I think this might be the secret sauce the Nomad always needed to be a real CNC monster. Ditch the wimpy Nema 17 BLCD and put a REAL motor on it. I had a spike of 697 Watts measured at the outlet. Mind you that the power supply here only powers the spindle. 16,000 RPM measured with a laser tach.
Next upgrades are a 10mm spindle shaft and better motor driver. I think with a new set of pulleys and a belt, a c10-er11-150l collet chuck holder, 10.2mm ID x .8mm thick Bellevue washers, and some 62900zz bearings will be a drop in replacement for the 8mm spindle shaft. Hopefully it will allow for larger endmills to be used with less stickout too. These parts I believe will install into an unmodified stock spindle barrel, making it a drop in replacement.
I’m needing to replace the BLDC driver as well. Its currently got a BLCD-8015C driver that is interfaced directly to the original Carbide3d motion board through the original spindle driver socket. I have also added an external power supply dedicated to the spindle that has an adjustable voltage. Right now I’m limited to 36v due to the motor driver. I will be putting a either an Odrive or a VESC in its place, probably the VESC 6.7. This will require an Arduino/esp32 to convert the PWM duty cycle spindle speed signal to a PMM signal for the VESC to have accurate speed control.
The motor is a Flipsky 5065 sensored outrunner. It has an 8mm shaft and I used the stock Nomad to mill out an adapter from 6mm aluminum plate. It currently has a 40t pulley on the motor and the stock 20t pulley on the spindle with a 150-2GT belt. The current setup is great if anyone chooses to use it. Overall, just with the motor upgrade the machine has been transformed. It does small intricate cuts better because of the higher RPM, and it has the power to really hog out material like no other desktop machine I’ve seen. It does require proper tool engagement and some adjustment of speeds and feeds to get a good cut, but once that stuff is figured out then its a solid upgrade.
On the power supply, I bought an adjustable 1000w 0-48v supply that cost $84.52. This isn’t needed, and is overkill. I think you could get away with a 36v 16.6A (600w) power supply that can be had for around $36.99.
Now, I do want to add that the BLDC-8015 driver is not working as expected, once I solve this part I’ll probably do a full writeup on how to perform the modifications with a complete parts list. I strongly discourage anyone from using that driver for now. Even though the PWM signal from the Carbide3d motion board is outputting the exact signal the motor driver expects, the driver seems to be limited by the pole count on the motor. There is a potentiometer that I messed with blindly to see what it did as a hail mary, and I did manage to get full speed, but that’s basically all it has right now. Leaving the factory GRBL parameters with $30=10000, that makes it easy to estimate the pwm duty cycle. Playing with that, I had the full 16000 rpm down to 90% duty cycle, then it suddenly dropped to 6000 rpm at 85%. 70% was something like 2000rpm. After doing some digging, its because the BLCD-8015C that I utilized is designed for a max of 6 poles, and I have a 14 pole motor. So I believe the better way to approach this is to use an Odrive. They also have an external power resistor that is used for braking that would help immensely. I believe this will get more RPM, more low speed torque, and a less noisy motor.
I did cut a part with the new motor though. If the tool loading is correct, it cuts like a champ. If its not, It has enough horsepower to chatter like never before . I could visibly watch the 5mm endmill deflect enough that I know its a spindle rigidity issue. The 10mm upgrade will likely solve this. I also found the Z axis needs a different antilash lead screw nut. I’m working on finding a suitable replacement. If I can find one, it will be a solid upgrade for all 3 axis.
The other weaknesses of the machine are showing up rapidly now but I don’t believe they’re anything that can’t be overcome. I think once the spindle rigidity issue and the Z-axis lead screw are corrected, this machine will hit well above its weight class consistently.
If spindle rigidity is an issue, will increasing the size of the cutter actually help? It seems like that would better transfer load to the spindle which would cause it to deflect more?
This is a cool project, I’ll probably tackle it myself as well at some point. So many thanks for sharing
Its more of an experiment to see where the issues of the machine start showing up first. I have the horsepower to drive it, might as well try it and see what can be improved. When the machine is then operated at a more reasonable level, it will be much less stressed and will cut better, with better finishes and more accuracy.
The stock spindle has an 8mm shaft up through the bearings. Looking at the stock nomad spindle, they used an off the shelf er11 collet extension with an 8mm shank to build the spindle. I’m just replacing that with new bearings and a 10mm shank collet extension. That extra 2mm will be roughly 50% more rigid. I get a kick out of how many common off-the-shelf parts they used in this machine. Easy stuff to source.
The part to look for is a C10-er11-150l ER extension. For the bearings, I ordered 62900zz bearings. The OD and width are identical to the 608 bearings but have a 10mm ID and are deep groove bearings. The belleville washers I mentioned in the first post are to preload the bearings the same way Carbide3d did.
Even with a much less powerful motor, I believe the spindle rigidity has always been an issue. Even with an 1/8" cutter. Any deflection anywhere is bad. So I’m hoping to solve it or at least chase it further up the chain, and might get a little more roughing speed and hopefully better finished surfaces out of it. I believe even a stock nomad would benefit from a more rigid spindle.
I might pick up an er16 version of that extension. With the beefed up spindle motor it might be worth having the extra tool size at our disposal. Even if it turns out that we lack rigidity up the chain, the larger tooling should be plenty useful for easier materials.
McFly cutter to surface hdpe or machinable wax blocks on a Nomad 3? Sounds like a good time to me
@k1n3k, out of curiosity where did you find those? All I see at Boca (my bearing “go to”) are 6900ZZ, which are 10mm ID, 22mm OD but only 6mm wide.
Having had an original 883 before my N4, I was struck by the “balanced-ness” of the Nomad. The 883 Classic was definitely spindle power limited, but also had the less-beefy round rails that were adopted in the Pro and N3. I’m interested to follow your development of your machine, given that you are starting with a good beefy chassis…
. I could visibly watch the 5mm endmill deflect enough that I know its a spindle rigidity issue. The 10mm upgrade will likely solve this. I also found the Z axis needs a different antilash lead screw nut. I’m working on finding a suitable replacement. If I can find one, it will be a solid upgrade for all 3 axis.
