Updated OLED menu (mlmCnc28.zip) to be able to test gcode/job before running it. (Must be running GRBL 1.1 on Arduinio)

Updated OLED menu (mlmCnc23.zip) to support grbl1.1. Added grbl1.1 zip with modified (grbl-1.1f.zip) step and direction pins. I was able to upload GRBL from PI using https://github.com/gnea/grbl/wiki/Compiling-Grbl .

Updated OLED menu (mlmcnc20.zip). Now you can connect to GRBL and jog/zero axis, home, and run a job. Any .nc/.ngc/.tap file in Documents show up in the jobs list. Run manually first to test new features and debug. Still not fully tested be very careful.

Updated OLED menu (mlmcnc13.zip). Now you can view PWM values if you have an attiny85 connected to the serial port. Also now displays RPM values if you have attiny85 connected to BCM pin 6. I have added code (connect ESC to BCM pin 12) to run the spindle manually or from PWM signal but I have not fully tested code, be VERY careful.

Yea i was also worried about that as well. Where there is a coupler there is a chance for the connection to be off center. For me the small motor i was using was soo crooked that i had no way of knowing the affects of the using a long shaft like i did. The ER11 chuck you listed bellow is the one I'm using now (look at the pictures at the end of the parts). The new large motor listed in the parts section with the chuck you have below is really nice and there is very little runout (did not measure). the chuck i got ended up fitting perfectly in the 5mm shaft of the motor. I don't have any milling experience... maybe zero, since i still haven't milled anything yet :)

I was looking around some more online for motors.

Have you considered something like this:

http://www.aliexpress.com/item/spindle-motor-clamp-rod-C16-ER11-35L-5mm-shaft-motor-lengthened-clamping-knife-engraving-machine-drill/32626384239.html

It is a collet that directly attaches to a motor shaft.

That would simply things a lot...

With your setups I would be afraid that small angular or displacement errors would result in large errors at the tip... Especially with the long shaft errors would amplify... I have the feeling that a regular coupling is also bound to introduce axial displacement errors because you tighten the nut to one side with no way to center the mill bit...

Do you have experience with mills or do you have knowledge of the above issue?

Interesting thoughts...

I'm looking forward to your findings...

Yea a 4mm lead might be a good choice. As for the motor thinking about it a bit more the biggest reason I changed the motor was that it vibrated way too much, I think the shaft was bent. I don’t know if it came like that or I bent it while trying to put the coupler on but I know I wanted to upgrade the motor anyway. So I just looked at 5mm shaft motors so I wouldn’t have that problem again and for easier assembly. The small motor would definitely etch the PCB fine, but I might affect your feed rate and cut depth settings. When it comes to etching both big and small motors will probably perform the same, but I think the bigger motor will have the edge with it comes to drilling and cutting the PCB out. Also the small motor seemed to run hotter so you might have to add a fan but I never ran it longer than 15 min. if I were to stick to a small motor I would probably redesign the Z carriage to hold the motor in a way that I would just use one of those 3mm couplers they use on those diy kits to reduce some friction:

https://www.amazon.com/3mm-3mm-Coupling-Coupler-Transmission-Connector/dp/B01D89CICC/ref=sr_1_7

I most likely will not be testing the small motor as the one I got is bad and I don’t see myself buying another one any time soon. I would say it’s worth getting a motor with a larger shaft, if only for a smoother rotation!

4mm lead might be a good middle way then...

I had not expected that the spindle would be a problem... It should take very little force to scratch off the top copper layer from copper clad pcb board I would think... Bldc might not be the best choice for a spindle I realise now. They are high speed low torque motors. I could consider mounting the cable driven extension piece to the carriage instead....

I'll wait for your updates before ordering parts then ;-)

Hi Sjoerd,

yup I didn't measure the lead screws and linear rods just put them next to the CNC and marked them. But looking at the numbers your right, the lead screw on the X can be as long as 200mm and 180mm on the Y. I don’t believe the coupler adds any length (if you don't want it too), but the motors I used do protrude 7mm the same depth as the bearing opposite to it. As for the linear rods I believe you can get away with 210mm on the X and 190mm on the Y. You will have to cut a bit off of the Y axis rods and lead screws but you can use the 200mm rods and lead screw for the X if you use the extension piece I added and you can get 100mm rods and lead screws for the Z axis.

