Integrated soft start for DC/DC converters

[arghpok]

Hi all,

I'm working designing a buck or boost (not buck/boost) DC/DC converter, and I have problems regarding the inrush current. The maximum tolerable is 1A but I'm achieving 2+ when powered on. How can I implement a soft start without introducing more elements in the power path ?

I've seen resistors in series in the power path for the start-up that are short-circuited after the required voltage has been reached. However, that would need a HUGE switch in parallel with the resistor (In fact the resistor has to be huge as well to let 500mA pass during the turn-on), and that would introduce another dissipative element in the power path.

Is there some way to control one of the powerMOS of the converter as a "LDO" during the start-up and then changing it to the regular control trough the drivers ? With you could clamp the Vgs in the transistor so you limit the current passing though it by modulating the resistance. It sounds too good to be true and I haven't seen it yet. Anybody has experience with this kind of circuits ? or another solution for this problem ?

Thank you,

 

[leebluer]

try to control the time of start up, but not directly control the current.

 

[arghpok]

Here's the deal.

The input power comes from a solar panel. When plugged in, the initial voltage can be anything between 0 and 7V. The circuit needs a voltage above 2V to start controlling the DC converter at the input and create an ouput of 4.3 at the MPP. When plugged in though, we need to let the energy pass through the switches on the DC converter to charge the output capacitor to at least 1.68V, because is this voltage the one that sources the controlling circuitry (it cannot be connected to the main input due to its variation between 0-7V and our tech is not high voltage. Moreover, the MPP can be below the voltage operating point of 1.8V). My question is if it is possible to control one of the switches resistance at the beginning to limit the inrush current, and after the control takes over, put it into its min resistance (Rdson) ???

 

[dick_freebird]

Many DC-DCs and PWM controllers have this feature.
It's actually abnormal, to not - it's a common problem
(not everybody can go slug the upstream source full-on
or something bad's gonna happen).

If you know Cload and Iload, and max allocated Iin, you
can figure the Iin allocatable to filter charging (Cload)
and hance dV/dt, dV is your output voltage so dT is
your soft start ramp time.

Now operating a converter below its undervoltage lockout,
that's not gonna work.

You might not want to bother with trying to cover the
whole range of output voltages. Look to your "use model"
for insolation and cell I, V variation. You might be willing
to just forget about the last (say) 5% of delivered energy
per day if it doubles material cost and triples design effort.
Just because there is some condition where MPP is 1.8V
doesn't mean anything in itself; maybe that's a case where
there's no meaningful power output anyway, so forget it.

But if your source is solely solar cells I don't think you
need soft start, because you want to take every bit of
their current and pump it out.

You might want to think a bit about architecture.
You might do better with a very crude boost at the
front to pull out all the energy you can to a working
semi-regulated voltage, and then a buck charge
controller to make the battery happy.

Further if you are unable to use all the panel output
and know this for sure, you might scrub off the high
end by cell shunting, pretty cheaply. Better to return
the current intra-cell at low voltage and power, than
send it out to stress other things and make more heat
everywhere. A good enough shunt FET is dirt cheap
and one more pad on a chip, close enough to free.