DC motor starting current

[hamdi.tn]

Hi,

Am trying to design where i control between 1 and 4 24V DC motor, with nominal current of 4A and ( as datasheet said 23A staring current ) and that's to damn high.
on the total cycle only one motor run at the time so typically i only need 120W power supply, am counting to use 200W one.
but am not sure how to limit motor inrush current without causing PS to shutdown or the damage the H-Bridge.

-NTC seems to me a nice easy way to do it, except they take sometime to cool down and am not sure yet if this ok, so this is a solution that i will hold until i test the motor.

-Does applying a PWM command signal on the H-Bridge help rising up the voltage smoothly and avoid inrush current.

-to keep price down am planing to design a single H-Bridge and switch between motor with relays , should i put some protection to detect if a relay contact are somehow soldered with current or that will be a little to much.

Any idea.

Thanks ?

[Ian.M]

Add a current sensor and inhibit the PWM on a cycle by cycle basis if the current exceeds 80% of the H-bridge or PSU rating.   Many MCU PWM units have this sort of inhibit functionality built in, if not, you can yse external logic.

[max_torque]

Understanding how a motor "works" is the first step to understanding how to control one!


When at rest (ie not spinning) a motor is just an inductor.  It has a small DC resistance, the larger (physically or powerfully) the motor, the smaller this DC resistance will be.  Typically, you'd be looking at a few ohms for a small motor, and just several milliohms for a big one.


So, if you apply a fixed dc voltage to that motor when stationary, V= IR, so you get a large current flowing.  However, due to the inductance of the motor (it is after all, designed to make a large magnetic field in the air gap)that current doesn't start flowing immediately, but takes time to build (how long depends on the inductance)

The important point however is that as a motor starts to turn, it generates a voltage in the opposite direction that the supply.  This "back EMF" fights against the supply voltage, and hence acts to reduce current flow.  The important point is that this back EMF is just proportional to the speed of the motor and nothing else (at a first order).


Imagine a motor that has a back emf characteristic of 1 volt per 1000rpm.  If you were to connect it to a 5v DC supply, it would start spinning until just before the back emf equaled the supply voltage (in this case, just below 5000rpm).  It cannot ever quite reach 5000rpm because if it did, the back emf would be exactly the same as the supply voltage, and so there would be no forward voltage left to drive current through the motor (and hence make torque).

But, lock that motor solid, and the "stall current" would just be set by the supply voltage and the motors DC resistance, because when it's stattionary there is no back emf.



So how does this all apply?  Well, if you have a system where the motor could be completely stalled, and your "controller" just applied 5v say, then yes, you need to be able to supply the full stall current.  But if your system uses some other technique to apply a proportion of the supply voltage (PWM for example) and can vary how much voltage it does apply (and even better measure the current that is created) then the stall current value is much less critical.

You can "soft start" the motor by only applying a small percentage of the supply voltage whilst the motor is stationary, and hence limit the current drawn.

Most motor controllers will have in "inner" current control loop (that drives the output voltage (via pwm) until the current reaches the target value) and an "outer" speed control loop that demands the current necessary to get to the speed target.  In this way, the inner loop prevents overcurrents and controls motor torque, and the outer loop prevents overspeeds and controls motor speed.

[hamdi.tn]

Add a current sensor and inhibit the PWM on a cycle by cycle basis if the current exceeds 80% of the H-bridge or PSU rating.   Many MCU PWM units have this sort of inhibit functionality built in, if not, you can yse external logic.

That's on the menu too, using MCU with inhibit  functionality is a nice idea thought  thanks

Understanding how a motor "works" is the first step to understanding how to control one!


When at rest (ie not spinning) a motor is just an inductor.  It has a small DC resistance, the larger (physically or powerfully) the motor, the smaller this DC resistance will be.  Typically, you'd be looking at a few ohms for a small motor, and just several milliohms for a big one.


