What part/circuit is needed to control in rush and standby/operating currents?

[KJ]

Pardon the likely lack of detailed info, I'm trying to get ideas to
help a colleague out and I'm in an area where I admittedly have very
little design experience.

The basic problem that he is trying to solve is managing in rush
current at device power up and then current spikes when transitioning
from standby to active mode. The parameters that I know about are:

Input Voltage = 24V (DC)
Peak in-rush current (measured) = 65 A for ~100 us
Standby current = 80 mA (measured)
Operating current 2 A (measured)

To limit the in-rush current at power on he's looking at a thermistor
which seems to limit the peak in-rush current nicely but then it
presents too high of a resistance when switching from standby to
operating mode.

What would be needed then to go along with the thermistor is something
that would turn on some time after the initial in-rush current has
subsided that could handle the 2A change in demand when going from
standby to operating mode.

A reed relay in parallel with the thermistor looks like it might do
the trick if there is one rated for ~3-4A (~50-100% margin guessing on
the 2A) and a turn on current around 20-40mA (~25-50% of the measured
standby).

Is this approach viable? If so, then are there any suggestions on
particular parts? Since I'm far from expert in this area I'm sure
there are better solutions to pursue. Any suggestions in that regard?

Thanks in advance.

KJ

 

[Paul Mathews]

Pardon the likely lack of detailed info, I'm trying to get ideas to
help a colleague out and I'm in an area where I admittedly have very
little design experience.

The basic problem that he is trying to solve is managing in rush
current at device power up and then current spikes when transitioning
from standby to active mode. The parameters that I know about are:

Input Voltage = 24V (DC)
Peak in-rush current (measured) = 65 A for ~100 us
Standby current = 80 mA (measured)
Operating current 2 A (measured)

To limit the in-rush current at power on he's looking at a thermistor
which seems to limit the peak in-rush current nicely but then it
presents too high of a resistance when switching from standby to
operating mode.

What would be needed then to go along with the thermistor is something
that would turn on some time after the initial in-rush current has
subsided that could handle the 2A change in demand when going from
standby to operating mode.

A reed relay in parallel with the thermistor looks like it might do
the trick if there is one rated for ~3-4A (~50-100% margin guessing on
the 2A) and a turn on current around 20-40mA (~25-50% of the measured
standby).

Is this approach viable? If so, then are there any suggestions on
particular parts? Since I'm far from expert in this area I'm sure
there are better solutions to pursue. Any suggestions in that regard?

Thanks in advance.

KJ

A very simple circuit to accomplish your goals would employ an N
MOSFET in the negative rail or a P MOSFET in the positive rail. The
gate signal for either type of FET would ramp the gate voltage up over
a few milliseconds by means of a resistor and capacitor. To complete
the picture, add a diode to discharge the timing capacitor when power
is removed. Should be less expensive than a relay and use almost no
bias power.
Paul Mathews

 

[KJ]

A very simple circuit to accomplish your goals would employ an N
MOSFET in the negative rail or a P MOSFET in the positive rail. The
gate signal for either type of FET would ramp the gate voltage up over
a few milliseconds by means of a resistor and capacitor. To complete
the picture, add a diode to discharge the timing capacitor when power
is removed. Should be less expensive than a relay and use almost no
bias power.
Paul Mathews- Hide quoted text -

- Show quoted text -

Paul,

I had considered that arrangement as well, but since the FET gate is
an R/C delay of the FET drain voltage it seemed to me like if the
external 24V power brick voltage blipped up a bit it would create a
negative gate to source voltage which would increase the source to
drain resistance which would then drop the drain voltage (and
therefore the gate voltage) which would make the gate to source
voltage even more negative than it started at. This would continue
and the feedback would cause the FET to shut down. If that happened,
the FET would only come back alive by the conduction through the
thermistor and the whole thing would be in some form of oscillation.
Or am I just completely missing something?

KJ

 

[[email protected]]

Pardon the likely lack of detailed info, I'm trying to get ideas to
help a colleague out and I'm in an area where I admittedly have very
little design experience.

The basic problem that he is trying to solve is managing in rush
current at device power up and then current spikes when transitioning
from standby to active mode. The parameters that I know about are:

Input Voltage = 24V (DC)
Peak in-rush current (measured) = 65 A for ~100 us
Standby current = 80 mA (measured)
Operating current 2 A (measured)

To limit the in-rush current at power on he's looking at a thermistor
which seems to limit the peak in-rush current nicely but then it
presents too high of a resistance when switching from standby to
operating mode.

