Switching power supplies and DC servo motors

[eldata]

Is this a match made in hell?

[Al_The_Man]

Some are using switching power supplies, I prefer the unregulated linear supply, much more rugged and easily repairable if anything does happen.
Al.

[eldata]

Originally Posted by Al_The_Man Some are using switching power supplies, I prefer the unregulated linear supply, much more rugged and easily repairable if anything does happen. Al.

Agreed, but the cost of a SPS nowadays can be much less than the cost of just the power transformer for a linear PS. It's tempting in my case. Besides ruggedness, the type of overcurrent limiting in the SPS seems to be a big factor. Those that shutdown or have hiccup mode current limiting are to be avoided IMO. I'm yet to try the constant current limiting types but that seems the way to go.

[Al_The_Man]

Originally Posted by eldata Agreed, but the cost of a SPS nowadays can be much less than the cost of just the power transformer for a linear PS.

Thats the chance you take, if it fails and is unrepairable, parts availability etc, then your back to square one, the other thing is how well it handles back EMF voltages?
The other advantage I like, especially with toroid type transformers, is an overwind can be added simply for auxiliary supplies for e.g.
Al.

[epineh]

Also a little bit of galvanic protection goes a long way...Transformers aren't all that expensive and with a little bit of planning ground noise can be eliminated quite nicely.

Russell.

[probinson]

Although a toroid is certainly my favorite supply, but I have a slightly different take on the subject. I prefer the switching supply with over-current shutdown, although I add my own shutdown circuit which I'll describe later.

For me, the high power servo motors are the most important component in the entire control system simply because they were the hardest for me to get at a price I was willing to pay. If something fails I want the supply to shut down and protect those motors at all costs. I can repair every drive circuit and control board faster and cheaper than I can rewind or replace one motor. I would probably even be willing to put up with the machine occasionally just stopping if it meant the motors were safe but, thankfully, I don't have to.

The biggest problem with the SMPS current protection is that it simply turns the supply off. Sure the motors are safe but the effects of uncontrolled power cycling on the drive circuitry can be devastating. The inductive spikes from hick-up mode pulsing power into a shorted bridge can easily wipe out lots of other components. A much better solution is to add a simple current regulator in series with the driver power supply. This is a piece of cake with the super low resistance of todays MOSFETs and you don't need to choose between saving the motors or saving the electronics. In normal operation this circuit is a short, applying full supply voltage to the circuit, just like a fuse. If there is a problem, however, a fuse would simply open up generating a huge inductive spike. The current regulator, on the other hand, maintains a preset maximum current to the load while also eliminating those sudden current peaks that can trigger a switching supply to shutdown.

Current regulation by itself, of course, will only protect things for a limited time before something overheats (most likely the current regulator that is suddenly dissipating 1KW or more). You really need to have a mechanism that ignores current spike and shuts down the supply gradually if an over current condition persists. Here is where the merits of the switching supply come to light.

Nearly every switching supply uses simple voltage comparator and an opto-coupler for isolation in the voltage control circuit so basically anyone who knows what an opto-coupler looks like can locate the correct place to add a controlled shutdown circuit to a SMPS. At a minimum, this circuit should ramp down the supply voltage if the current limiter is active for more than a few hundred milliseconds.

The design phase is my favorite part of any project so I tend to go overboard and I usually find a way to get the kitchen sink in there somewhere. My current control circuit is no different. I use a separate current regulator for each servo with a PIC16F688 so that I can log over current events via RS-232 and take different actions for different types of problems. For surges less 100ms, my circuit takes no action except to log the event. This shows me things like current surges from insufficient dead time or servo chatter. If one the current regulators is active more than100ms, then the power supply voltage is lowered to maintain 1 volts across the current regulator. This is enough voltage to keep the regulator stable and everything active for troubleshooting while reducing the power dissipation of the regulator. If the short lasts more than 10 seconds then the current regulator drops to 10% of normal output further reducing the power to levels that can be safely maintained indefinitely.

Pete

[Al_The_Man]

The makes of drives I use have current limit built in, so I never usually have a problem with motor damage.
Al.

[eldata]

Originally Posted by probinson The biggest problem with the SMPS current protection is that it simply turns the supply off. Sure the motors are safe but the effects of uncontrolled power cycling on the drive circuitry can be devastating. The inductive spikes from hick-up mode pulsing power into a shorted bridge can easily wipe out lots of other components. A much better solution is to add a simple current regulator in series with the driver power supply. Pete

They come in three flavors of OC limiting.....1) Shutdown.....2) Hiccup....3) Constant current. Constant current types are by far the best for servo motors IMO. Reason being that constant current types are the least likely to cause a fault on heavy acceleration or load spikes. This can happen with shutdown and hiccup types even if you've sized the PS right for continuous loads.