DIY CNC laser engraver

[blueskull]

----------Background----------

This thread is about a partially (mechanical, PCB, firmware, not PC software) open source DIY CNC laser engraver that's built completely from ground up.
The idea came from my failed purchase, a $129 "1500mW" laser engraver from Amazon, which turned up being a POS with inflated power rating, unintentional laser turning on and not very usable software.
I started another thread about pimping it up, but I then removed the thread due to the original laser engraver being destroyed.
There's no way I will buy another one jut for pimping it up, so I decided to make another from scratch.

----------Objectives----------

The goal of this project is to design and build an actually usable mini laser engraver that can mark on all common dark colored plastics, including high Tg/Td molding compound, Bakelite and glass fiber reinforced plastics.
Being able to mark on dark stainless steel and/or anodized aluminum is a bonus, but not a hard requirement.
Due to the application, Z axis space doesn't need to be large. As long as it can accommodate common IC/module packages, it is okay to me.
This machine is designed to take as little space as possible while being able to mark a relatively large area for marking chips in array for increased throughput.
Also, this machine shall have built-in air filter and smoke extraction system for indoor use.
Finally, inherent safety. This machine uses short, adjustable focal length to prevent reflected light from damaging operator. A safety door is provided to isolate internal cavity and external space.
Device under engraving is loaded in a jig, held in place by 4 corner magnets.
Another reason to use an adjustable focal length is to allow hands-free, indirect tuning of dangerous invisible laser beam using a sacrificial engraving stock.

----------Design target----------

Overall size: 150mm*150mm*100mm
Working area: 100mm*100mm*30mm
Minimum dot size: <=10um
Stepper resolution: <=10um
Stepper repeatability: <=20um
Maximum stepper speed: >=50mm/s
Maximum optical output power: >=5W
BOM cost: <=$300

----------Mechanical design----------

X axis: gantry, driven by stepper motor+lead screw, stabilized by dual 4mm steel rod
Y axis: plotter, hung on X gantry, driven by stepper motor+lead screw, stabilized by single 4mm steel rod
XY lead screws: T8/2mm metric ACME
Third axis: hung on Y plotter, micro stepper driving gear set for focal length adjustment
XY axis resolution: 1.8deg/pulse, 4 microstepping, 2mm per rev, resolution=2mm/200ppr/4=2.5um
Third axis resolution: 18 deg/pulse, 10:1 gear reduction, 4 microstepping, 0.5mm per rev, resolution=0.5mm/20ppr/10/4=0.625um
XY axis anti backlash design: horizontal compression nuts
Third axis anti backlash design: spring loaded linear actuator
Homing: micro switch for all axes
I chose to use XY plotter design rather than X gantry+Y moving bed design for maximized space utilization.
Since the required free space is moving platform size + moving range, by using a smaller moving platform I can have more moving range for a given space constraint.
With only 50mm margin on X and Y axes, minimizing platform size is crucial, and the use of a large 100mm moving bed simply won't work.



----------Optical design----------

Lens: single uncoated aspheric G2 lens with 6.33mm diameter and 4.02mm effective focal length
Calculated minimum spot size: assuming projected spot size on lens from fiber is 1mm, w0=2*915nm/3.14*4.02mm/1mm=2.34um (fundamental mode)
The optical subsystem is designed to achieve as high efficiency as possible with adjustable focal length to accommodate for different device under engraving height.
Adjustable focal length also allows for adjustable spot size, which is important for obtaining smooth edges while keeping internal features sharp.
Coating for 915nm is non standard, so I chose not to use a coated lens. 4% reflection is a loss that I can live with.

----------Mechanical/optical component selection----------

XY stepper: short body NEMA17 with 1.8deg stepping angle
Third axis stepper: 8mm micro stepper with 18 deg stepping angle (I would like better, but that's what the bay has)
Lead screw: T8/2mm metric ACME
Lead screw coupling: CNC part for D shaft to screw coupling and screw to opposite bearing coupling
Lead screw bearing: 10mm/22mm 6900Z sealed bearing
Anti backlash nuts: pair of CNC part with M2 screw for compression
Supporting rod: 4mm hardened steel rod
Rod bearing: LM4UU linear bearing with CNC part for bearing to platform connection
All metal panels: 3mm 7075 profiled and routed from 1/8'' stock
Third axis gear set: CNC parts from nylon, non lubricated
Laser source: 8W 915nm 105/125 fiber laser module (again, I would like better, but that's the only I have in stock)
Laser collimator: uncoated G2 lens permanently attached to Y moving platform
Laser fiber attachment: sliding mechanism machined from C360 brass