So, if you apply a fixed dc voltage to that motor when stationary, V= IR, so you get a large current flowing.  However, due to the inductance of the motor (it is after all, designed to make a large magnetic field in the air gap)that current doesn't start flowing immediately, but takes time to build (how long depends on the inductance)

The important point however is that as a motor starts to turn, it generates a voltage in the opposite direction that the supply.  This "back EMF" fights against the supply voltage, and hence acts to reduce current flow.  The important point is that this back EMF is just proportional to the speed of the motor and nothing else (at a first order).


Imagine a motor that has a back emf characteristic of 1 volt per 1000rpm.  If you were to connect it to a 5v DC supply, it would start spinning until just before the back emf equaled the supply voltage (in this case, just below 5000rpm).  It cannot ever quite reach 5000rpm because if it did, the back emf would be exactly the same as the supply voltage, and so there would be no forward voltage left to drive current through the motor (and hence make torque).

But, lock that motor solid, and the "stall current" would just be set by the supply voltage and the motors DC resistance, because when it's stattionary there is no back emf.



So how does this all apply?  Well, if you have a system where the motor could be completely stalled, and your "controller" just applied 5v say, then yes, you need to be able to supply the full stall current.  But if your system uses some other technique to apply a proportion of the supply voltage (PWM for example) and can vary how much voltage it does apply (and even better measure the current that is created) then the stall current value is much less critical.

You can "soft start" the motor by only applying a small percentage of the supply voltage whilst the motor is stationary, and hence limit the current drawn.

Most motor controllers will have in "inner" current control loop (that drives the output voltage (via pwm) until the current reaches the target value) and an "outer" speed control loop that demands the current necessary to get to the speed target.  In this way, the inner loop prevents overcurrents and controls motor torque, and the outer loop prevents overspeeds and controls motor speed.

Thanks for the explanation, am aware how motor works , if fact i did this kind of system before , but i did it the old way and for a smaller current. Am trying to build something that cover most problem that can happen with DC motors, and keep light BOM.
soft-start is exactly what am thinking of, do you have a motor controller chip in mind that you can recommend ? am thinking about DRV8701 from ti.

[max_torque]

You need to tell us more about the motors and the application ie:

1) Brushed or brushless motors,  with or without position feedback?
2) What is the load?(can they "stall", is exact speed or positional control required)
3) Cost aspirations? (if you're just making a few, then the price of the parts is pretty irrelevant, if you're gonna make 100s, 1000s or 10,000 then piece costs become important)
4) Interface?  How will the motors be commanded?
5) Robustness? Should the system be completely foolproof?

[hamdi.tn]

You need to tell us more about the motors and the application ie:

1) Brushed or brushless motors,  with or without position feedback?
2) What is the load?(can they "stall", is exact speed or positional control required)
3) Cost aspirations? (if you're just making a few, then the price of the parts is pretty irrelevant, if you're gonna make 100s, 1000s or 10,000 then piece costs become important)
4) Interface?  How will the motors be commanded?
5) Robustness? Should the system be completely foolproof?

1- Brushed , without position feedback
2- it will be used to turn a lead screw , holding some working platform. it will stall if stopped by an obstacle for sure.
3- not sure yet around 100 is the most probable
4- through H-bridge, and a H-bridge controller and uC obviously
5- completely foolproof, except some basic protection no , but am trying to give it thought so i maximize protection in theory , then i will choose what to keep.

i think what am asking for here, is to help decide witch soft-start mechanism is most practical considering the fact that am definitely unwilling to put something over 200W as a power supply. but the current could reach 23A and will probably cause an over-current fault.

[Ian.M]

If I was using a PIC with an ECCP module supporting  enhanced PWM modes, I'd use a 10A current sensor (hall effect or shunt + amplifier) in the supply to the H bridge.   I'd run the current sense output to a comparator input on the PIC, with the other input from the on-chop DAC.  The comparator output would be internally routed to the PWM auto-shutdown,   Each PWM cycle, the auto-shutdown would shut off the PWM pulse if the current exceeded the threshold set by the DAC and auto resume at the beginning of the next cycle so limiting the current.  The comparator input could also be read as an ADC channel for software current measurement, and the autoshutdown flag can be monitored to detect if its in current limiting mode.