About the motors don't really know how much more accurate they are to 1.8 or 3 degree motors in real world application, especially with microstepping. Initially I wanted to use the motors in full step mode to get the most torque out of them, but the vibration noise was just way to loud. What I ended up doing in full step mode was reducing the Vref value of the stepper driver. I found that I could actually get more torque with less noise by microstepping at a quarter step (you can see my GRBL settings) and increasing the Vref value. I didn’t do any torque measurements just used my finger to apply force (not too scientific).

Yeathe lead screw I got has an 8mm lead. When I chose the lead screw I thought the pitch was the lead :) and was confused why my numbers weren't working when setting up GRBL. Looking back at it now don't know with microstepping If I would want a 2mm lead since travel is pretty slow as it is now. But yea at 3 degrees a 2mm lead does sounds like a good idea.

This might be surprising but I still have not cut a single thing yet. All I have done is check and tune features/hardware as I add them. I found the small motor very limiting it seemed to weak for the ER11 chuck, that's why I just added a bigger motor. The Z axis does have some small wobble because it only has one set of bearings. Don’t know yet if the motors torque are a limiting factor, I have plotted some stuff but I haven't tried to do anything crazy yet. The last feature I want to add is an RPM sensor to monitor the motors RPM as it starts to cut in to material, I just ordered some stuff, so that will probably be the first PCB I mill. I will add some pictures when I do that but with aliexpress shipping it might be next year :p

Thanks for the elaborate yet quick reply!

That is a lot of useful info... You are right that I probably wouldn't need the extra 2cm in the Y direction...

I made an educated guess that the leadscrew in your design would be cut to approx. 200mm. I assume from your reply that that is incorrect?

I have a hard time guessing the length that is added by the coupling. I think I forgot to account for the fact that the coupling doesn't go all the way to the face of the stepper so that would mean your lead screws are actually less than 200mm? From your reply it seems that they are longer?

Do you know how long your leadscrews and linear rods are? That would be helpful info...

I realise now that the linear rods are a fixed amount longer than the lead screws so the probability that I can get away without cutting rods is low since I probably won't find the parts with the right length difference online...

I don't fancy a kit because I have many of the required parts already here (steppers, bearings, most of the electronics).


And you did most of the development work already anyway... ;-) For which I would like to thank you.

I agree that your motors are probably better. I measured mine and they have 3 degree steps :s

I can partially compensate that by choosing leadscrews with a smaller lead distance. Yours are 8mm lead but they go down to 2mm for the same price on aliexpress so that would give me a 4x precision gain (probably at the cost of speed).

This makes me wonder: do you use microstepping and at what divisor?

And: what is your general experience with the resulting quality?

What are the difficulties with this machine? which parts are limiting performance most?

Hi Sjoerd,

Yea I agree blender can definitely be tedious, but to be honest I probably spent 40% of my time on useless visual details, 40% on what parts I should go with, 10% revisiting everything on that 80% and changing my mind, and about 10% on the actual CAD/mesh work.

I went with the motors I chose because they seemed like the best combination of size/weight/torque/accuracy(.9 degree). One of my biggest concerns was the motor size and weight and that 3d printed material could not handle the wight of a nema 17 motor(without using up more space for extra support). I would have liked to use a nema 17 motor since they are more powerful and even cheaper. I pretty much built this machine around these motors :)

As for the rods and lead screws I wanted to go with a pre-cut sizes but I couldn’t find the right sizes for the print area. To cut the lead screws and shafts I just used a Dremel tool.

Basically the way I went about building the CNC was:

Q. What is the max size I can get out of the UM2?

A. Created cubes and opened it in Cura until it was gray. Didn't know at the time about messing with the machine settings.

Q. I know I want to use it primarily for PCBs, what is the biggest PCB I can make?

A. Worked out if I used 10mm walls for the CNC depending on the size of the X carriage I could at most get 150mm of travel in a single axis. Since PCBs generally come in rectangles I only cared about maximizing one of the X/Y axis's. As it is right now theoretically you can get 140mmx120mm (X/Y) of travel. Since the max X travel on the UM2 was 220mm I could either go down to 200mm on the CNC build and use pre-cut shafts and lead screws and sacrifice milling area or use the max printing size and go with larger shafts and lead screws that I could then cut down. I chose to go with the biggest shafts I could find that I could cut down. With 500mm shafts and lead screws I could cut them down to what I needed for the x,y, and z and use the max print size and it also saved some money instead buying separate pre-cut ones.


Q. How cheap can I make it without compromising quality.

A. This is where a lot of back and fourth happened. It reminded me of building a computer where you don't want to buy a super powerful processor with a crappy graphics card. So basically I chose mid range parts like ACME lead screws instead of low end hardware store screws or high end ball screw. Or linear ball bearings instead of those lower end brass or plastic bearings or high end linear rail guides and carriage.

