H-bridge problems

Discussion in 'General Electronics Chat' started by qweety, Jan 13, 2011.

  1. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    hello : ) i hope all is well.

    I wonder if anyone can help? I am looking for a h-bridge that can handle 32kh PWM and allow for a 10v, 800ma motor to be connected to the out.

    i have tested a few models but each was only able to handle up to a certain frequency before acting more like a switch than a variable.

    i've tried 3 models now and each was unfortunately a fail. although they could handle the lower frequency ( so i had them all wired correctly) and my generator is outputting at 32kh adequately as i can test this with an Led and see the pwm vary the brightness.

    any help would be great. : )
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    That's what the board is for. :)

    You mean 32kHz, right?
    Is there a particular reason you need to operate at that frequency?

    Which "models" have you tested (part numbers and circuits), and at what frequency did they start acting more like a switch?

    What were you using to drive these models?

    LEDs don't require much current.

    Your problems might be due to failure to include adequate bypass caps, long wires, etc. - but without more information, it's going to be a bit difficult to figure it out.
     
  3. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    hello thanks for your reply.


    "You mean 32kHz, right?
    Is there a particular reason you need to operate at that frequency?"

    yes 32kh is a frequency my micro controller can output ,just to keep majority of the noise above human hearing.

    "Which "models" have you tested (part numbers and circuits), and at what frequency did they start acting more like a switch?"

    a SN754410 which states in the pdf it is not capable.
    a l6202 which is not able to operate at 10v only 12 above.

    a a3953 which is the one i am working on now, i can make it work at 1kh but not at 32kh. i do not have an oscilloscope or a variable PWM generator to find the max frequency.

    here is the pdf on the h-bridge i am working on.

    http://www.google.co.uk/url?sa=t&source=web&cd=2&ved=0CB8QFjAB&url=http%3A%2F%2Fwww.allegromicro.com%2Fen%2FProducts%2FPart_Numbers%2F3953%2F3953.pdf&ei=oz8vTbr8Mdy4jAfduNCDBQ&usg=AFQjCNH4l5tlDb9TbpvffO7LcDuWZ6y5oA&sig2=75p6eLwC_wJhM-c-_4zr2w


    maybe i need to change the ref voltage? (at present it is at aprox 0.4v) or maybe a cap or resistor?

    here is the diagram i followed .
    http://www.fileden.com/files/2009/5/27/2457787/a3953-2.gif



    any help would be great, thanks.
    : )
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    OK then, use your uC's PWM output to charge/discharge a cap (say, 0.1uF) via a 4.7k resistor on the A3953's REF input. If your uC's Vcc is 5v, a 10% ON duty cycle will result in ~0.5v at the REF in.

    The driver IC already has it's own internal PWM (actually, a chopper driver) to take care of limiting the motor current.
     
  5. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    thanks again : )

    at the moment i am using a voltage divider on the ref input here is a pic
    http://www.fileden.com/files/2009/5/27/2457787/voltage divider for h bridge for REF in.gif

    you mention i need to pwm the ref input with a capacitor? is this instead of the enable pin or do i still need to pwm the enable pin? Or perhaps (with the 0.1uf cap ) this is just a simple method to reach to reach 0.5v?

    This IC has its own pwm? does that mean that it doesn't matter what frequency pwm i throw at it , it will always output the same frequency to the motor?

    thanks for your help sgtwookie : )
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Don't PWM the enable pin. You could use the enable pin to turn the driver on and off, but that's for relatively long periods of time.

    Have a look at the attached.

    On the left, the uC_out PWM signal is at ~9.6% duty cycle, 32kHz. It's feeding the other two circuits.

    The voltage on C1 in the middle circuit will eventually center around 5v*dutycycle, or in this case, 480mV. However, this limits the resolution your PWM circuit can provide.

    In the circuit on the right, C2 is charged via R2, and discharged via R2 and R3. If the PWM duty cycle is 100% (on full-time), then the voltage on C2 will eventually stabilize at 5v x R3 / (R2+R3), or 5 x 1k / (4.3k+1k) 0.94v.

    You can adjust the values of R2 and R3 to obtain whatever maximum reference voltage you wish on C2. I suggest keeping the total resistance of R2+R3 in the range of 4k to 15k Ohms. Less than that, and you'd have to increase the size of the cap. Much more than that, and you might wind up with too long of a response time - but you'll have to decide what response time you need.
     
