Building a CNC router (desktop)

[jacob]

I finally feel confident that I can build a CNC router + I have something to use it for.

My aim is to make this particular one: https://www.instructables.com/Homebuilt ... ased-GRBL/ ... using wood for all the components including the 3D printed ones. I want to hook it up to a dedicated Raspberry pi.

In the spirit of much increased synergy from making the same project, I invite people who are interested in making the same router to join me. It would be most helpful to be competent in at least one of the many skill components required. For example, given my background in mechanical clocks, I'm certain I can make all the wood+metal structures. OTOH, I know nothing about CNC programming.

To reiterate, the group strategy is to make the same widget. This will make it a lot easier to help each other. I would, for example, be able to make wood brackets for those who don't have access/skill to a scroll saw or a 3D printer, but this will only work if we're all making the same thing.

To start out, I hooked up the stepper motor from the Arduino starter kit (thanks again @Anesau) using the standard example here:
https://www.makerguides.com/28byj-48-st ... -tutorial/

It turns left and right in a very slow and convincing way. This is already halfway to building an Etch A Sketch with a sharpee. All it would require is a second motor ($1) + controller ($0.75) and a 2D gantry, alternatively hooking it directly up to the toy. I just might ...

The CNC router is only a difference of details and size.

I've been looking at starter kits, such as this one: https://www.amazon.com/dp/B072N4FMRN/

This seems to be one of those things, where lots of companies package a bunch of components together and sell them as a kit with some parts missing. For example, this one comes with 3 limit switches, but don't you need 6: One at each end of each axis?

It would be best if we all got the same kit. I've noticed that motors and motor controllers are slightly different.

Also TBD is:
1) What size wiring to use?
2) How to make a wiring harness that doesn't tie itself into a knot once things start moving?
3) What router to get? The router needs a relay to start/stop it from the Arduino ... or are they just always on?
4) The 3 nema17s need an external power supply (in the pic, the power supply is the black shield jammed into the protoboard). How big?
5) Where to get the metal rods?

[Sclass]

For example, this one comes with 3 limit switches, but don't you need 6: One at each end of each axis?

Also TBD is:
1) What size wiring to use?
2) How to make a wiring harness that doesn't tie itself into a knot once things start moving?
3) What router to get? The router needs a relay to start/stop it from the Arduino ... or are they just always on?
4) The 3 nema17s need an external power supply (in the pic, the power supply is the black shield jammed into the protoboard). How big?
5) Where to get the metal rods?

Three limit switches are for finding zero. You probably don’t need limits in both directions because steppers will slip if you slam into the stops. The little motors on these rigs won’t stretch your lead screws.

Wiring depends on how much current you’re going to use. I have a pile of steppers in my junk box that all seem to use 20-22 AWG on their harnesses. The motors are about the size of a tennis ball.

Power…you should see what the other hobby CNC people are using. Not sure what stepper and spindle motor you’re using but my Creality Ender 3 uses a 24V 15A supply for the entire rig with three small steppers + heating. It’s a small $30 unit. Switchmode used in all kinds of products. For my mill I ended up buying a power supply made to drive robotics. It was a surplus strip job and it was cheap. It was made by these guys. DIN rail mount specifically for automation systems. You can get these used for cheap at tech recyclers.

https://catalog.weidmueller.com/catalog ... 0457755045

I’d get the metal rods on aliexpress.

Sounds fun.

ETA - all this stuff has already been sorted out by the diy community. It’s all about Arduino, Marlin and RAMPs for the small machines. Best to replicate what they have settled on as it’ll be the cheapest and most readily available stuff.

I used Gecko Drive amps and they had recommendations for power and motors on their site at the time.

ETAx2 - ahh, had some lunch and recharged. I realized the system you use likely uses the three limit switches to zero out on one side (home) and then in SW you limit the number of pulses you put in from that home position based on the length of your bed. So you will never hit the stops unless you slip the motors. RAMPs boards have X min and X max inputs if you want to have wired limits.

[jacob]

In the spirit of much increased synergy from making the same project, I invite people who are interested in making the same router to join me.

[...]

To reiterate, the group strategy is to make the same widget. This will make it a lot easier to help each other. I would, for example, be able to make wood brackets for those who don't have access/skill to a scroll saw or a 3D printer, but this will only work if we're all making the same thing.