If your interested in using 200mm leaner rods for the x axis, I added an x axis shaft extension piece. As for the lead screw brass coupler, the regular one seems to work ok with little to no backlash when installed, but I did see from your link that there are some with anti backlash ones which look cool. Also seems like there are a lot of options for controllers, but yea I Just went with a nano GRBL controller because it was cheap and small. I have a blank electronics mount so you can add your own standoffs :)
If I had to do it all over again I probably would probably just go with an aliexpress/banggood diy cnc kits like

http://www.banggood.com/Triaxial-Desktop-DIY-CNC-Micro-Engraving-Machine-Assembling-Kits-p-1055377.html

since it was just way too much work to create my own. When I started I only know about those 3020/3040 kits that started like at 500/600 and up which seemed too pricey for what I wanted to use it for. I set out to just create a quick boxy CNC but … yea.

Hi Saul,

I set out remodelling things in freecad because there are some other things that I would like to change too. I have used blender in the past but it can be rather tedious for functional parts...

I have not yet settled on the steppers to be used because mine are actually a bit too large for the current model:s (i.e. they stick out ebove the edge of the base blocking the platform:s)

I would also like to change the dimensions of the machine such that I can used pre-cut rods and leadscrews like these: http://www.aliexpress.com/store/group/T8-trapezoidal-screw/834897_507763658.html

I was thinking 200mm leadscrews for both X and Y.

How did you cut things to size? Do you have special tools for this?

Is there a reason that you did not utilize all available Y-length of the um2?

I would also replace brass couplers with helical aluminium couplers because they are easier to obtain...

Any ideas on this?

(I would also change the controllers to an arduino mega+ramps because I have experience with that setup but that won't influence the 3d design too much)

As a side note:

Cura 2.1 is indeed beta software and may or may not behave as advertised. At some point they introduced the "feature" that the printable area is reduced by the brimsize... So if the model stays gray you can sometimes fix that by changing the bed adhesion...

Hi Sjoerd I tried Cura 2.1 (only using Linux) but it didn't even load STL files for me. Seems like the problem with scaling is with the Cura 2? Since Cura 1 and other programs load the STL files correctly. As for print area it seems to be a bit tricky since it looks like every few releases Cura changes its machine preferences defaults a little. The previous Cura release I used the defaults for x/y were 225mm I believe and the printer head size setting where such that I only had to change the y axis to 226mm. The current Cura I'm on (15.04.6) the defaults are 223mm and with the new default printer head size setting I have to increase the x/y to 230mm so the largest parts show as printable. I don't know what the best settings are as I didn't write down all the previous settings so I would try playing with the numbers and checking that your first layer is laid down correctly (as I know that the um2 has the available print area).

Sadly I'm using Blender it's far from the best tool but its the only 3d software I know how to use. I cleaned up the blender file for the top piece of the CNC(added to document section as top.blend). Check it out and if you are interested in the rest I can work on cleaning up the rest of the machine, its currently a mess. If you load up the file you might have to press the keyboard key “a” twice to select all the pieces and then ctrl+a and select scale for all the modifiers to take affect.

If you haven’t used Blender I can try adding the right mounting holes if you know the spacing or have an SVG of mounting holes. Don't know when I can get to it so if you have time I don't mind trying :)

Saul, I would like to change the type of steppers because I have a few of a different model lying around. Would you mind sharing the sources of this so that I can change it? (assuming that you used some cad software and not a mesh editing tool to design this...)

In the new 2.1 branch of cura they also show up as 100x too small. scaling to 10000% results in the sizes from your screenshot. However, the model exceeds the printable are in the x-axis.

Hi Børre thanks, I appreciate it. I haven’t used Simplify3d, only Cura. I created the model in mm so I don't know if there is a mm/inch issue? I just downloaded a few files to see if there was an issue after uploading but they show up fine in Cura. If its showing the objects to be scaled to 1 on all axis then you might have to scale the models up or maybe try re-installing Simplify3d? I know for the Ultimaker 2 I had to increase the print area “depth” (y-axis?) by 1mm (226mm total) for Cura to be able to load some objects. Don’t know what I can do on my side, but would like to know if you get it figured out :) (or if some one else passing by has any suggestions)

for example 1:1 scale of the base (base.stl) should be 220x200.843x75mm (pic uploaded)

Really good work

Downloaded this, and the parts is really small when I import in Simplify3d.

Is it possible to get the parts in right scale?

Love it. Awesome machine.