Space constraints are real with the Nomad, it’s not possible to flip the pulleys but you can fit a 1.25" OD max pulley on the spindle. It will need machining to have clearance.
Definitely a double O ring instead of the belt.
2 Likes
The belt speed limits are apparently dictated by the smallest diameter gear’s RPM in that chart. That motor and spindle running at 20 kRPM would have about the right amount of torque for a stock Nomad with a 1/4" endmill. Gates 3M HTD belts could do that based on the chart on page 41. If the spindle can handle that speed and both the spindle and motor can handle the required belt tension, that could double the MRR of stock Nomads without increasing forces. 50 kRPM would increase it by another factor of 2.5 - too bad the bearings in the motor aren’t well suited for direct use as a spindle!
That’s a swap to new pulleys as well as the belt then?
Did you check what static tension was going to be needed for those belts at that belt speed? Don’t want another wrecked motor from overtension situation like on the Shapeoko belts.
Nope, nor do I have any idea of how much speed or radial force the spindle is capable of, or if there is a way to get to 50 kRPM - which would be much better.
I hadn’t heard about that.
Hopefully in the motor specs there’ll be a max radial and axial load rating.
As for the spindle, the radial loads should be fine as it’s designed for cutting, if it can’t handle the belt static tension then the other end will be a problem. As for speed, as you say, we don’t know what it’s balanced to or what the bearings are good for.
There’s been a few posts from folks who’ve suffered broken stepper shafts after belt tightening, when I measured the actual belt tensions and then started checking the specs of similar stepper motors I came across warnings from manufactuers of fatigue failure causing shafts to snap if the radial load was exceeded peristently. We don’t want to do the same thing to the BLDC here.
Looks like some overshoot in the speed controller. Charts assume that Kv = 190 and force would be for milling with a 1/4" endmill (double it for 1/8").
1 Like
Going back to the “replace the motor” discussion briefly, does it make sense to replace the motor given that we don’t know what the spindle can and can’t take?
Replacing the controller is relatively straightforward and inexpensive, given that a VESC costs <$100 but those Maxon BLDC motors are getting dangerously close to the price of even a Mechatron spindle, disregarding the cheap stuff one can buy from AliExpress. It seems to me that rather than replacing the motor, I might be better off replacing the spindle and motor if I’m going to go down that path.
I could actually fit that spindle inside the Nomad pretty cleanly but I’d have to replace the Z carriage…
So I have the ODrive control basically working. Only problem is I cannot stop the spindle. If I issue an M5 in Carbide Motion the spindle keeps right on spinning. After quite a bit of banging my head into the desk I realized something. The PWM output on the Carbide Board is not going to 0. In fact an M5 command does not change anything. If I run:
M3 S2000
The PWM pulse width goes from 0 to 24. If I then run:
M5
The PWM pulse width stays at 24.
IMO replacing both the spindle and motor with a relatively compact, light weight, and inexpensive modified COTS BLDC router makes a lot more sense for Nomads and maybe even Shapeokos, if acceptable performance with them can be demonstrated.
The Carbide Motion board should bring it down to 0 IIRC. I had a problem with my other controller where if it wasn’t 0, the McGillicutty would keep spinning, sometimes in pulses if the voltage was low enough.
How are you determining that the pulse width isn’t changing? An oscilloscope, or something on the ODrive?
The carbide board is run into an arduino that is measuring the pulse width. That arduino is sending commands to the ODrive and reading values from the ODrive. It is also spitting out the values on a UART which is how I am gathering the data for the pretty graphs. Eventually I want to make a CNC.js widget that will take the data and spit it out in realtime. Right now though it is giving me fairly accurate data on the pulse width.
So CM -> Arduino ADC -> Processing -> ODrive and you’re looking at the Arduino’s ADC reading?
How are you powering everything? A problem I’ve run into in the past is that when using different power supplies, the grounds can be off by a little bit. In this case you can connect the GND from the CM board to GND on the Arduino.
I am not using an ADC. I am using interrupts on the arduino to measure the pulse width directly. I am not measuring the resulting voltage. The CM board and the Arduino are electrically isolated through one of these:
Well something’s up because the CM definitely shouldn’t output PWM when M5 is issued:
M3 S2000:
M3 S5000:
M5:
2 Likes
I might be totally off here but is it possible that your code is something like:
int pulse_width;
last_time = micros();
void interrupt() {
pulse_width = micros() - last_time;
}
void loop() {
Serial.print(pulse_width);
}
And since you’re not getting any pulses, the interrupt isn’t being triggered and you’re not updating the pulse width either?
That is possible, but I believe I already handled that situation. I’ll have to check when I get home.
Nice scope! Which Siglent is it?
I bought it as an SDS1104X-E but I hacked the firmware so now it’s essentially a SDS1204X-E.
1 Like
I have seen that hack a bunch.
1 Like