I got a big transformer from a big UPS and replaced the secondary with a single loop of heavy gauge wire. My problem is that whenever I plug it in it trips the breaker.
I guess the problem is too much inrush current. What can I do to reduce it?
Need more info. How many watts is the “big transformer”? What size breaker?
You can use inrush current limiters (that’s what they are for), but if you’ve modified the transformer, you may have changed the impedance enough that it is now drawing a lot of current, even though the secondary is unloaded.
ETA: Inrush current limiters at DigiKey.
What do you mean by “big UPS”? The kind one might use for their home PC, or a heavy-duty industrial grade UPS for a server rack? If the former, I wonder if the transformer got damaged (short in the primary; are you also sure it was good before you removed it, if it was from a discarded/trash pick UPS?) when you removed the secondary; otherwise, modifying only the secondary shouldn’t have any effect on the primary impedance. If the latter, than an inrush limiter is needed, as **beowulff **said (such a big transformer should also have a PFC capacitor across the primary if the circuit can barely supply the reactive current that otherwise flows; the easiest way to find the optimal value is to measure the unloaded current while adding capacitors).
The transformer is rated for 230V, 1000 to 1500 watts.
I am trying to build something like this: DIY Spot Welding Machine - YouTube
My transformer though is not like the core form shown in the video, it is a shell form.
Nothing fancy, just a cheap 2000VA dual battery 24V UPS.
It was working all right, just the batteries had gone bad. I had it forever in my junk bin and there was a time I needed some FETs and scrounged them from the board. I don’t know where the board is now, all I have is the metal shell and the transformer.
Nice little company in Carson City. They will send you FREE samples! Got one to fix my pinball machine.
Found some thermistors. I will test it later.
The thermistors looked wimpy to me but I tried them anyway. Turns out they were rated for a fraction of a Watt, and they blew right away. (I had two in parallel).
I’ve found some proper thermistors for the application in Digikey, but they charge something like $20 for shipping, so I don’t feel like paying that for a part that costs $1.
So I tried another thing. I put in series a 100W incandescent lamp. The current was 0.4A at 220V. But I measured zero volts on the secondary. It didn’t register anything even in the mV scale. How is this possible?
Parallel doesn’t sound right. I’d have to check the data sheet for the part for the proper circuit.
I’d expect there would be something. But lets think about this. Voltage is dropped across each of those lamps. If there was a switch parallel to each lamp and you closed the switches one at a time to by pass the lamps, then at the end all the voltage would be across the coil.
Sounds like a funky set-up.
I’d keep looking for a better circuit or product.
Sounds like your primary is shorted - did the lamp light up brightly? It shouldn’t have.
Negative temperature coefficient (NTC) thermistors have some serious limitations as inrush current limiters: they require a significant cooling down period between on/off/on cycles (maybe 10s of seconds for the bigger beasts), they need to run hot to pass the full operating current, and they’ll cause a power drop even in the hot low-resistance state.
Electric welders tend to switch across high surge capability resistors to limit the inrush current. Here’s a useful paper:
http://www.welwyn-tt.com/pdf/application_notes/Welding-Power-Supplies.pdf
You’ll need a meaty set of switch contacts for this. Look out for “contactors” rather than “relays”.
The lamp was at near full brightness
The best way is with a NTC thermistor and a relay.
Place the thermistor in series with the input.
Place the relay coil across an appropriate transformer winding. A secondary winding is preferable, but the primary will still work. Relay NO contacts go across the thermistor.
The thermistor limits inrush current, and goes to fairly low resistance when it heats. This allows the relay to pull in, which shorts out the thermistor, allowing it to cool and thus be ready for the next turn-on event.
If you can’t come up with an appropriately rated NTC, Then you can just use a fixed resistor, but you may have to fuss a bit to get the value right. It may be helpful to use a second set of relay contacts to remove the load during the soft-start time. With the load connected it is easy for the resistor to cause so much drop that the relay coil never sees enough voltage to pull-in. A time delay relay is another way to do this.
Transformer inrush is one of those things that is under appreciated. It bit my ass the first time I encountered it. Now when I design to account for it some engineers still think I am on crack.
For those not familiar with transformer inrush:
Transformers rely in inductance to limit input current. The iron core of a transformer only exhibits magnetic properties of iron up to a certain field strength. (saturation) Beyond on this it behaves similar to air. This results in a huge drop in inductance, and a huge increase in current.
Iron is heavy, so transformers are designed to have just enough iron to avoid saturation in normal operation.
Iron also exhibits a strong magnetic hysteresis. It tends to stay magnetized in whatever direction the last coil polarity established. (this was exploitetd as memory in early computers, some of us grey beards still refer to RAM as “core” memory) If the transformer is switched off when the magnetic field is at it’s peak, and then switched on during the same polarity of the AC waveform, then the iron will saturate almost instantly due to the iron’s memory. This will result in a huge current spike that will trip magnetic circuit breakers. Thermal breakers and slow-blow fuses will usually ride through the spike.
The spike only lasts a maximum of 1/2 AC cycle. The next half cycle resets the core, and it then operates normally.
Torroidal transformers are the worst for this IME, as they lack the air gap of E-I cored units, but I have still seen E-I transformers pop breakers due to inrush.
I once worked on a large, industrial electrical cabinet that contained a bunch of power supplies, and the problem of excessive inrush current was mitigated using the switched resistor technique. It used a bank of open-air power resistors and a heavy-duty contactor. At start up, all the current went through the resistors. After a few seconds, the contactor shorted-out the resistors, and thus all the current went through the contactor’s contacts.