  7. aussa

    Member

    Apr 27, 2009
    11
    0
    Hello Qweety,
    Try L298, its typical freq is 25Khz, and maximum supported is 40Khz, This is a two H Bridge driver that can hold up to 2Amps per motor, and outputs can be paralleled to give 4Amps. Make sure that your Chip is heat sinked as these chips are very sensitive to heat. Also don't forget the flywheel diodes to protect the chip from inductive kickbacks from your motor.

    Google the datasheet.

    Hope the above helps.
     
  8. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    The L298 requires the L297 controller in order to perform the automatic chopper driver function. The A3953 has the chopper driver built in. The L298 suffers from excessive voltage drop across both its' source and sink drivers; the A3953 at least saturates its' low side drivers; thus cutting its' power dissipation in half.

    Going to the L298 would be a step backwards towards a (basically obsolete) H-bridge motor driver.
     
  9. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    Thanks again.

    Sorry to be so slow on this, as i thought i was on the right track , i've just been derailed : ) Has this h-bridge a fixed pwm?at what frequency is this? i cant see it in the pdf. it appears from your pictures the reference voltage is used to tell the h-bridge what the duty cycle should be?

    thanks for your help.
     
  10. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    oh i see...we are reducing the PWM voltage down for the reference input.

    i saw the capacitor and thought we were smoothing the PWM out to dc. lol silly me.

    i think i understand now. : )
     
  11. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Did you read the datasheet?

    Start at page 6.

    The internal PWM has a fixed off-time set by Rt and Ct. It has two modes of decay; fast and slow. You'll likely generally want to use slow decay unless you're braking the motor, as it'll be more efficient. However, if you use small values for Rt/Ct, then you'll have to go to fast decay, otherwise you'll wind up with too much motor current.
     
  12. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    hi. maybe this is nearer the mark?

    "The internal PWM has a fixed off-time set by Rt and Ct."

    So by altering these values (rt & ct) i can alter the pwm frequency to the motor? i am using a 470 pf cap and 30kh resistor like in the diagram.

    rc appears to be half a cycle & toff is perhaps a full cycle?

    toff = rt x ct

    so perhaps

    (470pf) or 4.7 farads x 30000 ohm = 141000 toff

    seems a very high number , but could that be the way to calculate the pwm out from the a3953?
     
  13. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    I managed to make it work : ) i can use the micro-controllers pwm out to vary the reference voltage( using the circuit you sent on the far right), however the frequency out from the h-bridge is in the audible range, it is much lower than my 32khz from my micro-controller, i am not sure what frequency this h-bridge is outputting at the moment? as i do not have an oscilloscope.

    I feel I am near a solution but my calculation in the previous post must be missing something out?

    Any help on raising the frequency would be awesome : )
     
  14. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    What values are you using for Rt and Ct?

    The smaller they are, the shorter the off-time, and the higher the frequency will be when the motor is at speed.

    The frequency of the chopper drive is not fixed; only the off-time is fixed.
    Current to the motor is turned on until the sense input exceeds the reference level, and then is turned off for the fixed amount of time that you established with Rt & Ct.

    If you make Rt and Ct too small, you'll wind up with too much current through the motor.
     
  15. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    its o.k , i had my pwm at 1kh on my micro controller . i beleive i have it working now . : P
     
  16. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    hi, I didnt see your post there

    they are a 470 pf cap and a 30kh resistor , just like in the diagram.
     
  17. qweety

    Thread Starter New Member

    Jan 13, 2011
    14
    0
    The motor runs allot quieter now , which is great : ) It is not as strong as it once was at lower speed but I shall try to lower the values of rt & ct and see how this effects it.

    Sgtwookie you are awesome. : P
     
  18. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    If you're losing too much torque at low speeds, try decreasing Rt somewhat. That'll decrease the fixed off-time period.

    Since I know very little about your application, there's not a lot that I can do besides make some suggestions. You'll need to experiment with various values to come up with the best compromise.

    As far as the REF input, the noise at the input due to the PWMing will be at its' worst when your PWM input is at 50%. That might result in some "seeking" up and down of the motors' speed. If the response time to the REF input is not really important, you could add a resistor and another 0.1uF cap to greatly reduce the ripple on the input.

    See the attached.
     
Loading...