We definitely have seen people skipping steps with the default firmware, 12V and 1 start leadscrews. It is hard to tell if that is a fluke, or common among people using 1starts. But dropping the max speed, or going to 24V has solved all of them, I think.
It would be easier if I could actually get my hands on a build that is skipping steps so I could know for sure why.
Maybe we need a calibration routine gcode that proves out a build. Seeing that 80+ is possible, 10-15mm/s should be easy.
12V with 1 start leadscrews will slow jobs with a lot of up/down, to the point I feel itâs worth the upgrade (PCB drilling for example). My rapids slowed down well over 50% with 1 start, but I never actually tested the max speed with 4 start (was just 35mm/sec). It might also be fun for doing big rapids on a full size LR3.
I think testing should consider step rate. Velocity is the product of mechanical reduction. So it doesnât account for changes in drive ratio due to leadscrew pitch, pulley tooth count, or whatever. mm/sec does matter wrt the kinematics of the gantry though, and is a relevant measure of cnc machine performance. But for sure, if you include step rate in the dataset, there will be a more clear conclusion as far as what different drive supply volts can do.
My general understanding is the higher voltage allows higher switching frequency (thus more RPM), due to the lower rise time across the inductive coils in a motor. So I donât think 24V would cause more motor heat until it surpasses speeds that 12V could do. Iâm not sure what happens with the driver. It would seem there would have to be more losses with more voltage to drop at lower RPM, but Iâve seen many posts that modern chopper circuits handle that efficiently where the extra heat is not an issue.
[edit: âŚnumbers for a starting point. My 1 start primo is configured for 1600 steps/mm, and 12 max mm/s. Thatâs pushing it on 12V; 15mm/s and itâs definitely skipping a lot of steps. So around 19kHz, with tmc2209 at 1.4A.]
I would disagree. The driver controls switching frequency similar regardless.
The higher voltage allows the driver to fight back EMF. Any (fixed magnet) electric motor when spun becomes an electric generator. This is the intrinsic relationship between magnetism and electricity. The motor generates a magnetic field which moves the fixed magnets. Similarly moving the magnets past a coil of wire (Or the coil past the magnet, same thing) generates a current in the wire, even if the reason for the movement is a current in the wire. Since weâre moving the coil one way, the EMF generated in this manner opposes the incoming voltage.
The drivers do some interesting stuff with the effective voltage in order to keep the current limited. It switches the power on and off via PWM to keep the average voltage to the range where the current stays at the desired level. Because the load is an inductor, this really does act like an average, since the inductor serves to prevent changes in current that are too sudden.
Normally, current = voltage over resistance (I=V/R) so for a fixed value of I (eg: 900mA) and a fixed resistance, there is only a certain average voltage that youâll ever use. Where it gets jiggy is that the motor is now acting as a generator, and trying to reverse the current flow. This ends up expressed as a voltage, too, but in the negative. In order to maintain the average current flow, the driver increases the duty cycle, thereby increasing the average voltage, to a maximum of the power supply voltage. (100% duty cycle.) The faster you spin the motor, the greater the motorâs power generated is, and the more reduction in available voltage you get.
Higher power supply voltage therefore increases the maximum RPM that you can spin the motor at, but as long as the power is still throttled by the driver, it does nothing to increase available torque.
You can get that easily from the steps per mm, if you also consider the microsteps. The Z on Ryanâs machine is 400 microsteps/mm. The xy is 100. Both with 1/16th. So that is 25 steps per mm on Z and 6.25 steps/mm on xy. Just multiply Ryanâs numbers by those and you will get step rate.
what motors are you using? Thatâs an interesting fact!!
Doh. Text fixed.
Now you know my cnc machineâs secret.
The switching frequency I interpret as overcoming the motor inductance and in principle they can both happen at high speed. What I donât know is whether the inductance is quantitatively significant relative to the back EMF.
Everything you said about the back EMF is true. But even if the motor were stalled and not generating back EMF, the inductance of the motor would limit the ability to drive it at high frequency. As frequency increases, the motor will need higher voltage to overcome the inductance to reach the specified current. Once the required voltage exceeds the supply, the driver wonât be able to drive it hard enough and the current will start to drop.
