I'm doing something wrong... 120 Amps??

[odgonza]

I'm doing something wrong... and I dont knot what.

I'm designing (only simulations) a Power Supply. I have a small problem with the initial current peaks.
My circuit have a 35 volts input and a load of 6 Amps. I got 220 Amps peaks on the bridge diodes... Ihave read that the capacitors works like a short circuit what then are discharged, and this is the problem. I have 4x4700uF so... the peak ia very high.

I have tried a soft start circuit, eliminated 2 of the capacitors and I reduced the peak to 120 Amp. Afther this first pek, the following peaks are about 60-40 Amps. I thing this is too much if I only have a load of 6 Amps. When I eliminate the capácitors.

I suppose this is a tipical beginner question and I also suppose that it is explaines in some books... but I haven't read them

Is is normal? What I'm doing wrong?

I send two circuits with the results of the simulations (very similar circuits), and a Soft Start circuit that I took as a reference.

[darko31]

Don't worry, capacitors draw a lot of current while charging but for a very short period of time.

You won't have any problems while trying out on a real circuit.

[hammil]

Although, as an engineering tip, putting a high current through capacitors can shorten their life, so if you can, put multiple capacitors in parallel to share the current.

[odgonza]

It's interesting the idea of connecting the capacitors in parallel to share the current... I like it.

What about the repetitive peaks of 50 Amps? Is is normal? It's a lot of current. I know that it's only during a short period of time but... 50 Amps?

The rectifier bridge can deal with up to 220 or 240 Amps, but... I don't know if the electrical circuit of my house can support it

[c4757p]

Although, as an engineering tip, putting a high current through capacitors can shorten their life, so if you can, put multiple capacitors in parallel to share the current.

First figure out whose I/V characteristics are mostly responsible for the current, because if ESR is the dominant factor in determining the current, you'll just increase the peak and each cap will still have the same amount. If it is truly an issue you can add a tiny series resistor. It's almost never an issue.

The rectifier bridge can deal with up to 220 or 240 Amps, but... I don't know if the electrical circuit of my house can support it

Peak? Sure it can. You can easily get kiloamp peaks out of your household wiring. The inductance of the wiring will give a bit of a limiting effect, as well. You might want to add some filtering, though, because those fast high-current spikes can cause some nasty RFI.

If you really need high current with high efficiency and no nastiness, go for a pre-built SMPS with active PFC.

[mzzj]

Series inductance and series resistance of the transformer windings also limit the current somewhat.
If your transformer is something like 500VA the transformer winding resistance is probably roughly 0.1-0.2 ohms seen from the secondary side.

[c4757p]

Try adding realistic impedances to the circuit. Look up ESR for the capacitors you will be using, and per mzzj, add about 100-200mOhm to the supply (just use the "Series Resistance" field).

[mamalala]

Also, it's always a good idea to select "real" parts in LTSpice. What you have now are basically ideal parts which don't exist in reality. To do so, right-click on each part and then click on "Pick new ...." in the dialog. Then select a suitable part from the list that will be shown.

For example, you have no specs on the diodes, the caps have the symbol for ceramics, etc. Picking suitable parts means that a bunch of parameters are then used in the simulation that are otherwise set to defaults (or maybe even disabled).

As has been mentioned here already, you should also set a source impedance for the AC voltage source, since there are not transformers with 0 Ohms source impedance/resistance.

Greetings,

Chris

[Smokey]

Not that you necessarily need to worry about it, but when those cap banks initially charge on power on it's called inrush and can be a real issue.  Imagine after a power outage when the power comes back on and a whole building of equipment all get huge inrush spikes at once.  Now multiply that by all the other shops on that same line.  That can equal a lot of current really fast. 

There are a bunch of ways to limit inrush.  One way is to wait for a zero crossing of the AC input before turning on.  There are relays that do this for you.  This way the caps will charge as the AC ramps up instead of at the peak when the voltage is already high and able to supply a lot of current.  Another way to limit inrush is using a NTC (negative thermal coefficient) device in series.  They usually look like big black disks.  The way they work is that they start out high resistance, and as they heat up the resistance goes down to some really low value.  This way when you turn on your device the high resistance limits the inrush and it quickly heats up to the point where the resistance doesn't really effect the circuit.  Just watch out for these guys since they remain pretty hot depending on how much current is going through them.  Another thing to watch out for with NTC resistors is that they need time to cool down before they can limit inrush again.

[qno]

Insert a 1 ohm resistor in series with the transformer to simulate the real impedance of the transformer.

[odgonza]

mmmm....a lot of god sugestions.... thanks.
Thanks a lot!. Really :-)

[odgonza]

I have used a "soft start" circuit to reduce the inrush (only simulated).
It is not my idea. It's taken from the Internet. The idea is to make the current go across some series resistors during the first second and then change to another way withour resistors. It's not too bad I think.