The V0 of CNC machines - small envelope with 200x150x70mm (8x6x2.75") work area, aims to cut alu well (enough).
There's a small discord discussing this machine (and other) small CNCs.
A true small Desktop CNC machine that can be used inside an apartment in a pinch but can still cut softer metals. One that does not require a full shop due to space, mess and noise and does not break the bank. Current BOM cost is around $500 with most parts ordered on Amazon, excluding the price to cut the plywood panels (that can add another $100-200 if not diy'd). We're still working on the noise part currently.
CNC machines are a lot more dangerous than 3D printers. BUILD AT YOUR OWN RISK. DO NOT OPERATE "OPEN" OR WITHOUT POLYCARBONATE WINDOWS. TURN OFF THE MACHINE WHEN CHANGING END MILLS OR CHANGING WORK PIECES - DO NOT LEAVE IT POWERED ON WHILE HANDS ARE INSIDE THE ENCLOSURE.
The machine is past the main design phase and a prototype has been built and is about to get tested. But that does not mean everything is finalized.
35x15mm test cut on the first prototype machine (untrammed, excuse the bottom finish) - with 3D-printed ABS X- and Z-plates:
|
|
|
- Cuts aluminium well.
- Cost-efficient.
- Space-efficient.
- As quiet as possible.
- Not too messy.
This design was influenced by the Milk Cr8 (GH) CNC's use of plywood as a stiff box structure to enclose and stiffen the machine at the same time. The idea was to use the wooden structure and combine it with aluminium extrusions to simplify the build while retaining some of the improved dampening over just an alu extrusion structure.
- Cuts aluminium well =>
- Rigid (enough) => smaller size, slower feeds, single flute, smaller end mills,
- As it won’t be the stiffest machine, make it well-dampened => not fully made of alu, use steel, epoxy granite and wood to dampen some of the vibrations,
- Smaller, low-powered spindles 100-500W,
- chip-evacuation system => CPAP or chip fan,
- cooling via misting and air blast.
- Cost-efficient =>
- Standard components => MGN12/MGN9 rails, small 100-500W spindles, 20 series extrusions for some of the structure,
- smaller, lower-cost spindles,
- smaller size.
- Space-efficient =>
- gantry design => high space-efficiency, can be reasonably rigid in this smaller envelope
- Smaller size to cut smaller parts fitting target use-cases,
- Beginner-Friendly =>
- XYZ probe + tool setter + surface meshing,
- community presets and forums,
- custom interface for Duet RRF (much later),
- source + G-Code + Instructions packages like Coast Runner
- Integrated forums,
- g-code visualizer and simulated pre-runner including stock dimensions.
- As quiet as possible =>
- fully enclosed and ideally tightly sealed,
- interior/exterior sound insulation (?) => layer of foam + outer layer of ACM panels?,
- no compressor,
- BLDC spindles with reduced vibrations.
- Not too messy =>
- fully enclosed (!),
- dust extraction => CPAP or roborock fan,
- misting but no flood cooling.
- Makers and 3D printer enthusiasts but really anyone without a full (metal-working) shop and/or space for a larger machine.
- Anyone without the space for one of the larger machines.
The current wooden frame seems decently stiff at roughly 1N/µm in Y at the spindle collet even with 3D printed X/Y plates - needs more testing though but seems like a good start.
Some general numbers I've found for comparison:
- Siecni 12x30 Longmill MK2 ~ 0.07N/µm X and 0.05N/µm Y,
- MPCNC ~ 0.1-0.2N/µm,
- burly but large CNC build ~ 1-3N/µm,
- I've seen 20N/µm recommended for cutting alu very well,
- Milling Machines supposedly ~ 20-30N/µm,
- professional cast-iron or epoxy granite VMCs are on the other 70-150N/µm.
- (LET ME KNOW IF YOU HAVE OTHER/MORE DATA OR CORRECTIONS, I'D LOVE TO ADD IT)
NB: Load cell tool design to apply a measured/specific force to the spindle. NB: Some numbers and suggested values for stiffness - 25kgf ~> <0.1mm deflection seems like a decent goal (2.5N/µm).
Rough design for phase-2 buildout using a lasercut steel structure to increase stiffness if needed.
Finished a small prototype chip-vacuum based on a Roborock/Xiaomi/MiJia vacuum fan and Roborock-CPAP PCB. Additional cost is only ~$15 for the 2kPa fan versions or ~$30 for the (untested) 4-5kPa version.
V2 of the chip vac in a tubular form has been designed and tested, seems to work a tad better than V1 and less prone to clogging, easier to clean.
Supports different "duct attachments" held in place via magnets, which makes it easy to
- remove when changing end mills or cleaning parts of the machine,
- changing the type of attachment (there's currently an open and a closed version with a "curtain", one for better suction, one for better visibility).
Current V2:
Original V1:
- Small aluminium parts:
- high-speed 3D printer motion system reinforcements and brackets,
- higher-temp 3D printer component,
- toolhead components.
- Makers:
- Tools, …
- 🏗 low-profile modular vise system (untested design in V0.2),
- internal camera,
- laser Module Support,
- 🏗 internal LED lighting + toolhead lighting (basic design with room for improvement in V0.2),
- 4th axis?
- V0.1 - Initial Release.
- V0.2 - 07/2024
- Many, many minor improvements after alpha-testing prototype,
- roboroc-based chip-vac,
- interla left/right side LEDs,
- electronics mounts,
- support for 300W BLDC spindle (36-48V),
- prototype of MGN9-jaw-based low profile vise,
- prototype toolsetter (through-bed with fixed switch mounted to bottom panel.