To be clear(er) the plan is to make one CNC router each, but make the same one model with the same parts. Other synergies include sharing extras after "buying too many parts". Most of the time buying five instead of one only doubles the price, but then there are four to spare. It may also be cheaper to buy one big spool of wiring. Also, some may lack the tools+skills to make one particular part, which then can be made by others---it's usually far faster to repeat something than make it the first time.

This will involve some mailing so be prepared for that.

[Sclass]

Got it Jacob. I would love to join up but I have a machine already. No more shop space either.

Sounds like a fun opportunity.

ETA - for harnessing your looks I’d get a set of drag chains off aliexpress. Cable chains they’re called sometimes.





Here is my poor man’s version. Just zip tie your cables up and run a service loop (slack) right at the point they’ll move most. On my machine only the X axis harness moves given the mechanical layout. Check out the service loop on the X and how I secure it to the base. The spindle motor moves up and down so the AC power has a big service loop right next to the motor.

The wire is automotive grade stranded wire. The signaling wire is commercial stuff (RJ11) used communications devices for manual flexing. Bottom line is you want a lot of stranding. The cheap stuff online uses too few strands and it will fatigue after many bend cycles. Use the stuff from cars or phone handsets (I guess these don’t have cords anymore) that can take a lot of bend cycles. There are kits online for DIY harnesses for custom motor cycles and cars. They’re just a zip lock bag with several lengths/colors of automotive grade wires. Other way is go to the junkyard and strip a harness out of a minivan (they’re long). I like 1990s Honda Odyssey harnesses because they still used a lot of stranding and they didn’t go to biodegradable insulation yet. New cars have all gone serial and don’t have much wire.

I put cable chains on my 3D printer. It’s a big improvement. The nice thing is you can print out all the chain you want for a few cents of filament. Otherwise I’d just buy the drag chains on aliexpress.

I browsed around over there. They have a ton of CNC stuff. Some of they’re kits have motors, amps, precision lead screws and linear bearings/rails included. If I were building again I would seriously think about using these parts.

The wood stuff in your design looks too compliant and he motors look underpowered. I may be wrong but I think you will run up against their limitations quite quickly once you start making things. I think the kit parts on aliexpress are worth a look if you are really interested in building something that is able to make parts.

Your project’s big strength is its made from readily available lead screws and wood and it’ll get you learning and having fun today.

[jacob]

... on aliexpress.

I browsed around over there. They have a ton of CNC stuff. Some of they’re kits have motors, amps, precision lead screws and linear bearings/rails included. If I were building again I would seriously think about using these parts.

With those prices, they're practically giving the stuff away. I noticed that it's the same as what I've been looking at on amazon except much cheaper, so we should/could probably buy from there. I have no experience with aliexpress and how it goes through customs. I don't want to get slammed by crazy fees. (I'm still traumatized by ending up paying $100 for a pair for $20 sunglasses when I was 17.)

Your project’s big strength is its made from readily available lead screws and wood and it’ll get you learning and having fun today.

Wood and "common metal" also makes it easy to take it apart again and recycle parts into other things. (A lot of my constructions contains 2nd and 3rd generation material that has been used in previous projects.) This is why I'd prefer to avoid buying a very specialized "set" to construct a specific gantry. For all I know, the nema17s might eventually end up controlling my window shades

My use-project (excuse to build the router) is wanting to engrave white acrylic spraypainted black to make translucent instrument panels for my simpit. Also "decorate" wood for boxes and signs to make random presents for people.

[jacob]

@AH, @sodatrain (if still interested), and others -

We should start putting together a shopping list based on the instructables link since materials may take several weeks to deliver.

[Laura Ingalls]

My machinist offspring says you need $150k and phase 3 power for CNC milling.
I think he might be thinking bigger than your set up?

[Sclass]

@jacob I was stunned at how cheap the aliexpress stuff was. They have everything you need. Not sure how customs works on that stuff. On circuit boards nothing under $1000 is taxed. I regularly buy $700 shipments of PCBs and electronics components with no tax.

If I were doing this again I’d buy everything on aliexpress. It’s a nerd fantasyland in there.

@laura it depends on what you want to cut and how fast you want to cut. These little machines are basically a Dremel tool on a 3D printer. If you want to cut faster you get more spindle power and torquier axis motors. The sky is the limit.

$150k? Sure. Last time I checked you could get a nice Haas setup for $500k. But that being said the surplus market is full of machines under $20K. My BIL has two enclosed 3 axis mills Fanuc and Mori Seki. Each cost him less than $10,000 At bankruptcy auctions. They are from 1999 and they use floppy disks but they’ll cut like a $500k mill all day. He runs a little shop making parts for a bunch of different small businesses. Gun parts, custom car parts and outdoor toilet parts among other things. Cost him next to nothing to start the business.