You only need  five half bridges to control four motors as you can common one side of all the motors and drive that with PWM, and enable the opposite half bridge for one motor to drive it while leaving the other three hi-Z.  The extra half bridges only need GPIO pins to control them as you only change them to select a motor and forward or reverse. A couple of quad gate drivers and eight extra MOSFETS will probably be cheaper than four 10A relays and certainly will be more reliable.

[nuno]

i think what am asking for here, is to help decide witch soft-start mechanism is most practical considering the fact that am definitely unwilling to put something over 200W as a power supply. but the current could reach 23A and will probably cause an over-current fault.

You just need to limit current on the motor, using PWM and a cycle by cycle limiter (for 20A peak, and even higher, a resistor is acceptable as a current sensor, unless you have very specific needs). If you do that, in a start situation you can have those 23A at the motor but only 5A or 6A being pulled from the power supply. A motor controller it's a kind of a buck converter.

[Circlotron]

If you do that, in a start situation you can have those 23A at the motor but only 5A or 6A being pulled from the power supply. A motor controller it's a kind of a buck converter.

If you use that method, make sure you have a large, low ESR capacitor right at the H bridge so your power supply will only see the average current, not the peaks and maybe still trip.

[Circlotron]

-NTC seems to me a nice easy way to do it, except they take sometime to cool down and am not sure yet if this ok, so this is a solution that i will hold until i test the motor.

NTC thermistor would work okay if once the motor is up and running you bypass the NTC with relay contacts. It will then start cooling down right away, not waiting until the motor is powered down.

[hamdi.tn]

i think what am asking for here, is to help decide witch soft-start mechanism is most practical considering the fact that am definitely unwilling to put something over 200W as a power supply. but the current could reach 23A and will probably cause an over-current fault.

You just need to limit current on the motor, using PWM and a cycle by cycle limiter (for 20A peak, and even higher, a resistor is acceptable as a current sensor, unless you have very specific needs). If you do that, in a start situation you can have those 23A at the motor but only 5A or 6A being pulled from the power supply. A motor controller it's a kind of a buck converter.

totally convinced with this method, i will prototype a circuit using PIC with ECCP and try this. 

-NTC seems to me a nice easy way to do it, except they take sometime to cool down and am not sure yet if this ok, so this is a solution that i will hold until i test the motor.

NTC thermistor would work okay if once the motor is up and running you bypass the NTC with relay contacts. It will then start cooling down right away, not waiting until the motor is powered down.

Will keep space on the prototype board to add NTC in series with the motor so i can add it if needed. 

thanks everyone, really helpful 

[nuno]

i think what am asking for here, is to help decide witch soft-start mechanism is most practical considering the fact that am definitely unwilling to put something over 200W as a power supply. but the current could reach 23A and will probably cause an over-current fault.

You just need to limit current on the motor, using PWM and a cycle by cycle limiter (for 20A peak, and even higher, a resistor is acceptable as a current sensor, unless you have very specific needs). If you do that, in a start situation you can have those 23A at the motor but only 5A or 6A being pulled from the power supply. A motor controller it's a kind of a buck converter.

totally convinced with this method, i will prototype a circuit using PIC with ECCP and try this. 

Post your circuit before prototyping

[hamdi.tn]

i think what am asking for here, is to help decide witch soft-start mechanism is most practical considering the fact that am definitely unwilling to put something over 200W as a power supply. but the current could reach 23A and will probably cause an over-current fault.

You just need to limit current on the motor, using PWM and a cycle by cycle limiter (for 20A peak, and even higher, a resistor is acceptable as a current sensor, unless you have very specific needs). If you do that, in a start situation you can have those 23A at the motor but only 5A or 6A being pulled from the power supply. A motor controller it's a kind of a buck converter.

totally convinced with this method, i will prototype a circuit using PIC with ECCP and try this. 

Post your circuit before prototyping

will do