OPA549: Heat Sinking Requirements with DC Motor

Part Number: OPA549

Dear sirs,

I'm plannig using this power op amp as a power supply for a DC motor (it drives a water pump).

I'm briefly describing the circuitry (I tried using webench but I'm getting some issue, anyway is quite simple); I'll use the nomenclature you can find on page 10 of the datasheet:
non inverting configuration, I have a Vin=0-5 VDC signal in the input (this signal comes from a low pass filter that straighten a PWM 490Hz signal), I'll power up the amplifier with V(+)=+29 and V(-)=-29 V. in order to get an (ideal) 0-24VDC in the output, I selected R2=3.9 kOhm and R1=1 kOhm (first question, is that a nice choice? I calculated around 5 mA on this resistors when Vout=24V). I also limited the current with a Rcl=3.3 kOhm (I got a nice 6.9 Amps max).
I had to use the 29V supply beacause this is not a rail to rail op amp, so the maximum offset from V(+) is 5V (correct?).

Then I got into the calculation to know the Power Dissipation of the op amp ( Itried to take as a reference the SBOA022.pdf, from TI): I don't know the internal Resistance of my load varying the voltage, so to get the power dissipation I multiplied the Current with the difference between V Output and V(+):
PD=Io*(V(+)-Vo).
I got a peak of 39 Watts (!) with Vo=17V.

Now, it seems that I need heat sinking. I would like to keep the device at least under 85C, since my project invole Hydrogen, and I don't want to kill anyone (and myself).

Can you suggest some hardware or something else? Maybe a little of explanation on how to calculate the amount of heat sinking "power" required and which characteristic look for. I can use a fan to cool down the electronics, and I thought (hope) that I can use those aluminium heat sinker chips, glued with thermal glue.

Thankyou for the help, I'm almost new to electronics, beeing and aerospace engineer does not help a lot

10 Replies

  • Hi Giorgio,

    what about the DC motor? What current is it drawing at 24V and at lower voltages?

    Kai
  • In reply to kai klaas69:

    Hi Kai,

    it draws 2Amps at 11 Volt up to 5.2 Amps and 24V. It's more or less linear.
  • In reply to kai klaas69:

    Hi Kai,

    It's more or less linear, it's better that I'll show you a raw picture of Amps drawing vs voltage. Sorry, it's very raw

  • In reply to kai klaas69:

    sorry, I'm messing up a bit

  • Hello Giorgio,

    The information regarding the motor output current vs motor voltage is very useful in determining the worst-case power dissipation for the OPA549. The worst-case point is when Vo is 19 V, and Io is 4 Amps; that results in a power dissipation Pd equal to 40 Watts. All of the other points produce a Pd from 30, to almost 40 watts. Therefore, 40 Watts is what the thermal system must be capable of dissipating while keep in the OPA549 junction temperatures (Tj) below 150 °C. Keep in mind that 40 Watts is a lot of power to dissipate from such a small package area.

    The equation presented at the top of OPA549 datasheet Pg. 11, column 2 is useful in determining the maximum thermal resistance of the heat-sink:

    θHA = [(TJ – TA)/PD] – θJC – θCH

    Applying practical values as described in the paragraph associated with the equation:

    θHA = [(125 °C – 25 °C)/40W] – 1.4 °C/W – 0.5 °C/W

    θHA = 0.6 °C/W

    Unfortunately, that is a very low thermal resistance value and would necessitate a very large, expensive heat-sink. This application really calls for a low thermal resistance heat-sink, with a high CFM fan blowing across it; especially because you want to try and keep the OPA549 body at 85 °C or less. An example of the improvement in the thermal resistance of the heat-sink/fan combination, over a heat-sink alone can be seen in Fig. 6 of this Design World magazine article:

    I do think that if you try to employ just a heat-sink alone to remove the dissipated heat that it may be difficult to attain the 85 °C maximum device temperature you would like to achieve.

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • In reply to Thomas Kuehl:

    Thankyou Thomas, very usefull and complete reply. TI forum is really usefull.

    Do you think, thanks to your experience, for example, that this heat sink ( http://de.farnell.com/abl-heatsinks/bga-pp-030/heat-sink-bga-push-pin-7-c-w/dp/2084421 ) might be enough with its 7C/W couple with a fan (2.5 Watts). I know it's nearly impossible to answer without numerical data, I just would like having more infos about "forced convection" situation.

    From what I read on the usefull article you posted, should not be enough and I should go for a 2 C/W heat sink + fan

    I think I should think on a different way to power my pump.  

    thanks,

  • In reply to Giorgio Nicola:

    Hi Giorgio,

    7C/W is too high, I would say. It should be under 2C/W from my feel. And take a heat sink with a really thick back plate. Only then the heat can spread over the whole heat sink. With a too thin heat sink the heat is trapped close to the chip. Only when the heat sink can heat up as a whole you will profit from its actual size.

    Another advantage of heat sink with a thick back plate is the enormeous heat capacitiy. This helps to make the chip temperature rise slowly.

    Buy a heatsink onto you can directly mount a 120x120mm^2 fan.

    Another option could be the use of a switch mode power supply to drive the DC motor. This would result in much less heat dissipation.

    Kai
  • In reply to kai klaas69:

    Hi Giorgio,

    Kai has good advice; I agree with using a heat-sink having a thermal resistance (θHA) of less than 2 °C/W. Unfortunately, because of the numerous physical variables involved in the application your thermal system development will require some lab work. Certainly, the optimum fan size and speed, heat-sink effectiveness, and physical location of these devices will require some experimentation to achieve effective heat removal.

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • In reply to Thomas Kuehl:

    Hi guys,
    thank you for the various inputs. I´m sorry that I reply only now but I was out of the office.
    Without having to go for a complex heat sinking device (that now I understood how to implement), can you suggest one of these step up switching regulators? I think I should look into this category ( www.ti.com/.../products.html ), but I feel a bit lost. I don´t understand if the Vout is proportional to the Vin (controlling voltage), or they feature another way to control the Vout.
    Another thing that I´m a bit concerned about is that, if I cannot control the device with DC input (which I can provide with a PWM->DAC converter), I assume that I should drive the switch with the PWM signal; the arduino provide a signal that is up to 980 Hz, quite far away from the minimun switching frequency.
    thanks,
    Giorgio
  • In reply to Giorgio Nicola:

    Hi Giorgio,

    Kai's idea to use a switch-mode power supply has merit, but unfortunately I am unable to assist you with those devices. Talking with my colleagues who have experience with switch-mode devices they indicate that this would likely require a HV buck DC-DC switching supply that has a track pin, or soft start pin, that can override the internal reference.

    I suggest you contact TI's experts in switch-mode product area on their E2E forum. I am not certain of the exact forum, but I would start with:

    If you explain what it is you are attempting to accomplish as you did with us, they may be able to assist you directly or move you to the correct forum.

    Regards, Thomas

    Precision Amplifiers Applications Engineering