----------Electronic component selection----------

Controller: PSoC5LP with built-in USB and DAC
Laser driver: LMZ31710+OPA365+CSD16342Q5A (buck pre-regulator+IDAC+PWM, PWM for sub-threshold power operation)
Stepper driver: TMC2208 for all 3 axes, internal RDSon Isense, UART mode

----------CNC part list----------

Frames: top, bottom, left, right and back panels, top upper cover and top loading door
Reinforcement bars: 2 aluminum bars on each of left/right sides for holding linear bearing rods, one aluminum bar on the left side to hold X axis stepper, one aluminum bar on right side to hold lead screw gear
Platforms: one X axis platform with 4 vertical reinforcement bars and Y motor holder, one Y axis platform with lens holder machined on it, one magnet-embedded bed for material under engraving
Lead screw adapters: D shaft to flat cylinder adapter and cylinder to lead screw adapter
Lead screw nuts: 2 X axis nuts with compression screw threads, 2 Y axis nuts with compression screw threads
Linear bearing adapters: 6 cylinder to rectangular cube adapters for 2 X linear bearings and 1 Y linear bearing, 2 per bearing
Optical fiber adapter: SMA905 adapter with threads on each side, one for SMA905 connector, one for elevator platform
Fiber adapter elevator: internally threaded to house fiber adapter, externally machined with hear teeth to interface with gear set
Optical fiber cantilever: a holder to keep fiber adapter in place without tilting while allowing elevation
Gear plate: a thin machined part to house gear pins, bolted on Y platform

Optical fiber adapter and optical fiber adapter elevator are milled and threaded from 3/8'' C360 brass stock, optical fiber adapter cantilever and gears are milled from 1/32'' nylon sheet and all other parts are milled, routed threaded from 1/8'' 7075 stock.

----------To be continued----------

[moffy]

Very cool! I have bought a 7w blue laser engraver, build is good quality, and have made my own controller board which allows me to control the laser power. Uses the latest version of GRBL but with some extra code and hardware interlocks for controlling the 12v laser supply. I also had a 3mm thick orange perspex box made to enclose the laser which blocks the blue light, making it safe.
Look forward to hearing more!

[CM800]

Hi Blueskull,

How much experiance do you have with mechanical motion control design?

20 micron accuracy is -very- hard to achieve without proper linear rails, ground ballscrew etc.

I would be shying away from Steppers and looking towards servos, along with linear encoders.

RLS do some very nice magnetic encoders that are accurate to 20 microns over a meter, with sub micron resolution.

don't forget 200mm of Aluminium will expand 24um over a 5 degree temperature change.

If you have any motion questions, feel free to give me a poke, It's what I do for a job, anything I don't know I can ask one of my colleagues for you.

I've actually been thinking of building a unit myself for some time, slowly gathering parts
(including some linear motors, and a 200W 532nm Q-Switched medical laser) Waiting on some drives and a large granite surface plate before making a move (plus a bigger workshop!  )

Buying a linear stage second hand wouldn't be a bad idea. Check out your local surplus shop or Ebay:

https://www.ebay.co.uk/itm/IKO-Mikroskoptisch-Positioniertisch-microscope-linear-stage-Kreuztisch-X-Y-Tisch/352279001528?hash=item520576f9b8:g:DI4AAOSwvVNaHDjt

[CM800]

With ground circular rod rails, you won't even notice the thermal issues as those things are so sloppy.

You will -really- want to use square linear rails.

Hiwin is a reasonably well known (lower cost) linear rail manufacturer.

If you have a look over their datasheet:

https://www.hiwin.com/pdf/linear_guideways.pdf

Page 80 will be most relevant to you. It's the size of rails that would most fit.

They give you some really good numbers to use, along with mounting information etc.

I would completely stay away from ground circular rails if you're looking to do anything with some form of accuracy.

If you are to approach a proper linear rail distributor, I would suggest going for a light pre-load or very light pre-load.


If possible try to make your design in such a way you don't need two sets of rails on one axis, making them parallel nicely is a royal pain in the arse. ESPECIALLY if you don't have the good equipment like runout dials and gage blocks.

I've built a few stages in my time. The cheapest and best way is usually to buy pre-built, eBay is great if it's a home project. You can usually find something that works.