Back emf, is there a way to prevent?

[REVCAM_Bob]

O.K. I have got a 1Kw servo motor connected to a "manual" machine that
has some jogging capability provided by its "own" motors. The problem is I just realized that with my servo hooked up, if I run the machine from its original Jog mode, i.e. not running with my servo motor, that the system is going to create voltage on my servo. I did not figure this out on my own, but stumbled across it in the servo amp manual. i.e. "Do not drive servos or amp will be damaged" Oh Great! So basically this is a pretty major design flaw on my part. So I called the amp guys and they told be the voltage at which the drive would potentially get damaged. I unhooked the wires to the
amp and measured the back drive voltage and sure enough, when manually jogging the machine the voltage output of the servo is twice what will damage the amp. i.e. they told me 20 volts max, and it will put out 48.
So this is maybe similar to a back drive emf situation.

Does anyone have any quick fix or some way to dump the voltage so I don't blow my amp? Thanks a million!

[chaterlea25]

Hi Bob,
You may be able to fit what is known as a flywheel diode,
this diode is wired across the motor or coil which is causing the back emf problem when supply is disconnected.
the polarity of the back emf pulse is opposite to that of the supply voltage so the diode is wired so as not to conduct when the supply is "on"
The diode should have a voltage rating of 5 or 6 times the supply voltage
hope this may help
Regards
John

[Al_The_Man]

There are different ways to dump back EMF, but it seems to me that you are driving the servo from the other motorized source? If so there are a couple of conditions, if you apply dynamic or static load dumping your non-driven servo motor is going to appear as a load to the other "machine" driving motor.
One possible scenario is to electrically decouple the servo from the servo drive when the machine is being driven from the "manual" motor, the open circuit voltage should not present a problem, I would think.
A double pole or 3phase contactor in the motor leads that only pick up when the "machine" motor is running should work.
Al.

[vger]

When I think of Back EMF I see a coil of wire with current running through it producing a magnetic field and the current flow being interupted causing the magnetic field to collapse which in turn tries to create a current flow in the coil in the opposite direction of the original flow of current. A flywheel diode will generally take care of this problem as long as the driving current is always the same polarity. In your case you will probably be wanting to drive the servo in both directions which will preclude the use of a flywheel diode. If your servo system has position feedback and you disconnect the servo motor from the amp, then change position.... when you reconnect the servo it may try to drive back to the original position.

In your configuration the servo motor (most likely permanant magnet type) is acting as a generator producing the voltage when the machine is jogged with the other motor. Al_The_Man is correct about the servo becoming a mechanical load if you attempt to dump the generated current. If you have no position feedback from the machine to the servo system, then electrically decoupling the servo from the amp may work fine. Just make sure the machine shaft has come to a full stop before re-connecting. If the machine coasts and you reconnect, the servo system will act as a dynamic brake with the amp taking the load and being dammaged.

One other alternative might be to use an electric clutch between the servo motor and the machine shaft. Just disengage the clutch before jogging with the other motor, and re-engage after the machine shaft has stopped.

Steve

[Al_The_Man]

Originally Posted by vger If your servo system has position feedback and you disconnect the servo motor from the amp, then change position.... when you reconnect the servo it may try to drive back to the original position.

Depends on your system capabilities, all the systems I use have the feature of disableing the servo loop, but still retain the ability to record the encoder position and update this as current position until re-connected, hence avoiding the jump back to original position, or recording a position error.
If back EMF dumping is used in this example, it will be active all the time which can produce constant high current through the motor, as opposed to being used only in braking, where it would occur for a short term decaying slope.
Al.

[REVCAM_Bob]

Thanks guys, I have disconnected the manual motor for now. I am going to,
for the sake of simplicity, just skip retaining the manual capabilities for now.
This allows me to continue with the cnc version testing etc. I like the 3 pole
relay thing, it would be easy to activate and deactivate with a signal from my control card etc. Of coarse I will actually need 2 somehow, one for each servo....

[NC Cams]

I've seen drives that apply a shorting shunt across the DC motor drive poles and simultaneously disconnect one leg of the amp from the DC motor when the drive is disengaged.

This is done with a NC/NO switching relay quite effectively.

