Overvoltage protection against brushed DC motor

[EngKyr]

Hello everybody,

I'm following the EEVBlog from few months and I think its time for me to be active on the forum.)
I'm working on a streering wheel with force feedback (hobbystic project) and I need to protect the power supply from regenerative current from the DC motor.
If I'm not wrong, I need a shunt regulator to waste the extra energy from the motor, but I need to order it. So, in the meantime, I'm wondering if I can do something with the part I'm already using.
I'm using an AC adapter at 24V to power up the motor, a L298N, freewheeling diode and bypass caps.
The caps (2 x 100uF, 2 x 100nF) can handle some energy from the motor, but I won't rely on them, so I came up with the circuit I attached.
Can it do the trick? Am I missing something?

Thanks for the help!

Best Regards

PS: sorry for my bad English

[Benta]

What extra energy?
If you spin the motor (as a generator) it will provide a voltage at the terminals equal to the back EMF. But unless you load it, it will deliver no energy.

[EngKyr]

When I turn the steering wheel fast in the same direction of the motor there is a current flowing to the power supply. This make the voltage drop accross the caps to rise (in the worst case I measured ~ 28V)

[Zero999]

When I turn the steering wheel fast in the same direction of the motor there is a current flowing to the power supply. This make the voltage drop accross the caps to rise (in the worst case I measured ~ 28V)

Use capacitors rated to 35V or 50V then. It won't damage the L298 which can withstand up to 50V.
https://www.sparkfun.com/datasheets/Robotics/L298_H_Bridge.pdf

[Benta]

Could you explain what "force feedback" means in your case?
I was thinking servo, but that doesn't seem to be the case?

[Ian.M]

Put a blocking diode in series with the PSU and a beefy 29V Zener between the L298 Vs pin and Gnd, cathode to Vs to clamp the over-voltage.  You need enough margin for PSU voltage and Zener tolerances - don't try to clamp at a lower voltage!

 A 5W 1N5362B could handle up to 170mA of generated current - unlikely unless you are using a rather large motor.

[EngKyr]

When I turn the steering wheel fast in the same direction of the motor there is a current flowing to the power supply. This make the voltage drop accross the caps to rise (in the worst case I measured ~ 28V)

Use capacitors rated to 35V or 50V then. It won't damage the L298 which can withstand up to 50V.
https://www.sparkfun.com/datasheets/Robotics/L298_H_Bridge.pdf

As I said, I already have 4 capacitors in parallel (2x 100uF 35V e 2x 100nF 50V), but what happen if I try to spin the steering wheel as fast as I can?
Can I rely only on them?

Could you explain what "force feedback" means in your case?
I was thinking servo, but that doesn't seem to be the case?

I'm using a simple brushed DC motor. A rs-555sh20 was in my junk box.
The motor only brings the steering wheel at center position. For the moment I haven't implemented a real force feedback, because I'm following a step by step procedure.

Put a blocking diode in series with the PSU and a beefy 29V Zener between the L298 Vs pin and Gnd, cathode to Vs to clamp the over-voltage.  You need enough margin for PSU voltage and Zener tolerances - don't try to clamp at a lower voltage!

 A 5W 1N5362B could handle up to 170mA of generated current - unlikely unless you are using a rather large motor.

To put a zener I need also a resistor in series with the blocking diode. Am I right?

I already have a couple of 1N5359BG, maybe I can put them back-to-back for under voltage protection also. If I consider Vz(min) I can't do that.
But if the device actually has Vz(typ), considering also the voltage drop accross the blocking diode (1n4007), I can use them.
Or Do I need more margin?

[Ian.M]

As long as the Zener voltage is above about 120%  of the PSU voltage, you don't need a resistor.  1N5359BG is useless - its 24V nominal and if its 5% low and the PSU is 5% high it will pass lots of current and either burn out itself or the PSU or trip any PSU over current protection.   Alternatively, you could build an amplified adjustable 'Zener' using a TL431 and a Sziklai pair (as a PNP 'psuedo-Darlington) emitter follower on its Anode to boost its current and power capabilities, and set it to Vpsu +10%, using a current limited bench supply.

[Zero999]

When I turn the steering wheel fast in the same direction of the motor there is a current flowing to the power supply. This make the voltage drop accross the caps to rise (in the worst case I measured ~ 28V)

Use capacitors rated to 35V or 50V then. It won't damage the L298 which can withstand up to 50V.
https://www.sparkfun.com/datasheets/Robotics/L298_H_Bridge.pdf

As I said, I already have 4 capacitors in parallel (2x 100uF 35V e 2x 100nF 50V), but what happen if I try to spin the steering wheel as fast as I can?
Can I rely only on them?

That depends on the quality of the capacitors, the temperature, how fast you spin the motor (i.e. the voltage) and for how long. Good quality capacitors will be able to withstand over-voltage for short periods of time without any damage.

