Warning: Moving the steppers too fast will generate current, which goes back into the controller. If you are moving them fast enough for the lights to turn on in the controller, you are in danger of turning your controller to smoke.
I had really never thought of this but it makes sense. My simple question is what do most be do when setting up a carve (i.e. moving the gantry to the origin when its far far away) and perhaps the end of the cut (i.e. they have to move it out of the way to get the finished product).
BTW, I use a combination of manual movement with steppers off, and G0 commands with stepper on (I donât have end stops). However, I never really paid attention to how fast I was manually moving the gantry in the X/Y direction.
If you manually drag the tool head slower than the top speed that your CNC moves on its own, then you are generating voltages below the Vmot supply voltage and you are pretty safe. Itâs when you really yank it around fast that creates a danger.
IMO âturning the lights onâ is a bit conservative because the regulator can generate 5V for the CPU and lights when the induced Vmot voltage is still well below the maximum allowed Vmot.
First off, I havenât toasted any drivers this way myself. When I notice the lights come on, I slow it down. So thatâs why I wrote it this way. In fact, I wonder how much of this is really just theory vs. practice. But I also didnât want to know I told someone to push the gantry and they toasted their drivers.
I am not sure I agree with your theory completely Jamie. For one thing, when the current is generated by the motors, it may need to dump more current to reach 12V. The path that power takes may not be the same as the path it would take in normal operation. So even if the voltage was within normal limits, it may have more current, dumping through components that canât handle it. I donât think thatâs happening, but Iâm not sure it isnât.
Depends a lot on the board. Ramps certainly kills drivers this way, I hope most boards newer than that have reverse current protection. It is something to be aware of at least. Seeing the LCD light up is a good indicator of a speed limit, that is how I always handled it at least.
Reasonable enough. Let me elaborate on my mental image, which may or may not be right but it is perhaps informative.
My image of a motor driver is like an H bridge with diodes pointing from the motor terminals to V+, and also with diodes pointing from GND to the motor terminals, so the motor can pump power into the supply, even if all the transistors are off and behave like open circuit. Thatâs for a permanent magnet DC motor and a brushless motor is essentially the same except there are half H-bridges for each terminal and when back-driven it produces AC. The diodes effectively rectify the AC and backfeed the power into the supply rails as DC.
Meanwhile, a motor at a given speed behaves like a constant voltage device (AC), not a constant current device. The amount of current determines the braking torque. If the motor backfeeds into the power supply and the current has nowhere to go, then the voltage rises until the potential between V+ and GND matches the voltage that the motor is generating, and then it canât overcome that and doesnât generate any more current. It doesnât keep rising and rising until it overflows and breaks something.
I suppose there is an assumption that Vmot generated by back-driving the motors can be tolerated the same as Vmot from an external supply, itâs just an alternative source. There is an assumption that the high-voltage side of the drivers is isolated from the logic side with FETs or something in such a way that Vmot within limits is safe even if the 5V side is totally dead. There is also an assumption that a brown-out condition doesnât send the logic side or the Vmot side into a tizzy where it starts blowing up the world. I am also assuming that the diodes in the H-bridge exist according to my mental image and operate regardless of the status of the enable pin or whether external Vmot is being applied.
Now on a related note, my simplistic model of the motor drivers does not explain how a motor driver can die when an active motor is unplugged. The motor has high inductance, sure, but the transistors on an H-bridge are not inductance devices and the supply current should be absorbed by filtering capacitors. All the high voltages should stay on the motor side of the opened connection. Evidently this is not what happens so my simple model of how a motor driver works is clearly at least somewhat broken.
And in other news, I cooked another DRV8825 (different burn pattern this time). This was on my Z axis and most definitely not from dragging the machine. But like the first time, I donât know why it died. I wasnât doing anything. I just came back to it dead, and there was no dust this time. All the same symptoms except I guess I wasnât close enough at the right time to smell the smoke (I assume there was smoke).
So to sum up I think I have laid out a rather convincing case showing that I donât know what Iâm talking aboutâŚ
That all makes sense to me, except I think it is a constant current supply. So if there is very high resistance, the voltage will climb until the current finds a way through.
I am also thinking that the diodes (from gnd to motor to V+) wouldnât be moving much current normally, but they might in this situation.
As for the hot plugs, there is current going through a big inductor, so the inductor is going to want to keep that current going. But the connection is open (high resistance) so the voltage climbs until it arcs. The arcs is now flowing through the driver and is much higher voltage (like a large static shock). I think about when you unplug a vacuum cleaner that is running. It will arc, and even a few mm of air takes huge voltages to overcome.
Ok that makes sense that the high voltage is generated on the motor side and it then jumps to the driver side after it builds up high voltage. It would develop huge voltage in a tiny fraction of a second after itâs unplugged so the connector will still be extremely close and easy to arc.
And in that sense it behaves like constant current over a short timescale but Iâm still convinced over a longer timescale the voltage is linked to speed and torque is tied to current, regardless whether electrical power is going to mechanical or mechanical power is going to electrical, i.e. motor vs. generator.
With this back-driving of the motors higher in my consciousness, last night I saw something I had never noticed before. My big machine has a super heavy Z axis and when I cut power with M84, the Z axis drops, and it drops fast. I had seen this many times but it never registered until last night. This is what is killing my drivers! The symptoms are consistent with this hypothesis: both times it was the Z axis that died. Both times I âhadnât done anythingâ meaning I went to use the machine and found it inoperable, which could happen if I unknowingly killed it at the end of the previous time I used it.
Ah, the universe makes sense once again. Except we wont talk about that whole quantum mechanics/general relativity thing.
Thinking about it as a motor (in connection to being a generator) is interesting. When itâs a motor, the torque is definitely proportional to current. And the voltage is constant, but turned on and off to achieve that current. The resistance is also proportional to speed, so the duty cycle (which is analogous to voltage) is sort of proportional to the speed.
In a generator, if the resistance (from the load, which is now the microcontroller, and the generator) was constant, then youâd have voltage proportional to current. More RPMs would mean more of each. The resistance wouldnât be constant though, because when you went over 5V, the regulator would reduce the resistance by providing another path. I donât know enough about the electronics to really guess beyond that.
There is a detail here that I think is important, which is the âresistanceâ of the motor at speed is not like a resistor resistance. Itâs a back-emf which cancels some of the forward voltage resulting in a lower net forward voltage. This puts a hard upper limit on speed because when the back-emf meets the supply voltage, you canât push any current through.
And then the transfer from electrical power to mechanical is very much a two way street, where current flows in the same direction as voltage or opposite, and the torque is produced in the same direction as rotation or opposite.
I was having issues with the SD card on my RPI running Repetier server becoming corrupted:
I noticed that as I manually moved my xy that the lights on the RPI were turning on. He recommended a protection board if you have step stick modules like this one:
I havenât tried them but I found some on ebay, but only available from China. Iâm having a hard time picturing the circuit with freewheeling diodes as it seems to me that the polarity would be changing with direction of rotation. I ended up taking my buck converter out of the circuit and powering the RPI from a dedicated adapter.
So when the motor acts like an AC generator, it gets rectified to DC and sends voltage into the power supply rails. If the motor spins fast, this voltage can be high. I would think the corruption of the Pi comes from under-voltage or intermittent, crappy power being supplied by the motors-as-generators, and the additional âprotectionâ wonât get you anything.
Separate supply is safest and itâs what I do (mainly because I want to be able to shut down the motor power separately).