Tell your son to check out CNC on Craigslist. If you can work on a machine out of warranty there are deals to be had. Guys buying $500k machine centers go bankrupt all the time. Their loss your gain.

For a young guy like your offspring I’d say buy one of these cheap $300 machines on Aliexpress. The education in machine setup, gcode programming, and CAD will be many times the tuition. Then he’ll know what $150k can mean and when it’s an excuse for inactivity.

Oh yeah, and 3 phase power is nothing magical. My old colleagues made 3 phase in their homes to run their shops using residential power and DSP power conditioning equipment. They basically roll their own 3 phase so they could run surplus equipment nobody wanted…because it was 3 phase.

ETA this one looks nice. Ouch that had to hurt.
https://orangecounty.craigslist.org/hvo ... 09822.html

This too: ouch.
https://inlandempire.craigslist.org/hvo ... 77948.html

[ffj]

@Jacob

What are you wanting to ultimately make with this machine? You are going to be quite limited with a Dremel tool like Sclass pointed out. Unless you just want to engrave soft plastics?

[jacob]

What are you wanting to ultimately make with this machine? You are going to be quite limited with a Dremel tool like Sclass pointed out. Unless you just want to engrave soft plastics?

Primarily [carving black spray paint on white acrylic] panels like this: https://thewarthogproject.com/gallery (yes, this is someone's room)... (it's either that or a laser cutter)

Engraving patterns into wood to make signs or personalize other projects would be a bonus. I primarily consider this a learning project both in terms of building and "what not to do", but also figuring out what I actually need rather than committing to an oversized model that I'm not going to use. Also, it's easier to recycle generic parts into other constructions than optimized parts.

I'd prefer using a small handheld router rather than a Dremel. According to the interwebs, Dremels or drills won't last long if pushed sidewards too hard. This has also been my experience with my [naive] powerdrill-based lathe. Not a good idea.

[Sclass]

I actually think Jacob has it spec’d right. Wood and plastic. This will be a great setup to do small items made of compliant materials and with low part throughput.

Cost is low. Learning dividend is high. I get it.

Making the machine stiffer can be done with add on reinforcements made by the machine itself at a later time. Stronger steppers are a dime a dozen on aliexpress. Nema17 -> Nema23 adapters can be made on the machine itself at a later time. The machine will be self replicating to a certain extent.

This is actually sounding like a pretty cool project.

I had a night of sleep to process this a bit.

The skills you pick up on these little machines are directly applicable to big machines. Big $$$ machines. Take a look at this training school for CAD/CAM. Look at what they have on the table.



Skip the class. Make your own class. There’ll be lessons beyond machining here - like engineering, robotics, CAD/CAM, wire harness design, SW, embedded HW, Aliexpress supply chain management.

He and the ERE2 CNC gang are going to learn a lot from this. It’s a very big world in there. There are oversized rewards for people who take risks.

I’m really liking this attitude of build and learn.

Those little machines in the photo above were nowhere to be seen in CAD/CAM classes 20 years ago. Things are changing.

ETA - forgot to mention I watched the instructables videos of the wooden mill you are going to build. I noticed he cuts styrofoam and melamine foam in the vids. Very soft stuff. Caveat emptor.

[Gilberto de Piento]

There are products that are intended for those that need something bigger than what Jacob is talking about but not as big as the $150k that was mentioned and are also not used from Craigslist. Example: https://cnc.tormach.com/cnc_mills

[Laura Ingalls]

@sclass

Machinist kid’s most recent job he had a formerly self-employed colleague that had sold his mill to the company that owned shop they both worked in and several other businesses in the area.

Colleague bought more mill than was probably prudent. DS would like to be self-employed. He had an idea for a business but it was oversaturated. He likes metal, high tolerances, and complicated (three axis meh).

[jacob]

A suggestion for the cutting tool: https://www.harborfreight.com/power-too ... 62659.html

It's $30 and easily gettable. I have no idea if this one is any good(*) though, but I suspect it's more durable than a drill or a Dremel. OTOH, I don't think the specifics of this particular component is all that important for the build as long as it's not "too heavy". If you already have a "trim router" (<- keyword) or you can get one elsewhere for $30 or less, I don't think the particular choice matters all that much.

(*) One concern: This uses 1/4" bits. There are different sizes and bit collections are "an investment". The other bit-size is "1/2". Further research is required.

