I just ordered this...

[throbscottle]

https://www.banggood.com/3018-3-Axi...=145885357&sc_uid=ZuQ5tbS19B&cur_warehouse=CN

I always wanted a CNC machine. Never imagined I'd be able to afford one! We will See What Happens...

 

[Pommie]

We need a full review as I may have to buy one.

Mike.

 

[sagor1]

I have one very similar. What I found wrong with it was the "chuck" for holding the bits are only 2 set screws (on mine anyways...). One on the shaft, one holding the bit. This allowed a major wobble in the bit due to play in the diameter of the shaft/chuck/bit, making it rather useless for fine things like isolation routing on a PCB. Even with 2 opposing set screws, you still get wobble in the side direction. What I ended up doing was ordering a small adjustable "drill" chuck on Ebay, with a press fitted brass bushing over the shaft (heat the bushing and slide on, then cool). The adjustable chuck is much more stable (straight) than what they supply. These chucks are made for these hobby milling machines.
Other limitation was there are no limit switches. If you have to reset while doing some milling, you cannot necessarily find true "zero" again, without some form of limit switches (optical are best, micro switches next best). The controller has inputs for limit switches.
Your controller board is a little different than mine. On mine, there was no PWM for the spindle motor, only for the "laser" output. On mine, the PWM would only work when it was high enough to trigger a relay to turn on the spindle full power. I worry about that relay on yours, it may be the same "idea". You can easily tap into the PWM signal and feed a MOSFET driver to get true PWM speed control on the motor. BTW, my motor is not a real high speed one, and not very high torque, but enough to mill things like PCBs.
Finally, I found the gantry too low, mostly due to the extra chuck length. By the time I added a sacrificial board on the table, the limit of vertical travel (Z axis) was less than 1/2 inch. I ended up ordering new aluminum frame members (they come in 3 foot or 6 foot lengths from CNC shops), and extended the height that extra 1/2 inch+ (with spare length on top in case I mill something thick and have to raise the motor).
Check the firmware in the controller. If it uses GRBL firmware, download the latest version (1.1f?) and flash it.
https://github.com/gnea/grbl/releases

Those are my comments. You may find your model has some slight differences or features fixed (or different ones broken). Good luck and have fun with it...


EDIT:
PS: if ordering frame members, check out their 90 deg angle brackets. The grey items in your picture are plastic (including the motor mount), and that 45 deg support on the sides does not provide the best support. Using steel 90deg brackets on the outside (horizontal side to vertical member) locks it in a lot better with less flexing.

 

[throbscottle]

That's interesting to know, sagorl. Some of the negative reviews mentioned issues including the lack of limit switches (also missing/incorrect parts being supplied). I guessed the parts are standard enough to improve and adjust - so it's good to know I was right about that! I may end up getting a collet chuck, but we will see how it goes.
Thanks for the "preview"!

 

[sagor1]

Throb, you can also use a regular Adruino Uno board with a add-on daughter board (CNC Shield) with higher capacity drivers (Polulu), and with better connections for the limit switches and PWM, etc...
I found that RFI from the motor itself (noise from brushes) caused something to crash (board on machine, or laptop USB port). Make sure you use a well shielded USB cable, preferably one with ferrite beads on each end. That reduced the number of hangs/crashes I've had. Add RF bypass caps to the motor leads, at the motor, if it does not have any yet.

 

[throbscottle]

Well, it arrived today, found to be fairly easy to assemble. Couple of minor niggles, some of the screw holes in 3D printed parts weren't quite big enough, 1 was missing counter-bores for screw heads, no spacers to go between plastic panel the control boards sits on, and the frame (so control board mounting screw heads are against frame). Instructions left something to be desired.
The cable they supply for the spindle motor just goes to a 12v supply. No speed control. There's a 2 pin socket marked "12-36v" so I guess that's the speed controlled one. The PSU supplied is 24v so I doubt if it actually goes to 36v with that one.
Came with what I think is a US mains plug. Grrr.
Otherwise a very pleasant afternoon.
The coupler the supplied as a "chuck" is a very snug fit on the motor spindle, with 4 screws.
There are no limit switches. I discovered the importance of these when trying out moving things around and crashed. Arrrrgh! Anyway I see header pins for them so will fit some micro-switches. I have a load of them pulled from various things just waiting to be useful...

