Making a CNC Router from very thin plywood is a big challenge for you. You will need to produce stiff box sections, and use a good quality glue. There are a few Plywood CNC builds on this site but using much thicker, hardwood Plywood. I guess the CNC size will be small?
Anyway good luck with your project.

 

That's good. The model that you have attached seems to be incomplete, I'm guessing that the gantry is not cantilevered?

 

The 400 watt spindle will be a great starting point for your router as long as you are satisfied with it's abilities. Limit your cutter diameter to no bigger than 1/4 inch ( or 6mm) and you will be fine. There is a ton of import 1/8" Cutters on ebay that are very cheap and sharp enough for starting out. When you decide to start cutting aluminum you will need better quality bits, but my son cut up to 1/4" aluminum with a 400w spindle motor and very cheap carbide bits. Just limit your speeds an feeds to what the setup you build can handle with good results. 400 m/minute might be a good rate on wood with one cutter but it might be a disaster when cutting aluminum.

 

You'll almost certainly get better results by relying on design rather than material properties on this one. Use thinner plywood for the moving parts, more like 10-15mm rather than 25mm-class stuff, and make sure all of it is in torsion box sections; the base, the columns, the gantry bridge, everything. The thinner wood will save a lot of non-useful weight, whilst the design will add significantly more rigidity than you would have gotten with the thicker material. You could also screw on some cheap aluminum angle on the edges to add some extra stiffness and rail mounting surfaces.

Use a high performance glue that will also add significant structural support to the whole thing whilst keeping weight down- and can be spread in very thin sections to keep joints from floating too much and won't flow out under screw tension. Not 100% sure what that would be, off-hand, (polyurethane could be a good option?) but there are likely more options in Germany than a lot of other places anyway for those high performance materials.

At some point, put blasting sand or pool filter sand inside the torsion box sections to help with damping (plywood is fairly resonant for a composite wood product- look at DIY audio applications and musical instrument bodies!). If your axis motors can handle the weight of it, put some in the gantry and carriage too. Adding the MDF for the spoilboard will help there, it's quite damping too.

Consider your options for not using wood for the gantry carriage/z-axis body; it may be too small a volume to practically use wood. To create a stiff enough structure, you may end up creating too large an overhang, creating a significant moment on your gantry axis mechanicals. It's such a small area that you could probably just use laminated aluminum square tube or something along those lines and get a similar rigidity to plywood in 1/3 the volume, or something. On the other hand, maybe a simple slab of 24mm plywood would be sufficient to do everything because it's so short. Maybe you could use 5mm plywood with the right structure and keep it all shrunk down flat. There are a few possible options, and you can't really know unless you a) try them, or b) run some calculations. Just something that jumps out to me to be aware of though.

That spindle actually looks pretty decent for what you're doing. 500W, if that's close to actual usable power out and not just marketing, is plenty for basically any material that a plywood machine is gonna cut. Brushless is almost certainly going to be higher power density, likely smoother running and quieter, more usable torque, all that good stuff. I'd keep poking around a little more to see if you can find a 240V EU model, but it may just be the 110V US models that can be found in the lower ~$100-150 range. For cutting wood, you'll definitely appreciate the larger collet options.

Gearing down your steppers between 2:1 and 4:1 will improve both absolute torque and acceleration (at the cost of top speed rapids) for smoother interpolation and contouring, so maybe consider how feasible it would be to do pulley plates and motor standoffs instead of direct mounting if you're intending on doing a lot of arcs, circles and organic forms vs simple plate-making.

 

With 2mm / rev you certainly don't want to step down. The typical openbuild screws are 8mm/rev.

When picking a spindle you'll want something that goes low in the RPMs. Especially if you can't get to run faster than 400mm/min. Ideally, when running 10k RPM you'll want to go about 3000mm/min to get a decent chip. At 10k RPM and 400mm/min you'll be rubbing the bit and generating a lot of toasty dust and burning your carbide...

What drivers are you using? If you add a fan to them you can get pretty far with DRV8825 drivers. I ran my machine at 4000mm/min with DRV8825 drivers. With 1/8 micro stepping you should be able to hit roughly 2000mm/min before the GRBL/Arduino is maxed out.

So, add a fan, max your currents and see how high you can push things...

Maybe I missed it, but do you have a plan for tackling backlash?

(Btw, in the description you still state that it's 2.4mm plywood, not cm).

 

60C should be within operating range. I think that steppers are ok even at 90-100C.

 

First of all I am with you on every cost saving step you have taken and applaud your efforts to build what you need while finding alternative materials that might have ended with filling up more landfills. You are proving that spending loads of money is NOT required to build a CNC. Great Job!!