My belief is that this is a minority effect and the back EMF is the main cause of the speed limitation. I could be convinced that it is 50/50 (back EMF vs. inductance) or 90/10 or 99.9/0.1. I guess it could be calculated from the motor inductance and the âkVâ (RPM per Volt) of the motor but I would need to do some research to calculate it myself.
Not hard to put the motor coils on a scope and look at them. I think that the practical speed limits are limited to the ability of the driver to distinguish individual pulses on the step input from the control circuits. I think I remember a RepRap discussion about some driver needing a minimum of 4us pulse duration (And presumably a minimum 4us rest duration) which means that each segment needs a minimum of 8us. At 1.8° per step, with 16X microstepping, that means 3200 pulses per revolution, 8us each, is a shade better than 1500RPM, if the coils can energize that fast. Iâm pretty sure that this speed would absolutely choke on a Mega2560 processor though. Just for reference, with a 16T pulley this represents 819.2mm/s travel speed, which I flatly refuse to believe possible with either 12V or 24V, so this step rate is clearly in excess of what we need.
Measuring the inductance, about 4.8mH per winding is a reasonable figure for most NEMA17 motors. 1 Henry of inductance would mean that the circuit would need 1 second at 1 volt to raise to 1A into a dead short. Well, say weâve for our drivers set to 1A current, so thatâs one of the variables out of the way. So 24V should (presumably) need 1/2 the time to rise to 1A than 12V (though our drivers donât really allow that. 4.8mH means a delay of about 4.8us (at 1V) which is still well under our 8us at maximum possible commanded step rate, so even at 1V potential, this isnât really a consideration in terms of electrical laws. It does mean that a 1kHz switching frequency at the driver end really does mean that weâre getting the exact same average voltage across the motor circuit, rise time and all when it comes to limiting the amount of current.
I suppose that itâs possible that we get âfull torqueâ out of the motor a few picoseconds faster at each step, but itâs more than reasonable to assume that at any reasonable operating frequency, the difference is indeed negligible.
Okay, this popped straight back into the nit-picking theoretical, but I think that it reinforces that back EMF is the #1 preventer of motor speed. Who posted that insanely fast printer running with 48V on the motor drivers again? Even agt those speeds, I think that the back EMF would be the need for the 48V much more than making the motor respond faster.
4.8mH means a delay of about 4.8us (at 1V)
Would this be 4.8 ms (not us) at 1V, which would drop to 0.4 ms (400 us) at 12V or 0.2 ms (200 us) at 24V?
Iâm (still) not arguing the significance, just want to check my understanding.
The back emf and coil inductance are both directly proportional to the winding turns and canât be separated. So you are kind of right, about inductance being neglible, but because double inductance means double back emfâŚ
Which came first, the current or the magnetic field? I remember understanding this so clearly in physics and ee classes. I was ok at the math, but the imaginary part of AC circuit analysis was something I have never been very fluent at translating to real world (should have been labs with 480VAC experiments lol). This discussion definitely is helping clarify.
Did this ever end up with a solid conclusion? Or is it still being tested?
Same as always, Faster travel rates are possible, but for the speeds we typically use, not a huge difference.
PS I can not find any more affordable 24V PS. I am willing to change since there is no solid reason to stick to 12v anymore (other than fans) but I am not finding small ones for a decent price. 24V2A would be plenty.
so is 24v better if I get 24v fans ?
Ryan doesnât sell 12V any more, pretty much everyone now builds with 24V. I have a board that can use 12V fans despite 24V input (and even 5V), so you have to see what suits you best.
On paper, 24V is better, but you wonât see any difference in day-to-day cutting. Ryan, in one of his posts above, sums it up like this:
So lets discuss this a little bit. Clearly the 24v PS allows for faster rapids. Now in terms of actual real world use, rapids will help overall but 90% of the time is spent cutting, These numbers have no effect on this.
So to even take use of this we can not really speed up cutting accel, we could speed up travel accel though. I just donât think it makes too much real world difference.
Emphasis is mine. So 24V is not worth spending extra money on, but if you can get parts for a similar price, then it is a better choice.
Thanks