Can someone possibly tell me why these other headers are there? They are marked in pairs:
5v G
3.3v G
13v Ref
A0 ...
... A7
Tx Rx

Well, we'll see how long it takes me to actually make something, now!

 

[Mickster]

Isn't that header just the Nano PCB pinout?
The Nano is 15 x 2 rows, but GND & RST are both duplicated on each row, so a 14 x 2 header is sufficient to break out all the pins.

 

[throbscottle]

I thought so but know nothing of Arduinos so had to ask.

The X and Z axes run backwards and I can't find any reference to these machines doing that. I also learned it is AKA the "woodpecker". Awww...

The software Bang-good supply turns out to be an open source project that hasn't been active for 2 years. So although it's cross-platform (yay!), I'm having trouble getting the Linux binary to run properly because of some weird GTK glitch. So I need to install it the old fashioned way and hope it works, wahhh. Reason being, I want to see if the axes run normally with it. I was testing with bcnc and can't find an option to reverse an axis. All I can think of is to specify a negative dimension.

 

[Mickster]

Isn't that board running GRBL?
If so, this may be helpful to you (scroll down to the $3 section)
https://github.com/gnea/grbl/wiki/Grbl-v1.1-Configuration

I think you can also use Chilipeppr in Linux??

 

[throbscottle]

That's interesting. Obviously what the tick box sets in other software.
Been having a look at Chilipepr. Will try it connected to machine tomorrow.

 

[throbscottle]

Aha! I found the invert mask in bcnc. So if I set that, and move the Y motor connector to the other Y port so it's also reversed, they should all go the right way. And now I have to walk the doggies and go to bed

 

[Mickster]

Haven't looked at bcnc, so not sure what options are available to you for doing the configuration.
One other prog, which I failed to remember earlier, is that you can use Universal G-Code Sender too.
You can set up the GRBL firmware on your Arduino using either Universal G-Code Sender:
https://winder.github.io/ugs_website/
...Or Chilipeppr.

If Y was going the right way and X & Z were not, you would enter $2=5 (step port invert mask:00000101) into the "Command" box in UGS, or enter the same in the "Type serial port command" box in Chilipeppr.

If you are inverting Y in bcnc, aren't all the axes now going in reverse?

My confession:
It's better to get each axis moving in the correct direction before you start making any chips/prints/laser cuts/etc, and also to confirm you are moving the correct axis too.
I built a Banggood Prusa i3 clone 3D printer and symmetrical parts turned out great. It was only when an asymmetrical part was printed, that I found out the part ended up mirrored. Cue a bit of head-scratching and I eventually found out that what I thought was the X and Y, were actually the Y and X.... Swapped the stepper motor cables and all was well.

 

[throbscottle]

Lol. At least I got my cables right! There are 2 Y outputs, which turn out to be Y and !Y. I couldn't install UGS, couldn't be bothered doing more with chilipepr for now. After you pointed it out to me I found you can set the $ settings in bCNC directly, which I've now done and the axes are now correct. bCNC has got a handy control pad to you can move the head/table around.
Now cutting the "Transformer" test image in a bit of plastic. Weird how it pulses the spindle motor. Some kind of substitute for PWM I wonder?

 

[throbscottle]

Huh. Test files are getting mangled somehow, not coming out right. Tried both absolute and incremental setting, just got 2 versions of bad. Did a gear though! Auto generated shape, just sent it as is, came out perfect.
I obviously have a great deal to learn....
Altogether very happy with this little machine anyway, it's going to keep me occupied for quite a while...