[jacob]

1/4" engraver bits: https://www.aliexpress.us/item/3256805949774024.html

[Sclass]

Looks like a great choice. A cheap speed control can be useful on plastic.

Look into 6mm to 3mm router bit adapters. These will allow you to use 1/8” carbide tools that are mass produced for the circuit board industry. You can get a boatload of carbide drills, gravers and end mills very cheap because they’re made for the circuit board industry. There everywhere for cheap as long as they’re 1/8” dia.

ETA - the photos of the warthog project are amazing.

[ba199]

I went down a similar rabbit hole two years ago and settled on another popular project - MPCNC (https://docs.v1e.com/mpcnc/intro/) with a build size of 36"x24". Cost was roughly $450 and the time sink wasn't too bad because there's a community built around it with a lot of information. It could've been as cheap as $300 if I would've stuck with electrical conduit of the budget builds but I upgraded to stainless steel tubing. I've been able to cut 1/4" aluminum and 3/4" plywood but the cutting times are very slow. This past year I also added an interchangeable lasercutter that can handle 3mm plywood.

For wiring harness I'd recommend - https://hackaday.com/2017/12/17/diy-cab ... ays-cheap/
Back then the go to routers for these small diy machines were Makita RT0701C, the dewalt equivalent, or a generic chinese router from aliexpress that I can't find. I think one of those would work better for this project than a dremel or harbor freight tool with poor QC. I found the makita for $75 on black friday.

I agree with Sclass on the 1/8" end mills.

[Sclass]

@ba199 whoa! Love that cable chain from a tape measure hack!

ETA - just an added note.

I am currently exploring CAD/CAM sw. There was a question several posts up regarding how to generate gcode.

I’m having really good results generating gcode with FreeCAD and their Path workbench. I haven’t tested the gcode on my controller (actual mill) yet so that is another hurdle.

There’s this multi dimensional space around gcode generation SW. Price, complexity, compatibility with different controllers, training can make or break it for your particular application. There is nothing like the environment that surrounds 3D printers and slicer SW. GCode for milling is a more painful process. And once you get it working something inevitably changes.

I stumbled on to FreeCAD Path after false starts with CAMBAM, Carbide Create and MasterCAM. Unlike those, FreeCAD is free and there is a lot of training on YouTube. I’ve gotten pretty far making tool toolpaths over the last day.

Just mentioning it because it will come up once you complete your machines.

ETA2 - ok I finally got FreeCAD Path working. It is a little buggy but overall it is an excellent choice for doing home CNC gcode generation. Took a day of watching youtube vids and tinkering with the sw. I made a simple part - a disk with a stepped flange on the edge. Taught me FreeCAD 3D modeling. This will be useful for 3D printing since it has STL output.

FreeCAD also has the gcode generator in the same application. Path Workbench. I strongly recommend this tool. It is free and it works surprisingly well for free stuff. Took a day to get it all working correctly but mostly because it is a powerful tool with lotsa settings. It works well with LinuxCNC and GRBL which you guys will be using on your mills. I used the GRBL output “dialect” and my mill understands it. My mill is an ancient DOS machine that is unsupported for the most part. The Arduino hw you guys are building has a huge ecosystem so this won’t be an issue.

So that’s it. Once you get your mills moving around, download FreeCAD and start watching the videos. It is hard to understand if you aren’t coming in with parametric cad experience but the videos help educate all levels. I’ll wait till tomorrow to cut the disk out and I’ll post up the results.

12hrs later…And here they are. FreeCAD is a hit. It’s a one stop SW for 3D modeling and Gcode/toolpath generation.

[jacob]

Alright, time to get this going ...

I am very interested. Wouldn't be able to start until January of February. I'd like to make it one of my Skillathon blocks. If it works out maybe those interested could pick the same time chunk?

I've kinda wanted a CNC machine for s while so the topic is interesting!

Anyone else?

I figure we're generally agreed on building something very close to this? https://www.instructables.com/Homebuilt ... ased-GRBL/

Step 1: Get the plywood.
Step 2: Get the controllers and motors.
Step 3: Get the rods and bearings.

Step 1 can be done (in the US) by going to the nearest HomeDepot or Mernards and buying sheets of 3/4" sanded plywood in 2x4 foot sheets.
Step 2 is a decision between going with amazon and aliexpress. The latter is way cheaper.
Step 3 seems to be a decision between metric and fractional.

I've already done (1). Now looking for (2).