Power supply and circuit breaker

[spletcher1]

Hi All,
I just finished building my first CNC machine. It seems to be working well, but I seem to have a problem with my power supply. I have put together a standard linear power supply using a large toroid transformer (about 45V and 18A), two large capacitors, etc. When I power it up, my 20A circuit breaker (in the mains box) kicks out about 50-60% of the time with 1 sec. If I get past 1 sec, everything works fine and there is no problem. I am not an expert in this area, but I am rather sure everything is wired correctly. I am thinking that the power draw during the charging of the caps is what is doing it. Does this sound reasonable. If so, how might I add some inductance to slow the power draw early on? I am happy to provide more details on the individual components and the design of the supply if needed.

Thanks!
Scott

[Jesse B]

I'm not sure if this is a legitimate reason, but I feel like it's possible that power supply has an amperage surge during power-up? I don't know, just an idea. It's close enough to the breakers limit that it seems possible.

[bharbour]

If your power supply is not tripping the breaker every time you switch it on, there is a good chance that is is wired correctly. You are correct that there is a large current draw while the caps charge up. If you disconnect the load from the power supply and it still trips the breaker, that is a really good indication that it is the capacitor charging current. Keep in mind that with no load, those caps are going to stay charged for a long time, so you should discharge between tests through a lightbulb or something similar that will deal with the current in a controlled way.

When the power supply does come on successfully, are the DC output voltages measuring what you are expecting for the transformer? If the primary were jumpered incorrectly, you could be getting twice the output voltage you are expecting, along with twice the inrush current. That could make for some major drama with the electrolytic caps in time.

As for ways to solve it, I have seen designs that power up in a two step process. The first part is done with a series resistance in the AC line to limit the current while the caps charge up, then bypass the resistance. That can be done automatically with a part called a Negative Temperature Coefficient Resistor (NTC) or a fixed resistor with a relay that switches in to bypass the resistor when the caps charge. You could do it manually with two switches, the first to apply power to the transformer through a power resistor and a second switch to bypass the resistor.

Good Luck,
BobH

[vger]

Scott,

Yep, most likely the caps initial charge current. When starting cold (no charge on the caps) they initially present a direct short to the power supply. This can also be pretty hard on rectifiers as the current can overheat them and burn them out.
One method I have seen to get around this initial surge involves a resistor and a relay. A resistor is placed in the circuit between the recitfiers and the capacitors to limit the current. For your supply I would say maybe 50 ohm 50 watt would be safe. You could probably use a lower wattage as the power through the resistor will only be there for a short time. Then the relay... The relay should have a 48 volt coil with the coil connected across the capacitors. The contacts of the relay that close when energized are connected across the resistor.
This way the caps charge through the resistor initially and when charged to about 30 volts or so, the relay closes and the contacts make a direct connection between the rectifier and the capacitors.
Additionally, there is what is called "Transformer inrush current" and torroids are nutorious for high inrush currents. With some transformers you will actually hear a Thump sound when initially turned on.

Hope I've helped
Steve

[spletcher1]

Thanks Steve and BobH! Your replies were very helpful and informative. It is also reassuring that I have some idea what is going on! I double-checked the connections on the transformer, and I believe that they are correct.

Steve, you are right, that toriod transformer is a beast and it often "clunks" when I turn it on.

I wonder whether is it the transformer that is causing the problem or the caps? I guess trial and error will tell me. If it is the caps, then Steve, your solution of a relay across the caps should work. If it is the transformer "inrush", then I would implement BobH's resistor and switch on the mains side of the transformer. Does that sound correct?

Thanks again for your help.

Scott

[Al_The_Man]

The inrush to the transformer itself should not be a problem to a normal breaker as there is a time delay where the inrush will ride through. Normally 1 cycle max.
If you disconnect the caps from the bridge, it will show whether the Toroid is the main problem.
I have used Toroids extensively and so far have not yet had to resort to any kind of inrush protection, I would check whether your breaker is a normal time delay type.
Al.

[bharbour]

Originally Posted by spletcher1 I wonder whether is it the transformer that is causing the problem or the caps? I guess trial and error will tell me. If it is the caps, then Steve, your solution of a relay across the caps should work. If it is the transformer "inrush", then I would implement BobH's resistor and switch on the mains side of the transformer. Does that sound correct?

Hi Scott,
Assuming that there are no defective parts, it doesn't really make much difference which side of the transformer you put the relay on. The transformer should not have enough inrush current to pop a breaker by itself. I prefer putting the relay on the mains side of the transformer for two reasons:
1) The current is lower on the mains side which will make the relay last longer.
2) DC has arcing problems with relays if you are switching under a load. AC will extinguish an arc as the voltage passes through 0. The circuit both Steve and I have described should not be switching when heavy current is flowing, so this is more of a general practice thing than an issue in this circuit.

Good Luck,

BobH