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.

LM3478 - 12V to 20V (4A) Boost Controller

Other Parts Discussed in Thread: LM3478

I am having issues with the 12V to 20V 4A boost controller circuit that I have designed using the LM3478. I used the guidance of the LM3478 datasheet and AN-1286 Compensation for the LM3478 Boost Controller. The issue is that the output voltage drops from 20V unloaded to 17V loaded with a 5ohm 200W resistor trying to pull 4A. Also, the output is very noisy when loaded down. The circuit can source 2A without any noticeable voltage drop on the output. Also, the AC-DC converter can source over 8 amps of current. Below is the list of components that I have used for my circuit. Please refer to the LM3478 datasheet Figure 27 for the references.

Vin - Cosel TUNS100F12 12/100W ACDC Converter
Cin - AVX TPSD107M020R0085 100uF (2 EA), TDK C1608X7R1H104K080AA 0.1uF
L - Wurth 7447709150 15uH
D - Diodes Inc. PDS1040-13 Schottky Diode
Q1 - Vishay SI4420BDY-T1-E3, N-Channel MOSFET
Cout - AVX TPSV107K025R0100 100uF (2 EA)
Rfa - 39.2Kohms (for 400KHz) Rf1 - 120Kohms Rf2 - 7.87Kohms
Rsn - Vishay WSL2816R0200FEH 0.02ohm
Csn - 0.01uF
Rsl - 0ohms
Rc - 1.3Kohms
Cc - 0.15uF

My calculated poles and zeros for a 4A load
fz1 = 15.9KHz
fz2 = 18.6KHz
fz3 = 816Hz
fp1 - 155.2Hz
fp2 - 21.2Hz

I would appreciate any insight or recommendations. Please advise. Thanks.

  • Hi Shane,

    Are you using Ti EVM(modified) or using your own layout, Re figure 27 in the datasheet, does your layout have the feedback connection directly on COUT cap? when measuring 17V are you measuring at that point

  • I am not using an EVM. This is my own layout. The FB resistor is tied directly to COUT. My circuit schematic is identical to Figure 27 in the LM3478 datasheet. I am measuring the output of the LM3478 there at COUT.

    I put a 1x2 jumper block in the design so I can disconnect the 20V booster from the Li-ion Battery Charger IC that I am powering. This is how I am able to provide a 5ohm load to the Booster circuit.

    I am not certain if my problem has anything to do with frequency stability or compensation. I tried a 470ohm resistor for RSL and I tried a 10Kohm trimpot for RC and there was no change in the output behavior. Could it be that I need a larger value for the current sensing resistor RSN? If you are able to contact me through my registered email, I can send you the schematic and my Mathcad worksheet that I made for the component calculations and frequency response. Please advise. Thank you.
  • Hi can you send your schematic please.
    john.bowie@ti.com
  • Update: I decreased Rsn from 0.02ohm to 0.01ohm and now it can source 4A and maintain 20.5V. However, I am seeing plus/minus 1.5V spikes riding the 20.5V signal occurring at 1/400,00KHz. I know that these are caused by the inductor but are the spike levels excessive or are they normal?

    The LM3478 20V Boost Controller on my PCB is providing the power for a 16.8V Li-ion charger IC. There is a series Schottky to isolate the LM3478 from the charger IC. During the initial Fast Charge cycle when the battery current draw is at a maximum of ~3.8A, the charging IC cycles on and off continuously (i.e 3s ON and 3s OFF). During the maximum current draw I am seeing extra noise spikes on the o-scope besides the ones from the LM3478.

    The LM3478 is switching at 400KHz and the Li-ion battery charger IC is switching at 300KHz. Could this cause a conflict? I need to solve this noise issue with the LM3478 on my PCB. I am running out of ideas and parts to swap. Can anyone provide me with any insight of what to do next? Thanks