I am designing a board to convert high voltage (~400V) DC down to medium voltage (~30V) DC to provide power to subsequent boards. In its simplest form, my board will consist of the DC/DC converter and a large capacitor bank for large transients.

The DC/DC converter I am using has a limitation on it, where it can only supply 10% of its maximum rated current until it has reached 50% of its nominal output voltage. The capacitor bank consists of ~20mF of capacitance. This caused the need for my soft-start circuit.

My soft start circuit is made of an high power N-channel MOSFET (FDMS86540) and a high side gate driver (LTC7003). It is designed to limit the inrush current to below 750mA until the capacitors are fully charged and then cause minimal power dissipation (R_on of MOSFET). Below is a schematic for it.

enter image description here

The LIMITER_IN pin ties to the output of the DC/DC converter, while the LIMITER_OUT pin feeds into the capacitor bank, as well as the subsequent boards.


The issue I am running into is that the subsequent boards have components, such as linear regulators, that will turn on(enable/run pins tied to output voltage through resistive divider) before the capacitors are fully charged. Because of this, they are drawing loads of larger than 750mA, causing the capacitors to discharge and the soft start circuit is never bypassed. The idea is that once, the capacitors are fully charged, the current is no longer limited and the DC/DC converter will supply the current to the loads, and the capacitors are solely there for transients.

I could potentially put a switch at between the capacitor bank and the subsequent loads, but for my current application that would dissipate an unnecessary amount of power. Is there a way for me to adjust my soft-start circuit to account for the premature loading on the cap bank? If not, are there other architectures I can look at to limit the inrush current of the cap bank? The primary goal is to minimize dissipated power after the soft-start completes.

  • \$\begingroup\$ Is the 750mA your constraint or one set by the LT part? Seems like, if you wanted higher in-rush, you could reduce the value of R_SNS. Or modify the resistive divider values on your Run/Enable pins to activate at a higher voltage? \$\endgroup\$ – calcium3000 Nov 15 '17 at 19:03
  • \$\begingroup\$ 750mA is a limitation I have set to ensure safe charging of the caps and due to the current capabilities of the DC/DC converter. It is adjusted by Rg and Cg. The issue is that the resistive divider already sets the required voltage close to the nominal output voltage. Even though the components only turn on a few volts below the nominal, the current they draw is larger than the 750mA, so it draws it from the cap bank and discharges it. \$\endgroup\$ – GdB Nov 15 '17 at 20:12
  • \$\begingroup\$ I take it a relay between the cap bank and loads is out of the question? I know you said your application is especially power-loss-averse, but there are power-sequencing ICs, e.g. Linear's LTC2924, that could help. I'll also keep thinking... \$\endgroup\$ – calcium3000 Nov 15 '17 at 20:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.