
Accepts Standard Gcode
Power Source: 90-264 VAC, 127-370 VDC, 47-63Hz

Components
7 NEMA 17 Motors
3 linear lead screws with preloaded nuts
4 rotary 8/1 belt drives
Linear Bearings, 9 & 42mm, 10% preload
Angular contact rotary bearings
6061 Aluminum Frame, +/-0.002in squareness
Machine Dimensions:
Fully extended: 17.5 in wide(444.5 mm) x 12.5 in (317.5 mm) tall x 12.5 in (279.4 mm) deep
Fully compressed: 13 in wide (330.2 mm) x 9.5 in (214.3 mm) tall x 9 in (228.6 mm) deep

Software
Autodesk Fusion360, post processor included

[Windshadow]

I can't speak to this company (I have to say their pitch is very slick and a full 5 axis machine at $4,000 sounds amazing)... but do keep in mind your tooling costs as we'll, the solid carbide cutters this seems to use they have for sale at prices from $25 to $60 per cutter and these are thin and long mostly and if not used properly and with care they break early and are expensive to sharpen if you send them out to a shop to sharpen... or you have the expense and learning curve of your own tool and cutter grinder to make or even just resharpen the cutters on your own. make sure the folks you are rating time on the machine are using their own tooling and cutters or ar paying you of them as well as for the time on the machine.

[Captain Starfish]

My impression watching it as it's developed are:

It's a cute toy. And expensive for what you get.

Very small work envelope, gutless spindle, slow as hell (40ipm will be for rapids, and it's pretty slow for them), gutless motors (NEMA17).
Leadscrews? FFS - those figures for accuracy etc are pretty optimistic new, let alone a couple hundred hours of working time down the road. And they're only at 0 load. As soon as you start feeding it into a job they're going to go to crap.

Lots of functionality with 5 axis etc but not a lot of utility when you move from cutesy demonstrations to actually using it to make stuff.

If you like making chess pieces out of balsa wood, it'd be awesome. Anything useful? Nah.

[Xenocrates]

Not to mention that 6061 is the wrong material for a frame. Cast iron would be far better for a base (Granite epoxy or actual granite even more so), and a steel frame is almost essential. You should not be using a tool made from things on the soft side of what you might want to cut on it. They also don't mention if they hardened the aluminum at all. Other issues are things like using steppers for the rotary axis. Unless they have encoders on there, that's a bad plan (Especially with them being belt driven 17's. I assume they used a timing belt like GT2 (The belt the Max uses), rather than a power transmission belt (which would be far less accurate). That means it is very possible for it to slip, grind, or simply overwhelm the stepper. Take a look at the load ratings for A and B. That's two pounds, at the center of the table, and it's already giving measurable deviation. If I were to put a decent vice on it, it would probably fall off (I work primarily with Kurt vice's weighing 30+ pounds, and those are small ones.).

Put simply, it's worth about as much as a CNC mill as a Cartesian printer with a few lazy susan's strapped on. That's honestly what it reminds me of. Someone looked at their Taz, and went "I could put this on kit starter with a vacuum motor (Actually less power of a spindle than many vacuums), and some of those turntables I got at the garage sale", and then they turned it into an actual product.

[Captain Starfish]

Jeez, 626pilot: not feeling the love in here!

Damn it's tempting. ITS GOT FIVE AXES FOR LESS THAN $30,000!!!!

And they've obviously worked hard to get a small machine cheap. But seriously, if I were in that price range and that teensy size were acceptable, I'd be looking at a Taig CNC setup for 3 or 4 axes and forget the 5th for as long as you can.

[JFettig]

My impression watching it as it's developed are:

It's a cute toy. And expensive for what you get.

Very small work envelope, gutless spindle, slow as hell (40ipm will be for rapids, and it's pretty slow for them), gutless motors (NEMA17).
Leadscrews? FFS - those figures for accuracy etc are pretty optimistic new, let alone a couple hundred hours of working time down the road. And they're only at 0 load. As soon as you start feeding it into a job they're going to go to crap.

Lots of functionality with 5 axis etc but not a lot of utility when you move from cutesy demonstrations to actually using it to make stuff.