One place I'd urge you to re-examine is your stepper mount. The photo on the Build page shows a gap between your mount and the face of the stepper motor flange. I think that you need to either put a spacer between the motor flange and the mount, cut open the mounts inner bearing space open to clear clear it to give you a flat mount or use washers to fill the gap on each bolt you use. My thoughs are that mounting like you show puts lots of stress one the stepper motor flange that may cause it to part off. The biggest issue is that I doubt those bolts can ever be tightened enough as you show them.

Just some thoughts you did'nt ask for.

Larry

 

I made my first cnc this year, used 18mm ply throughout and it’s a pretty solid build. I’ve looked into using x-frame rails but the cost for some pieces are extortionate and to be honest don’t look rigid enough for what I need.

I’ve used both my Mikita trimming router but have now opted for a cheap Chinese spindle to cut down on nose and vibration, and it worked well. Although people have mentioned sbr20/25 rails, I opted for sbr12, I believe unless you are mating a monster of a machine, anything larger that this would be overkill.



These are just some cuts I’ve managed to achieve with my machine for some Christmas lanterns I’m making and with the setup I’ve used I think they turned out really well considering I’ve only spent £150 to make the machine

 

Some roughness as you'd expect, but that's super impressive, actually, for 150 quid.

I think people wildly underestimate how much force a large number of ball bearings spread over multiple planes of a fully supported rail can take. People are out here using 20 and 25mm THK/Hiwin rails when 12mm and sometimes even 9mm rail would be just fine for their machine. It's hard to judge visually when things are engineered like that, common sense has to take a back seat to actual numbers.

(Also, YORKSHER! )

Yes and no. I'd use full or half-stepping on 2mm pitch screws- that's 0.01mm or 0.005mm resolution, respectively- not that you'll actually see that out of the machine, it's just the theoretical output on the nuts themselves. This will also reduce the frequency that your controller has to send at, which is usually a good thing for reliability. As you say, it'll also give you more torque.

However, gearing achieves more than purely adding torque, it also balances the power transfer between the moving carriage and the motor rotor- you don't want them to get too far away from each other in terms of momentum, or you end up forcing the motor to do a lot of work that you don't actually see out of the machine. Remember, every time you decelerate the carriage, you're basically trying to force its kinetic energy back into the motor... Which the motor doesn't like very much. Gearing helps even the two parts out so it can speed up and slow down much more quickly and efficiently (which it's doing a lot any time you're cutting curves!). On your machine, it might be overkill, but just thought I'd mention it anyway, since inch-thick plywood could be a lot for the motors you're using.

 

I think any unreliability is more a product of the variability in packaging and distribution than anything; I've had zero issues with the "Beauty Star" importer versions on Amazon, for instance. And at $11 a piece, you could afford two dead ones out of every three before you start getting into DM542 territory. If they're sufficient for the application (and there isn't a NEMA 17 project they couldn't handle, along with a fair few low/mid-grade NEMA 23 and even some small NEMA 34 projects) the cost efficiency is hard to beat. As with any project, if you need guaranteed functionality the first time out the gate, it's worth paying the extra for, but frequently with electronic projects comprehensive QA and warranties are an expense that may not be worth it.

There's also the fact that "TB6600" is kind of a genericised term, and what driver IC is actually in there is open to interpretation. I haven't been aware of any where they were severely down-rated to the point that they effectively weren't TB6600-equivalents, though.

I drive my laser NEMA17s (typical mid-range ones, 50-70oz-in) at 1.2A, I believe 15000mm/s^2 acceleration and 8000-10000mm/min top speed. I always use 8x microstepping because it's the best trade-off between torque and precision.

Cheaper drivers (and even expensive ones in some circumstances, physics being what it is) may have issues regaining position after a disabling cycle with microstepping, so if you're testing with stops make sure your enable setting is on "always on", it's like $4=256 or something, I forget now.

Stepper drivers also want specific shapes and timings on their input pulse trains. If, for example, the step pulse duration is near the minimum value, it may sometimes be missed by the driver. Or if there isn't the specified gap between step pulses, or the direction pulse doesn't switch early enough before the next step pulse, etc etc. This shouldn't typically be an issue for Arduino-based controls, but it can come up occasionally.

It would be very odd for all three/four/whatever drivers to be bad. If all axes are skipping steps, that sounds like a control issue or even a structural issue. If it's just one, make sure it's definitely skipped steps and not lash. This is a wooden machine, remember! Try driving different axes with different drivers, too, see if the problem follows the driver or if it stays on the axis. All the usual troubleshooting steps.

Double-check that the problem persists with a desktop, too. Likely not an issue with buffered serial commands, but just in case. Laptop I/O can get spotty at precisely the wrong times and often for imperceptibly short durations. Saves battery, not so great for machinery.

I dunno. Hard to gauge what it might be from just this description, but hopefully it can be fixed without buying a bunch of expensive drivers.

 

Wow, I can't even measure 5 microns. If you can achieve +-0.1mm on your machine you will be doing fine.