5
\$\begingroup\$

I have been looking for ways to limit inrush current in a power supply circuit I am designing. In LTspice, with no current limiting, I see a spike of around 24A before my caps are charged.

Looking around, I see NTC inrush current limiters as a common solution as well as some ideas with inductors and resistors inline with the capacitor, each with its own advantages and disadvantages.

My idea though was to put a resistor before the bridge rectifier in my circuit, so the resistor would always provide some resistance before the capacitors got charged. Playing around again in LTspice, this drops my inrush current to about 4A (and changes with the resistor value obviously).

Is this ever done? Are there any major good or bad reasons to do this? I suspect no one does this due to decreased power efficiency, but is that the only reason?

\$\endgroup\$
5
  • \$\begingroup\$ en.wikipedia.org/wiki/Inrush_current_limiter This article may be of some small help. \$\endgroup\$ Oct 27, 2015 at 1:41
  • \$\begingroup\$ In common cases, the inrush current isn't severe enough to warrant mitigation. As long as you're not using oversized capacitors, your components will handle the inrush well enough. \$\endgroup\$ Oct 27, 2015 at 5:28
  • \$\begingroup\$ How does your simulation look like, does it contain a realistic enough model of a transformer (I am assuming you recitify some transformer output)? If you just rectify after some AC voltage source, then you get unrealistic inrush values. In a lot of cases the DC resistance and the voltage drop due to excessive current is enough to reasonably limit the inrush into your smoothing caps. \$\endgroup\$
    – PlasmaHH
    Oct 27, 2015 at 9:07
  • \$\begingroup\$ What is the Issc (single cycle surge current) rating of your rectifier? Studying the datasheet may show it's already over 24A. If not, upgrading the rectifier may be the most cost effective (and efficient) solution. \$\endgroup\$ Oct 27, 2015 at 11:49
  • \$\begingroup\$ An important point that I've not noticed being made is that an iron core transforme feeding a capacitor will produce very short large I spikes as Vmains rises above Vcap and this can cause bad EMI . Adding even quite a small series resistor spreads and rounds the peaks. \$\endgroup\$
    – Russell McMahon
    Oct 27, 2015 at 13:56

5 Answers 5

6
\$\begingroup\$

There are good and complicated solutions, relay, FET, and simpler, thermistor, but it still may be possible to use the simplest, fixed resistor, without too much dissipation for your application.

If you get an inrush, does it matter? Well, only if it breaks something. So what could break? Supply fuse, transformer, or diode rectifiers.

Fuses have 'T' rated versions that take a long time to blow, for just this application. The transformer is a heavy lump of copper, that's not going to fail.

Read the specification of your diodes carefully. You may be surprised at how much the 'single cycle surge current' is. In a 1N40xx (cheapo workhorse mains diode), the continuous current is 1A, the surge is 30A. For 1N54xx series, the figures are 3A and 200A. This is specifically to allow them to survive inrush. You may find you already have enough stray resistance in your circuit to limit the current to the safe surge value. If not, you maybe won't need much more, and it still may give you acceptable efficiency. If it doesn't, then try smarter solutions.

\$\endgroup\$
2
  • \$\begingroup\$ And can get very "smart" and very expensive: electronicdesign.com/power/… \$\endgroup\$
    – Fizz
    Oct 27, 2015 at 6:19
  • \$\begingroup\$ Add to list of "what could break?": anything on the same AC phase that wants a sinusoidal input. Inrush current can be responsible for ugly harmonics. \$\endgroup\$
    – Ben Voigt
    Oct 27, 2015 at 13:05
3
\$\begingroup\$

Using a resistive element to limit inrush current is done.

However, placing a permanent resistor in the AC (or DC ) lines is very inefficient as it is always in-circuit.

The use of a Negative temperature coefficient (NTC) thermistors is often done instead as it provides a level of resistance during the precharge event that is then reduced due to heating. This however does impose a "cooldown" period to facilitate the NTC returning to room temperature resistance in the event of a power loss or power cycle.

An established method is to use a form of soft-start

http://www.ti.com/lit/an/slva156/slva156.pdf

schematic

simulate this circuit – Schematic created using CircuitLab

There is equally monolithic chips that provide the same functionality.

http://www.onsemi.com/pub_link/Collateral/NCP330-D.PDF

\$\endgroup\$
1
  • \$\begingroup\$ Those things you link to are normally used to protect voltage regulators, not rectifiers. \$\endgroup\$
    – Fizz
    Oct 27, 2015 at 6:06
2
\$\begingroup\$

Adding resistors will create a permanent power dissipation. This is usually undesired. It will also create fluctuating input voltages caused by changes in current draw (voltage drop at resistor).

\$\endgroup\$
0
\$\begingroup\$

A quite simple solution (I'm surprised it's not mentioned already) is to add a small coil after the rectifier:

schematic

simulate this circuit – Schematic created using CircuitLab

A coil doesn't add much active resistance and doesn't practically affect the efficiency under "steady state" operation. The current spike is suppressed by the inductance which limits the current rising slope by $$L=\frac{dI}{dt}$$

\$\endgroup\$
4
  • 1
    \$\begingroup\$ "is to add a small self" - a small self what? \$\endgroup\$
    – abligh
    Oct 27, 2015 at 8:48
  • \$\begingroup\$ Sorry, I mixed English and French. I meant to say "coil" or "self-inductance" \$\endgroup\$ Oct 27, 2015 at 8:58
  • \$\begingroup\$ Merci beaucoup! \$\endgroup\$
    – abligh
    Oct 27, 2015 at 9:05
  • \$\begingroup\$ The coil is good .It also helps distortion factor to some extent.Beware of voltage spike at turnon. \$\endgroup\$
    – Autistic
    Oct 27, 2015 at 9:37
0
\$\begingroup\$

There are fancier ways of limiting inrush current than the simple resistor ,Some of which JonRB has stated. If you are using a resistor there are a few more things to watch out for .Firstly the surge power dissipation of the resistor can be incredible .I have seen Christchurch design engineers blow up such resistors .Its really important to get specifications for peak power dissipation .Melf are good that way but some manufacturers are vauge.Secondly the average power dissipation is dependant on the RMS current not the average current .RMS current can be and often is significantly more than the average .I have seen people unwittingly undersize these resistors . The horrible part about this is that the resistors take some time to fail open circuit .The way the resistor fails is also important .Some resistors will go on fire which is bad .Some are flame retardant and go out ,some dont .If you still want to use a resistor test it under fault conditions .

\$\endgroup\$
2
  • \$\begingroup\$ Is "Christchurch design" a pejorative term I'm not aware of? \$\endgroup\$ Oct 27, 2015 at 9:06
  • \$\begingroup\$ Oh dear I didnt know what Pejorative meant,I suppose that RETARD would be a Pejorative for my screenname which is not entirely innacurate. I am sure that CHCH doesnt have a monopoly in blowing up current limiting resistors on the input of bridges.CHCH punches above its weight as far as electrical engineering talent BUT sometimes they cant see the wood for the trees which is good in a way because it keeps me in work. \$\endgroup\$
    – Autistic
    Oct 27, 2015 at 9:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.