What would be needed then to go along with the thermistor is something
that would turn on some time after the initial in-rush current has
subsided that could handle the 2A change in demand when going from
standby to operating mode.

A reed relay in parallel with the thermistor looks like it might do
the trick if there is one rated for ~3-4A (~50-100% margin guessing on
the 2A) and a turn on current around 20-40mA (~25-50% of the measured
standby).

Is this approach viable? If so, then are there any suggestions on
particular parts? Since I'm far from expert in this area I'm sure
there are better solutions to pursue. Any suggestions in that regard?

Thanks in advance.

KJ

Have you looked into hot swap controller chips? They sometimes
(usually) have a soft start circuit.
One I've used is the LTC4211, it won't do for your voltage but take a
look at the idea.

 

[D from BC]

Pardon the likely lack of detailed info, I'm trying to get ideas to
help a colleague out and I'm in an area where I admittedly have very
little design experience.

The basic problem that he is trying to solve is managing in rush
current at device power up and then current spikes when transitioning
from standby to active mode. The parameters that I know about are:

Input Voltage = 24V (DC)
Peak in-rush current (measured) = 65 A for ~100 us
Standby current = 80 mA (measured)
Operating current 2 A (measured)

To limit the in-rush current at power on he's looking at a thermistor
which seems to limit the peak in-rush current nicely but then it
presents too high of a resistance when switching from standby to
operating mode.

What would be needed then to go along with the thermistor is something
that would turn on some time after the initial in-rush current has
subsided that could handle the 2A change in demand when going from
standby to operating mode.

A reed relay in parallel with the thermistor looks like it might do
the trick if there is one rated for ~3-4A (~50-100% margin guessing on
the 2A) and a turn on current around 20-40mA (~25-50% of the measured
standby).

Is this approach viable? If so, then are there any suggestions on
particular parts? Since I'm far from expert in this area I'm sure
there are better solutions to pursue. Any suggestions in that regard?

Thanks in advance.

KJ

Power mosfet + Isense resistor + op amp

Chose high side or low side for limiting.

Loads of fun to design..especially high side design..

As a bonus, the circuit can act like an electronic fuse if a time out
circuit is added too..


D from BC

 

[PeteS]

Paul,

I had considered that arrangement as well, but since the FET gate is
an R/C delay of the FET drain voltage it seemed to me like if the
external 24V power brick voltage blipped up a bit it would create a
negative gate to source voltage which would increase the source to
drain resistance which would then drop the drain voltage (and
therefore the gate voltage) which would make the gate to source
voltage even more negative than it started at. This would continue
and the feedback would cause the FET to shut down. If that happened,
the FET would only come back alive by the conduction through the
thermistor and the whole thing would be in some form of oscillation.
Or am I just completely missing something?

KJ

Hot swap controllers are designed specifically for this issue. You could
design your own, but why when there are plenty of single chip (well,
with a few support components) out there?

The large initial inrush is because the output (to the unit) is not
being ramped (soft start); it's charging up caps that are either large
or have low esr or both.
You don't have to be a hot-swappable unit to use a hot swap controller.

The best offerings are from Linear Tech and maybe TI. Analog devices
does some as well IIRC. For a few bucks you get what you need.

Cheers

PeteS

 

[Paul Mathews]

Paul,

I had considered that arrangement as well, but since the FET gate is
an R/C delay of the FET drain voltage it seemed to me like if the
external 24V power brick voltage blipped up a bit it would create a
negative gate to source voltage which would increase the source to
drain resistance which would then drop the drain voltage (and
therefore the gate voltage) which would make the gate to source
voltage even more negative than it started at. This would continue
and the feedback would cause the FET to shut down. If that happened,
the FET would only come back alive by the conduction through the
thermistor and the whole thing would be in some form of oscillation.
Or am I just completely missing something?

KJ- Hide quoted text -

- Show quoted text -

You don't need a thermistor...just the FET in series. On power up, the
RC delays the gate voltage and the FET transconductance increases
slowly. On power down, a diode in parallel with the R quickly
discharges the C across the gate. Problems can arise if the gate
voltage stays too low too long, since power dissipation in the FET is
then high. The fancy hot-swap chips can take care of that kind of
condition.
Paul Mathews