If you like making chess pieces out of balsa wood, it'd be awesome. Anything useful? Nah.

I have nothing more to add

[bot]

Safe up and buy a 440

[DeltaCon]

I'm looking for feedback from people who have experience working with CNC mills, in terms of build quality, choice of drivetrain components, backlash, accuracy, and how well it's likely to stand up to machining forces.

Hmm, I guess you got what you asked for...

[Eaglezsoar]

Expensive toy that even a five year old would become frustrated at because it would not cut anything but fingers.

[Xenocrates]

Not quite that bad eagle. If you wanted something that could cut balsa/foam/basswood/wax/plastic, it would work (As the material imprecision's would be larger than the machine induced ones, and they would be light enough). But it's kind of expensive for those materials.

[Eaglezsoar]

Not quite that bad eagle. If you wanted something that could cut balsa/foam/basswood/wax/plastic, it would work (As the material imprecision's would be larger than the machine induced ones, and they would be light enough). But it's kind of expensive for those materials.

Agreed, it would cut those materials. I will try not to be so hard on these manufacturers.

[nebbian]

Would it be able to produce injection moulds for something like normal Lego pieces, out of aluminium?

I'm actually looking for a CNC machine that would do this.

[Eaglezsoar]

Would it be able to produce injection moulds for something like normal Lego pieces, out of aluminium?

I'm actually looking for a CNC machine that would do this.

I think the general consensus is that it would have a tough time doing aluminum.
Look at the Xenocrates threads for what he says.

[Generic Default]

I have a Taig CNC mill running off of a smoothieboard with steppers, and a 4 axis rotary table on it. Small machines like these work great with steppers; I'd actually rather have steppers than servos to avoid brush wear and or burnouts.


The rigidity of this PocketNC machine isn't that great because the cutting forces are far away from the linear carriages, plus the carriages are tiny. The rotary axes look pretty good though, 0.025 degree resolution with even less backlash is excellent for small things like this.

The spindle says it has about 100 watts for 1/8th inch endmills, which is more than enough power. My Taig has ~180 watts and can take a decent cut with 3/8 endmills in aluminum. They say that the spindle on the PocketNC has about 2 microns of runout, which is on par with full sized industrial mills and lathes. Low runout like that is important when using small endmills.

1/8th carbide endmills aren't expensive, you can get them for about 10 dollars each. HSS or cobalt endmills in this size are only a few dollars each.



Overall, I'd say this 5 axis mill is decent for the money. For someone who makes small, complex metal parts, it would be very useful. That tiny engine block they machine is a good example.

I still won't be buying one though, but I would consider it if I wasn't building a bigger 9 axis machine.

[626Pilot]

I still won't be buying one though, but I would consider it if I wasn't building a bigger 9 axis machine.

Nine

I'm getting the impression that this device could be alright for prototyping, but as far as aluminum is concerned, not for routine production work. They have milled what looks to be a decent (if tiny) V8 engine block with it, but perhaps that was with a new machine, and one with a thousand hours on it might not be quite so precise or have as nice a finish. They are telling people in their Google group that they want to test it on steel and titanium, but that they don't expect good results.

Based on comments here, I see the Pocket NC as something a mechanical engineer might use to create a few iterations of a part in-house before sending the G-code to a firm that has a "real" 5-axis.

It's a bit like buying a car in its first model year. It looks shiny now, but the production models have been "in the wild" for far less than a year.

[cope413]

I heard much of the same criticism above directed towards the Othermill - to which I reply - it appears to do exactly what it purports to do, and seems adequately designed and built for said applications.

I've gotten great results from the Othermill. It has a puny, toothless belt-driven(!!!) outrunner brushless motor, with an HDPE frame, stepper motors for XY and Z movement, and a TinyG board, and it's cut 6061 beautifully.

If you're expecting to get $30k+ machine results from a $4k machine, then you'll be disappointed.

If the build envelope is adequate for your needs, and you don't mind slow feed rates (or if you're not planning to cut aluminum), then it probably makes a lot of sense - and if that's the case, most of the criticism here is irrelevant.