When the relay is disengaged (drive not being powered) any voltage out of the motor is shunted across the NC path. Coincidentally, the NO path isolates the motor from ones side of the drive which is all that's needed.

As soon as the drive goes active, the state of the NO/NC changes and the motor functions normally.

In such cases the NC side of the switch can act as the brake coil for the DC motor when the drive is disengaged.

Depending on the drive, there SHOULD be bypassing shunting diodes already in it to deal with the coasting back EMF that gets generated - this is simply good design practice. Again, depending on the robustness of the "design", the drive should be somewhat self protective of back EMF already.

As far as simply hooking up a diode across the motor terminals, this will work as long as you only go in the one direction - reverse direction and the polarity of the back EMF changes and the diode won't conduct.

[Al_The_Man]

The only problem here as I would see it, is that the generated power will flow continuously in the driven motor and be a shorted circuit load depending on the resistance of the shunt, This will occur always when the driving motor is providing the motion. This will overheat the driven motor.
It will also appear as a braking load to the driving motor.
Al.

[NC Cams]

Shunting across the driven motor with a NC connector will result in the back EMF being dumped across a dead short which would act as a pretty hefty brake - of course, that would depend on the speed of the motor being driven and the generated back EMF.

The option would be to shunt the NC connector into a resistive load (essentially what Fanuc does on their DC drives) as well as what diesel locomotives to with their regnerateive braking systems (resistively dump the back generated EMF).

You're not going to stop the back EMF from being generated, especially when you use an external motor to jog the servo - why you don't simply use the servo to jog is not understood but, be that the case - the alteratives are thus few.

You can either divorce the drives from the overdriven motors via mechancical disconnect or electrically disconnect them with an NO relay system (what Fanuc does). Bridgeport also does this on their V2XT's as I recall.

As far as the EMF being generated is concernced, you can either allow it to internally dissipate within the motor instead of shunting it anyplace. Or you can shunt the EMF current across an NC terminal. Finallly, you can shunt it to/thru a braking resistor when the shunted NC connector/contactor is engaged. The size of the resistor (ohms and watts) depends on how much power needs to be dissipated and how fast.

IF the servo drive has the appropriate and adequate flyback diodes built in, this could all be of needless concern.

[Al_The_Man]

There is obviously a reason that two motors are used, the actual machine details are not given.
If the motor is open circuited as I previously suggested, the open circuit of 48v volts produced should cause no problem, I see no need in this case to cause unnecessary motor current and extra load on the driven motor, regenerative or non-regenerative emf braking is used on a coasting motor and/or an overhauling load in a single servo application.
An exception to the rule calls for other practical measures to suit the demands of the machine.

Incidentally on the subject of Locomotive Dynamic braking, I recently converted two retro-style 1950's style locomotives to dynamic braking that came from the Mid-west prairies, for scenic tour Rocky mountain use.
The braking resistors were 125kw x8!

[Geof]

Originally Posted by NC Cams .....You're not going to stop the back EMF from being generated, especially when you use an external motor to jog the servo...

When there is no current flow there is no back EMF. Why not simply leave all the driven motor lines open? You do creat a potential across these lines and if you spin the motor fast enough this potential could be very large because it is not being dropped across anything. But that can be easily measured.

[Mariss Freimanis]

I don't think there is any alternative except to disconnect the servomotor while it's being back-driven with another motor.

A motor becomes a generator when it's back-driven. Think of a motor as a bidirectional transducer that converts electrical power to mechanical power and vice versa. A back-driven 1kW servomotor is perfectly happy being a 1kW generator and 1kW at 48V means over 20 Amps of generator current capability. No 50 Ohm resistor is going to stop it from cranking-out 48VDC at the speed you are turning it at.

Disconnecting the servomotor is the only way to safety it against the drive.

Comment: That's an interesting drive you have there if it's limited to a back-drive voltage of only 20 VDC. Normally servodrives are full-bridge designs and a 1kW motor usually requires a 60 to 90 VDC drive to keep motor drive currents reasonable. On a full-bridge drive, the motor can be back-driven with no harm to anything (and no load on the motor-generator) up to the power supply voltage applied to the drive.

Mariss