LTC4020 Strong heating of the inductor and ITH compensation

Good day.
My project uses Vin 9 ... 28 V. Lead acid battery 6 cells .. Charge current 5 A. Load current 10A.
I am faced with the problem of strong heating of the coil with a load of more than 3 A, the coil heats up> 70 degrees. The coil used is such a SRP1770C-4R7M. All mosfet Sir664.

Could it be caused by incorrect compensation work? And can I calculate the denominations of the compensation components for the LTC4020 in the LTPowerCAD program using a project for example for the LTC3780, how true will this be?

Thank you!

Parents Reply Children
    •  Analog Employees 
    on Oct 14, 2020 4:07 PM in reply to instalator

    Thank you for posting the schematic.  I have some comments on the circuit.

    • The schematic shows the MODE pin is open-circuit.  For lead-acid charging the MODE pin should be tied to INT_VCC.
    • The voltage at SHDN pin will always be 0V if R24 is not installed.  The voltage at SHDN must be greater than 1.225V for the converter to operate.
    • The connection to VIN_REG pin will affect the converter operation when VIN<10V (VIN_REG<2.5V).  Is this what you want?  See VIN_REG pin description, datasheet p. 8.
    • It will be very difficult to control the voltage at RNG/SS with the circuit U3+Q15.  A colleague has suggested the circuit shown in the attached LTspice file.  Note that the LTspice simulation shows good behaviour, but I have not actually tried this circuit in the lab.

    For the compensation I suggest you try these values for the compensation at VC.

    • R42 - 97kΩ
    • C30 - 180pF
    • C32 - do not install

    If you still have problems, please send oscilloscope pictures of SW1, SW2, inductor current.  A horizontal sweep of 1µs/division is a good setting to analyze the waveforms.

    LTC4020 RNG_SS Control_gbcomp.asc

  • Thank you WATaylor for your reply.

    • MODE pin is connected to INT_VCC, it is marked with gray text on the diagram and there is no problem with this.
    • The SHDN input is controlled from an external controller, this is also no problem.
    • VIN_REG also does not cause problems and is made according to the technical specifications.
    • Thanks for the diagram, I will test it in practice and give feedback.

    I installed R42 - 97kΩ and C30 - 180pF, diagrams in the screenshots below.

    All measurements were made with an input voltage of 27 Volts, a load current of 8 Amperes, and a charge current of 4 Amperes.

    I tried two kinds of transistors this SiR664DP and BSC016N06.

    VC pin:


    Fig. 1 - VC pin



    Fig. 2 - SW1 (yeelow) SW2 (blue)


    Fig. 3 - Gate SiR664DP


    Fig. 4 - inductor current (I measure across a 0.01Ω resistor)



    Fig. 5 - SW1 (yeelow) SW2 (blue) 


    Fig. 6 - Gate BSC016N06


    Fig. 7 - inductor current (I measure across a 0.01Ω resistor)

    The transistors heat up to 70 degrees, the coil now heats up most likely from the transistors. After charging the battery and dropping the charging current to ~ 0, at a load of 8 Amperes, the temperature stabilizes at 55 degrees. Is this a normal temperature? I think it should be less.

    The printed circuit board is 4 layers. 

    •  Analog Employees 
    on Oct 19, 2020 8:27 PM in reply to instalator

    The voltage waveforms (Fig 1, 2, 3, 5, 6) look good.  The current waveform in fig 4 shows no current.  It looks like a measurements error. 

    The high frequency ripple in fig 7 must be a measurement artifact.  In fig 7 I see ~100mVp-p (10A), ~4MHz oscillation.  This would imply several hundred volts at 4MHz across the inductor.  4MHz is far from inductor resonant frequency (30MHz), so resonance is not an issue.  The high frequency ripple current waveform is not real.

    Have you measured the efficiency of the converter?  That is the best way to see whether there is excessive power loss.