LM2585 12V to 24V @ 1A Step-up switching regulator

Michail Papadimitriou   mixos@otenet.gr     





This is a DC-DC step-up converter based on LM2585-ADJ regulator manufactured by Texas Instruments. This IC was choosen for it’s simplicity of use, requiring minimal external components and for it’s ability to control the output voltage by defining the feedback resistors (R1,R2). NPN switching/power transistor is intergrated inside the regulator and is able to withstand 3A maximum current and 65V maximum voltage. Switching frequency is defined by internal oscillator and it’s fixed at 100KHz.


Additional features include soft-start circuit to eliminate current spikes during start-up and internal current limit. Output voltage regulation is 4% within input voltage and load specifications.


Vin 10-15V DC
Vout 24V
Iout 1A
frequency 100KHz

Schematic is a simple boost topology arranement based on datasheet. Input capacitors and diode should be placed close enought to the regulator to minimize inductance effects of pcb traces. IC1, L1, D1, C1,C2 and C5,C6 are the main parts used in voltage convertion. Capacitor C3 is a high frequency bypass capacitor and should be placed as close to IC1 as possible.

All components are selected for their low loss characteristics. So capacitors selected have low ESR and inductor selected has low DC resistance.

At maximum output power there is significant heat produced by IC1 and for that reason we mounted it directly on the ground plane to achieve maximum heat radiation.




If you would like to receive a PCB, we can ship you one for 6$ (worldwide shipping) click here to contact us

Parts List


Part Value Package Part Number Manufacter
C1,C2 33uF, 25V, 1Ω EEV-FC1E330P Panasonic
C3 0.1uF, 50V, 0Ω 1206 C1206C104J5RACTU Kemet
C4 1 uF, 25V 1206 ECJ-3YB1E105K Panasonic
C5,C6 220uF, 35V, 0.15Ω 10×10.2mm EEV-FC1V221P Panasonic
D1 0.45 V, 3A, 40V Schottky Diode B340LB-13 Diodes Inc.
IC1 LM2585S-ADJ  TO-263 LM2585S-ADJ Texas Instruments
L1 120 uH, 0.04Ω PM2120-121K JW Miller
R1 28 KΩ 1206 ERJ-8ENF2802V Panasonic
R2 1.5 KΩ 1206 ERJ-8ENF1501V Panasonic
R3 1.5 KΩ 1206 ERJ-8ENF1501V Panasonic
R4 8.2 KΩ 1206 Panasonic
LED white SMD led 1206/1812


We’ve done a simulation of the LM2585 step-up DC-DC converter using the TI’s WEBENCH online softwate tools and some of the results are presented here.

The first graph is the open-loop BODE graph. In this graph we see a plot of GAIN vs FREQUENCY in the range 1Hz – 1M and PHASE vs FREQUENCY in the same range. This plot is usefull as it gives us a detailed view of the stability of the loop and thus the stability and performance of our DC-DC converter.


Bode plot of open control loop

What’s interesting on this plot is the “phase margin” and “gain margin“. Gain margin is the gain for -180deg phase shift and phase margin is the phase difference from 180deg for 0db gain. For the system to be considered stable there should be enough phase margin (>30deg) for 0db gain or when phase is -180deg the gain should be less than 0db.

On the plot above we see that the phase margin is ~90deg and that ensure us that the DC-DC converter will be stable over the measured range.

The next simulation graph is the Input Transient plot over time.


Input Transient simulation

In this plot we see how the output voltage is recovering when input voltage is stepped from 10V to 15V. We see that 4ms after the input voltage is stepped the output has recovered to normal output voltage of 24V.

The next graph is the Load Transient.


Load Transient simulation

Load transient is the response of output voltage to sudden changes of load or Iout. We see that the ouput current suddenly changes from 0,1A to 1A and that the output voltage drops down to 23,2V until it recovers in about 3ms. We also see that when the load is reduced from 1A to 0,1A, output voltage spikes up to ~25,5V, then rings until it recovers to 24V in about 4ms.

The last graph shows Steady State operation of DC-DC converter @ 1A ouput.


This graph show the simulated output voltage ripple and inductor current. We see that output voltage ripple is ~0,6Vpp and the the inductor current has a peak current of 2,4A. The inductor we used is rated at max 5,6A DC current so it can easily withstand such operating current and without much heating of the coil.

Operating point data (Vin=13V, Iout=1A)

Operating Values
Pulse Width Modulation (PWM) frequency Frequency 100 kHz
Continuous or Discontinuous Conduction mode Mode Cont
Total Output Power Pout 24.0 W
Vin operating point Vin Op 13.00 V
Iout operating point Iout Op 1.00 A
Operating Point at Vin= 13.00 V,1.00 A
Bode Plot Crossover Frequency, indication of bandwidth of supply Cross Freq 819 Hz
Steady State PWM Duty Cycle, range limits from 0 to 100 Duty Cycle 48.3 %
Steady State Efficiency Efficiency 93.2 %
IC Junction Temperature IC Tj 65.2 °C
IC Junction to Ambient Thermal Resistance IC ThetaJA 34.9 °C/W
Current Analysis
Input Capacitor RMS ripple current Cin IRMS 0.14 A
Output Capacitor RMS ripple current Cout IRMS 0.48 A
Peak Current in IC for Steady State Operating Point IC Ipk 2.2 A
ICs Maximum rated peak current IC Ipk Max 3.0 A
Average input current Iin Avg 2.0 A
Inductor ripple current, peak-to-peak value L Ipp 0.50 A
Power Dissipation Analysis
Input Capacitor Power Dissipation Cin Pd 0.01 W
Output Capacitor Power Dissipation Cout Pd 0.035 W
Diode Power Dissipation Diode Pd 0.45 W
IC Power Dissipation IC Pd 1.0 W
Inductor Power Dissipation L Pd 0.16 W

Configuring Output Voltage

Output voltage is configured by R1, R2 according to the following expression (Vref=1,23V)

VOUT = VREF (1 + R1/R2)

If R2 has a value between 1k and 5k we can use this expression to calculate R1:

R1 = R2 (VOUT/VREF − 1)

For better thermal response and stability it is suggested to use 1% metal film resistors.





4 Response on “LM2585 12V to 24V @ 1A Step-up switching regulator

  1. Hi! Thank you for this schematic. My only question is about PCB size needed for this (too huge for my application). I need a 24V output from a 12V battery, at 1A, but only by peaks (ie. 10ms of 24V@1A every second, max).
    After a simulation on WEBENCH, I need a 220uH@3,5A inductor (why don’t I get the same results as yours?). What if I use a 220uH@2,8A inductor? It should be OK given that the average output current should be less than 1% of 1A (10mA), no?
    Thank you

    1. Hi Thomas, I reviewed the simulation on my end and a 150uH/4.3A inductor has been proposed as ideal. I don’t know why this is different from your results. Probably both solutions works quite well. I also checked the inductor current on steady state operation and it has a max current of 2.35A, so probably there is some room to use a lower value inductor.

Leave a Reply