Amplification of a dc current

[cc85]

Hello everybody,

I am using an l293d h bridge to drive a dc motor. This works happily and I can control the speed using pwm from a microcontroller.

I also want to determine the speed of the motor without an encoder (due to space restrictions). I can do this using the method described at http://www.romanblack.com/encoder.htm

This method doesn't work with pwm because the output to the motor isn't constant but an amplified PWM.

What I think could work is to make an amplifier that takes a pwm input and provides enough power to run the l293d motor supply. I haven't had any luck doing this. I have tried transistors and op-amps but neither provides enough current.

I appreciate any suggestions.

 

[Harald Kapp]

The idea of PWM is to minimize power lost in the driving circuit. As the driver is either on or off, it has either full voltage across the driver, but no current through the output (aka to the motor), or it has ~0V across the driver and full current through the output. The effective power delivered to the output is proportional to the duty cycle.

You can convert PWM to a DC control signal suing a low pass filter . You can then amplify this voltage to drive the motor. The cost is an increased power lost in the driver circuit (amplifier).

You'd be better off using a tachometer generator to sense the speed of the motor.

One idea (untried): The output signal of the sense circuit you linked to will contain two components:
- one frequency component from the PWM
- one frequency component from the motor's rotation
If the PWM frequency is considerably away from the frequency component caused by the motor, you may be able to separate the two using a filter. Assuming the PWM's frequency is much higher than the frequency from the rotation, a low pass after the detector may be able to blcok the PWM component and let pass only the speed component.

Alternatively you could use the PWM to drive a simple sample and hold circuit that samples the motor's voltage synchronously to the PWM. This creates a sampled system similar to a lock-in amplifier. By evaluating the sampled voltage you should be able to detect the ripple in motor current caused by the brushed, provided the PWM frequency is much higher than the frequency from the rotation.

One note: The autor of the linked page states that "GNDs do not have to be connected". I disagree. Without a GND connection there is no way for current flowing through C1 to return to the motor.

 

[CDRIVE]

A hall effect or photo transistor mechanically interfaced to the armature would work with any kind of drive, including PWM. Just input the signal to the uC and let it do the work.

Chris

 

[cc85]

Thank you Herald, I am currently investigating your suggestions and will let you know if anything works.

Thank you also CDRIVE but I am very limited with space, hence the sensorless requirement.