Has anyone seen protection for the MOSFETs built into any circuits or plugs? Maybe using electrical paths that are close enough to have the arcs from disconnecting stepper motors go through the gap or zener diodes to have the high voltage go through them?
Can you describe the connection between MOSFETs and stepper motors in your case? Also, in what case are you disconnecting powered stepper motors?
When the stepper motors are powered by the MOSFET and there is a loose cable or crimp job, the MOSFET gets fried real quick. I would like it so you could even unplug the stepper motors completely without it frying the MOSFET. This fragile MOSFET issue is annoying. I think the issue could be resolved with 2 zener diodes or maybe just some unusually close traces on the circuit board for arcs to go across instead of through the MOSFET.
I have actually fried a few MOSFETs because the crimping job did not go as well as I thought it did.
Different MOSFETs have different characteristics, but a spark gap won’t help you any.
Interesting never heard of this problem, tell us more for how the crimping job killed the mosfet. Or in general how you think it died??
Could it in any way be static.?
@NathanielStenzel is this something you’ve run into with your own cables on a 3D printer? If so, what crimp tools have you used, and what wire? Or were they third-party cables that came with the printer?
I could imagine if a crimp didn’t actually hold wires, they could work loose as the machine runs, to the point where they start pulling apart while the stepper is powered; similarly, if the conductors were severed but the insulation held in place, they could momentarily pull apart; in either case back-emf could build up to a voltage that makes a tiny spark gap and the momentary voltage exceed the max instantaneous Vds?
I don’t think that a MOV would be a fast enough crowbar to protect. I think a TVS diode (about an order of magnitude faster than a MOV) might be fast enough but I don’t know. I do know they are generally used for electrostatic transient voltage suppression though.
Unipolar steppers could be protected by serially connected diodes because each pole has current in only one direction, and current sensing makes up for the diode forward voltage drop. The difficulty for bipolar steppers, used universally in at least consumer 3D printers, is that current has to flow in both directions (so each side has a full H-bridge).
- If you have a stepper driver with external MOSFETs, I suppose it would be possible to bridge each individual MOSFET with a TVS diode. The larger trinamic drivers use external MOSFETs to drive large steppers.
- For arbitrary drivers with integrated MOSFETs, it might be possible to put a TVS diode from each of the four stepper motor connections to ground. Sometimes these are packaged in multiples. For example, you could use one ESDCAN03-2BM3Y (41¢ in onsies, 31¢ in tensies, about 10¢ in bulk) per motor pair. If you ran them to ground through a current sense resistor, you could use that to trip an error too. One quad comparator or op-amp with latch would do it.
To be reliably effective they would have to be designed into the board of course.
Most Mosfets now days have internal snubber diodes that should protect the device.
Of course you can add them outside the unit…
People keep cooking the internal MOSFETs in stepper drivers though, through mistakes, inattention, or lack of knowledge. I’m now curious whether this approach might be useful. Also wondering whether this could be useful as a RED for small servos.
I definitely exceeded the capacity of the internal snubber on a power MOSFET when I was controlling a solenoid and forgot to put an external flyback diode in place before testing.
It happened to me on two different controller boards. One was about 2 years ago. The older was much older. The crimping job was not done with the proper tool and at the time that I considered getting the tools it was too expensive to consider. The crimp job was insufficient so it had a poor connection and arced. The current shorting between the two wires of a coil is good enough. It should not need to go to ground where it could potentially affect other things else than the MOSFET. Sure, I might be a bit thrifty buying cheap controllers but if a plug with diodes on it or some other solution like that helps to stop people from having their MOSFETs die from something else than heat then controller boards will last much longer.
I had the issue with 2 different types of connectors by the way. Not that it matters. What I came asking for was ideas on what can protect the MOSFETs and what boards could tolerate poor connections or outright unplugging. donkjr seems to think there are diodes in modern MOSFETs to protect them but you mentioned you surpassed those diodes yourself. Those coils can have a nasty kick back when power is removed from them and they discharge.
I tried looking at that and had trouble understanding if it would properly protect the MOSFET or not but if I understood the diagram right, two of those as part of a plug with an extra wire (besides the 4 motor wires) going to ground would send voltage spikes to ground or ignore pin 3 and let the two ends of the coil short to each other. I would say the closer to the coil, the safer the MOSFET is.
No, @donkjr actually knows. As do I. They show on the datasheets, including their limits.
If you want to go across the coil without going to ground, you could try Bourns SMBJ28CA-Q or so (28 there I think would work with a 24V system IIUC). About the same cost. You could put those at the motor without running a ground line to the motor.
I could imagine that some stepper motors would actually have enough room inside to add these. <2.5mm high, <4mm wide, <6mm long.
If your crimping skills are insufficient then soldering is the way to go. Just add some mechanical support to the sides of the wires next to the solder to the wires don’t fracture at the solder joint.
Personally, I’ve blown a couple of stepper motor drivers over the years and learned how to not do that any more. One was unplugging a powered motor having thought the recommendations were not to unplug while they were moving and even silent motors can be powered. The other way I blew a driver was adjusting the current limit Vref POT using a metal screwdriver. One slip off the metal cap of the POT can short the board out. I learned to use non-conductive screwdrivers for this task.
For crimping, there are definitely two schools of thought. One is to get the IWISS SN-28B which does the whole crimp in one go. The other is the Engineer PA-09 which requires separate crimps for the insulation and wire tangs.
I, of course, tried both. I go back and forth. The PA-09 is more flexible and can crimp a lot more types of connectors, but it tends to bend the pins and need some straightening. Also, the pins often stick in the jaws and bend when I try to remove them. The IWISS is a little tricky to get things aligned perfectly in one go, so it’s easy to slip and destroy the pin, but when it works the pin is usually straighter and easier to slide into the housing.
All of my stepper driver boards have adjustable pots to balance the drivers. I assume you’re supposed to set them up properly to avoid this.
@jkwilborn That’s for adjusting current. Doesn’t help any for mistakes like broken wires or knocking loose connectors out of sockets with the unit powered on.
I guess I thought that’s what generally took them out.
You could add snubbers at the motor side of the connecter soldered at the connectors and tied to a local ground.
Seems to me trying to circumvent poor fabrication is a complex endeavor.
You could crimp then solder if you want foolproof connections. Admittedly this would require precision soldering.