# Inrush current and BFC's

Discussion in 'Electronic Design' started by John Devereux, Oct 23, 2007.

1. ### John DevereuxGuest

Hi,

I am putting together a high current supply and wondering about inrush
current.

The situation is a center-tapped 110V 3kVA transformer, feeding a
bridge rectifier and a large, high-ripple current capacitance:

..
.. -------- +70V
.. ----| |-------o------> 20A rms >---
..|| / | AC + | | |
..|| / 55 | | ----- 20,000uF [LOAD]
..|| / | | ----- |
..|| /-- | | | |
..|| / | | | |
..|| / 55 | | | |
..|| / | AC - | | |
..|| /----| |-------o------< 20A rms <---
.. -------- -70V
..
..

Simulation shows hundreds of amps for the first mains cycle!

Do I need to do anything to stop the local substation tripping out
when I plug it in?

Or will the transformer limit the current to a sane value?

I have looked at thermistor based "inrush current limiters" but they
don't seem to protect against short disconnections.

2. ### Paul MathewsGuest

Energy storage in 20K uF at 140 V is 196 Joules (i.e., watt-sec),
which means that charging the cap is probably not going to blow a
thermal breaker rated at 30 amps or more (which is what you'll need
for your 2.8 kW load). You can't count on leakage inductance to limit
the current much, either, because the initial core state can include a
flux offset at the moment that you energize the circuit.
A common way to limit this type of inrush is to add a series power
resistor and a time delay relay. The relay contacts short out the
resistor after a delay. The relay can be electromechanical or SSR
type.
Paul Mathews

[...]

The good solution is a resistor + triac. The resistor limits the inrush
current to a reasonable value. After that the resistor gets shorted by a
triac.

The substation will survive; the diode bridge and the power switch will
burn out eventually.
I wouldn't rely on that.
True. The thermistor limiter is a simple solution for the low power.

VLV

4. ### john jardineGuest

A much bigger problem is the switch on transients of the big transformer
(occasional core saturation). Spice does not model this. Two switch on's out
of five maybe OK, the other 3 can (try to) draw many thousands of amps .
If it's a one off design, I'd be inclined to fit a small, current limit
series resistor in one of the supply leads, which is shorted out by a
relay, 1/4 second delayed (say) after power up. You then kill 2 birds with
one stone.
For production, a softly ramped up triac in place of the resistor, would be
better

5. ### Phil AllisonGuest

"John Devereux"

** See:

http://sound.westhost.com/project39.htm

....... Phil

6. ### JamieGuest

might be a good idea to use a time delay on set of contacts that would
bride a power resistor for a soft start.

7. ### Spehro PefhanyGuest

Yes (bridge) a power resistor. You should also turn the load during
this process to avoid the potential of welding the contacts. I guess
you could use a SPDT relay if there is no electronic way to turn the

Best regards,
Spehro Pefhany

8. ### John LarkinGuest

I've used the disc-type thermistor inrush limiters at the kilowatt
level, two units, one in series with each of the two primary windings
of a 120/240 power supply. They seemed to work fine, even when I
teased the power switch various ways; don't quite know why.

John

9. ### Paul Hovnanian P.E.Guest

That's what they do for power factor correction caps.