• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 63 subscribers
  • Views 1181 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Current Mode Buck Converter Sense Resistor vs. Efficiency

A A2
Prodigy 60 points


Hi Fellow EE's,

            I have an old buck converter design that uses an off-the-shelf current mode controller IC . A sense resistor is used to measure the inductor current and compensate accordingly. The original design is rated for a current that is much higher than necessary. As a result the efficiency is close to 50-60% with the typical small load used in the application

           Since the sense resistor was chosen for a current about 14 times greater than the nominal in the application, I'm considering using a larger sense resistor to bring the current range closer  to the range of currents drawn by  the actual load. This is per the datasheet formula that relates the Rsense value to IOUTmax. Will this ultimately increase my efficiency, and why?

           I have done an experiment in which I substituted a larger sense resistor value and the efficiency improved about 10-20% with no other changes. I really would like to know the mechanism behind this.

Thanks,

New Power Supply Designer

  • 0
    TI__Guru**** 191445 points

    I fond that somewhat implausible.  In general, the the sense resistor is used to measure the peak current for each pulse and compare it to the current determined by the converter needed for that particular load.  The only losses associated with it are the I^2 R losses in teh resistor.  A smaller resistor value should have less loss and higher efficiency but the resistor values are typically very low, in the order of milliohms.  Of course you do not state what controller you are using so there may be something else going on.  Perhaps you could post your schematic.

  • 0 in reply to JohnTucker
    Prodigy 60 points

    Hi John,

               Thanks for the reply. The controller is an LTC1629-PG (http://cds.linear.com/docs/en/datasheet/1629f.pdf)  and the circuit is essentially the same as the app circuit in the datasheet. The equation for Rsense is Rsense=(50mV*N)/(IOUTmax) where N is the number of phases used.

                 I figured since the equation includes IOUTmax that somehow Rsense controls the current range. I know that if the Iout current for a given Rsense is exceeded the controller won't operate(I think it goes into hiccup mode). The efficiency vs. output current graph is given in the datasheet and the efficiency at nominal current falls in the 50-60% range. I was figuring that by changing the sense resistor I could possibly shift the efficiency to our range of operation.

                I know a higher resistance for Rsense will dissipate more power leading to reduced efficiency. However, the nominal output current is low and the resistance might be at most 13 milliohms, so that source of power loss isn't significant.

                I would just like to determine if there is a way to change component values to improve the controller efficiency in my operating range with a 1.5 A  upper limit. Being new to SMPS i'm not really sure if this is possible without going for a new controller which is a really undesirable option since a new layout and design will be required.

       

  • 0 in reply to A A2
    TI__Guru**** 191445 points

    I'm afraid I can't help you too much with an LTC device, except to recommend you use TI power for your application.  If you post your input and output requirements I can suggest a suitable device.

  • 0 in reply to JohnTucker
    Prodigy 60 points

    Hi John,

                   Thanks for the reply. I understand that you cannot help with LTC issues; but I'm sure the basic CCM operation is similar in TI devices. In a comparable current controlled buck regulator that is made by TI, can the sense resistor be changed such that the nominal current falls in a higher efficiency range. Or is the efficiency vs. output current graph fixed for each particular controller, and the only option is to swap controllers?

  • 0 in reply to A A2
    TI__Guru**** 191445 points

    For current mode control, there are essentially two control paths.  The PWM duty cycle is directly controlled by the current.  So for constant load current, there is a required duty cycle for a given Vin and Vout.  When the load current changes, either increasing or decreasing from the current supplied by the converter, the output voltage will decrease with increasing load or increase with decreasing load.  A portion of that voltage is fed back to a transconductance error amplifier that will convert the error votage to a new current set point.  Most controllers use peak rather than average current detect, so the load current is I peak - 1/2 ILp-p.  So the "nominal current" is really set by your load condition, not the sense resistor.  That sens resistor is just sized so that the maximum threshold voltage of the comparator is not exceeded at your maximum current conditions.  For 10 A max load, the resistor is only 5 mohm and the losses are .5 W at full load.  That is a sizeable loss. So if you lower the resistance value, you will get better efficiency but give up dynamic range in your current sense loop.

    In general, the controller does not affect efficiency too much.  Efficiency is equal to Pout / (Pdiss + Pout).  Power out is Vout * Iout so ingeneral you will always get higher efficiencies for higher output voltages.  The power dissipation can generally be considered as conduction losses + switching losses + quiescent losses + losses in the various passives.  Conduction losses dominate at higher currents and switching losses dominate at lower currents and are mostly determined by the switching FETs (and catch diode for non synchronous converters).  Higher operating frequencies and higher input voltages have higher switching losses, so in general, lower frequency and lower input voltage converters will have higher efficiency. Quiescent losses are mostly constant and those losses are the main contribution of the controller.  You can look for controllers that feature low Iq specs.   Most of the passive losses are relatively small except the losses in the inductor.  Low value inductors generally are wound with lower resistance wire than higher values in the same package and offer higher efficiency.

  • 0 in reply to JohnTucker
    Prodigy 60 points

    Hi John,

              Thanks for the extremely detailed explanation. It makes sense that the resistor is sized in a way to avoid exceeding the comparator threshold at high output currents. That answered my question completely.