Actions
Forum Thread Details
  • State Verified Answer
  • Replies 20 replies
  • Answers 2 answers
  • Subscribers 77 subscribers
  • Views 2914 views
  • Users 0 members are here
Related

L298 current sensor pin, convert pulsed current signal to consumption flag

modelrailwaycat

Hello,

 

I am designing a small, simple 12V motor controller fed by a L298 H-bridge driver. The H-bridge input pins are supplied by means of PWM at 40 Hz. The motor can be connected or disconnected from the output even with the input and enable on. For this reason, I am interested in adding additional circuitry in order to obtain a binary signal of consumption using the current sensor pins; that is obtain a permanent high level while some current is detected and a permanent low level when no current is detected (after at least 1 period, but this is flexible).

 

The idea is to use a 0.5 ohm resistor to connect the sense pin to ground, and then use the voltage across the resistor to obtain the signal. I consider that the voltage measured will be from 5 mV to 1V and the Duty Cycle can be down to 1%. Because the sensor signal is also pulsed, I need somehow to filter it and obtain an stable output while consumption is 'on'. The problem is that I had not success to set properly a low-pass filter plus op-amp buffer configuration from this signal working reliably with all the mentioned conditions. I also tried to pre-amplify the sensor signal with another op-amp, then output that into the low-pass filter plus buffer, but had no success too.

 

Any idea is appreciated. Thank you very much!

  • 0

    Can you draw it up?  What op-amp are you using?  Not all op amps have common mode input range all of the way to

    BoS.

  • 0 in reply to D_Hersey

    Hi Don, thank you for your advice.

     

    Here's the circuit which I'm currently simulating. The op-amps are LM324. Maybe there's something related with what you commented regarding common mode input range in the first op-amp.

     

    Even before the PWM input sends the first pulse, the output of it is in high state. Then, when simulating a certain Duty Cycle, for instance 50%, the filter doesn't hold the value properly, causing output CONSUMTION_FLAG to switch at the PWM frequency. Maybe the design of the filter itself is wrong too.

     

    Any further comment is appreciated.

     

    image

  • 0 in reply to modelrailwaycat

    Hi John, is your motor turning ok?

    If you put a DVM across Rsense what voltage do you see when turning and when stationary?

    Do you have a scope to look at waveforms?

     

    Peter

  • 0 in reply to peterjcs23

    Hi Peter, thank you for your reply.

     

    Yes, the motor turns OK according to the Duty Cycle defined.

     

    The voltage across Rsense (R3), Vsens, is according to the current on pin 1 (divided by 2). Attached is a capture of the waveforms for a simulation of 110 ms. PWM start is delayed 50ms. As you can see, the output of the op-amp (Vsat in blue, on the right axis) has some disturbance, but its value is in high state even at the start of PWM. Vsens is in green and Vthreshold in red. Duty Cycle is 10% in this case.

     

    image

  • 0 in reply to modelrailwaycat

    The LM324 has a typical input offset voltage of 2mV and  a max of 3mV. You shouldn't really use it as a comparator if you care about much less than 10mV threshold uncertainty. (And it makes  a poor comparator any way (slow, high input bias current etc ))

     

    Just integrate the signal across R3 with an RC filter (set the time constant for at least 20 pwm periods). Then use  a low offset amp or comparator to compare the signal with a threshold. If you want to stick with a really cheap op amp like the LM324 (which does have input common made range to below ground) then add a pot to trim the offset. Having the integrator before the comparator means that you can get away with using a slow cheap op amp if you like.

     

    MK

  • 0 in reply to modelrailwaycat

    Hi John

     

    okay this is what I'd do.

    1. The LM324 is a very old fashioned opamp with poor offset and does not operate rail to rail. You might like to choose a better one like OPA4241.
    2. An op amp is a poor comparator, the LM324 is just not working in this mode at a voltage near the rail.
    3. I would make the first stage of the processing a positive gain amplifier to get a bigger signal (you need to know your max current to set the gain for full scale)
    4. The next stage could be your RC filter which needs to be around 4Hz for a 40Hz pwm say 39k with 1uF.
    5. The next stage should be your comparator, which is now working with a higher more manageable voltage. Use a LM339 for a cheap standard part. That will drive your led directly on the output which needs a pull up to make it work.
    6. If you are going to reverse the current, you will need to set up a different reference voltage as with everything referred to 0V you will not get a negative swing. And your comparator will have to work with + an - current as both represent the on state.

    Peter

  • 0 in reply to modelrailwaycat

    ...oh yes and one more thing...

     

    You show zener diodes on the motor drive for the fly back current. These should be rectifier diodes 2A fast recovery types. For example ES2A.

     

    Peter

  • 0 in reply to peterjcs23

    I'm pretty sure those are schottky diodes not zeners. The symbols used are very similar however usually with zeners the "flags" on the ends are at an angle, not straight up and down.

  • 0

    "If you are going to reverse the current, you will need to set up a different reference voltage as with everything referred to 0V you will not get a negative swing. And your comparator will have to work with + an - current as both represent the on state." -Peter

     

    Not so, the way the LM298 is set up, there will be a positive potential on the sense resistor regardless of the direction of the current flow in the load.  It is an H-driver.  Uses a unipolar supply to drive a bipolar load.

  • 0 in reply to D_Hersey

    Yep, that's true, the current always flows through the RSense the same way. You might want to invert the signal to indicate reverse operation, but for a simple led indicator, that's unlikely. Peter