[EngKyr]

As long as the Zener voltage is above about 120%  of the PSU voltage, you don't need a resistor.  1N5359BG is useless - its 24V nominal and if its 5% low and the PSU is 5% high it will pass lots of current and either burn out itself or the PSU or trip any PSU over current protection.   Alternatively, you could build an amplified adjustable 'Zener' using a TL431 and a Sziklai pair (as a PNP 'psuedo-Darlington) emitter follower on its Anode to boost its current and power capabilities, and set it to Vpsu +10%, using a current limited bench supply.

Thank you Ian.M!
I've seen the TL431 configuration, but I don't have it at the moment and I was trying to do something with the components I already have.
I'll go with the 1N5362, maybe two back-to-back

When I turn the steering wheel fast in the same direction of the motor there is a current flowing to the power supply. This make the voltage drop accross the caps to rise (in the worst case I measured ~ 28V)

Use capacitors rated to 35V or 50V then. It won't damage the L298 which can withstand up to 50V.
https://www.sparkfun.com/datasheets/Robotics/L298_H_Bridge.pdf

As I said, I already have 4 capacitors in parallel (2x 100uF 35V e 2x 100nF 50V), but what happen if I try to spin the steering wheel as fast as I can?
Can I rely only on them?

That depends on the quality of the capacitors, the temperature, how fast you spin the motor (i.e. the voltage) and for how long. Good quality capacitors will be able to withstand over-voltage for short periods of time without any damage.

Thank you Hero999!
I leave the capacitors until I get some 1N5362 from the supplier

[Ian.M]

You don't need a back-to-back pair - its going across the DC bus and the free-wheeling diodes in the L298 H-bridge will act as a bridge rectifier.  Simply use a single Zener (or a series chain of lower voltage Zeners summing to between 28 and 30V to split up the peak dissipation), cathode positive.   You would only use a back-to-back pair if you were putting them directly across the motor.

[james_s]

Could you explain what "force feedback" means in your case?
I was thinking servo, but that doesn't seem to be the case?

Force Feedback means the steering wheel is motor driven to simulate the bumps, vibration and varying resistance felt in the steering wheel of a real car. It's been around for a long time, the earliest widespread example I'm aware of was Atari's Hard Drivin' arcade game in the late 80s.

[tatus1969]

Could you explain what "force feedback" means in your case?
I was thinking servo, but that doesn't seem to be the case?

Force Feedback means the steering wheel is motor driven to simulate the bumps, vibration and varying resistance felt in the steering wheel of a real car. It's been around for a long time, the earliest widespread example I'm aware of was Atari's Hard Drivin' arcade game in the late 80s.

Force feedback is like a servo application: the PWM controller controls motor torque (instead of rpm), which simulates what you feel when driving a real car - the steering wheel tries to go straight when driving around a corner. Back to the game controller: whenever the controller wants to turn the motor in one direction, and the user forcedly turns it in the other, you will have energy recuperation. That energy charges the bus capacitors. A voltage limiter must be capable of dumping this energy. The question: how much energy does it need to handle? A very rough estimate: with my hands, I can turn the steering wheel with maybe 3 rotations per second. My hands can create maybe 10Nm of torque during that. This is the input power to the system: P = 2*PI*M*f = 188W. There is a gearbox which will probably waste quite some of that energy, the motor and controller will also heat up. So you will have to expect something in the order of 100W that you need to be able to dissipate for a short time. A zener diode, or maybe a TVS, should be large enough to handle that.

[tatus1969]

What extra energy?
If you spin the motor (as a generator) it will provide a voltage at the terminals equal to the back EMF. But unless you load it, it will deliver no energy.

That is not valid anymore, once you have a PWMed controller. These form a buck/boost converter with the motor inductance, boosting the back EMF voltage to charge the bus capacitors.

Similar to this:

[Ian.M]

Yes.  If you are doing force feedback properly, you need to be able to dump all the energy the user inputs turning the wheel against the feedback.  Quite a bit will be lost in Joule heating of the motor windings and the H-bridge transistors, but the rest gets dumped into the DC bus. If the suggested Zener isn't man enough for the job, you can boost it with a well heatsinked  NPN power transistor -  Connect the Zener beween base and collector, with a pulldown resistor base to emitter chosen so the Zener runs about 1Vbe drop below its nominal Vz, and the transisstor is *JUST* starting to turn on at Vz,

[EngKyr]

Yes.  If you are doing force feedback properly, you need to be able to dump all the energy the user inputs turning the wheel against the feedback.  Quite a bit will be lost in Joule heating of the motor windings and the H-bridge transistors, but the rest gets dumped into the DC bus. If the suggested Zener isn't man enough for the job, you can boost it with a well heatsinked  NPN power transistor -  Connect the Zener beween base and collector, with a pulldown resistor base to emitter chosen so the Zener runs about 1Vbe drop below its nominal Vz, and the transisstor is *JUST* starting to turn on at Vz,

Can I ask you a schematic to understand better the configuration?

[Ian.M]

The only things I didn't mention are Zener anode to collector, collector to +DC_bus and emitter to 0V/Gnd.  Its only three parts and three nodes, so now you can draw it yourself . . .

You wont lean if you don't try!