Large transformers have a huge current demand when they are initially turned on. This is because, until the magnetic field and inductive resistance builds, they are essentially short circuits.   For example, you may have turned on some large tool or appliance and heard an initial large "HUMMMMMMM".  That is the transformer say "Ow".  The circuit breaker for that outlet might also go "Whoa, what are you doing!"

The transformer above (Avel Y236907 800VA 45V+45V Toroidal Transformer), for example, will try to draw over 100 Amps on the first cycle of 60 Hz Power.

To keep a large transformer from being damaged at turn-on (and to keep it from saying "ow"), or to keep a breaker from popping, you put in an inrush current limiter circuit.  This Instructable will detail how to do that. 

Step 1:

The thing we are after is a way to limit the current initially at turn on, and then to not limit current afterwards at all.

The circuit I use contains a thermistor, a relay, and some resistors, a capacitor and a couple transistors.

Step 2:

A thermistor of the type used here is a temperature sensitive electronic device that at low temperature, presents a resistance to a  circuit.  As temperatures rise, it gradually becomes closer to a short circuit.

This particular one here is the Amtherm MS35 20010 and can be bought from Digikey (570-1026-ND).  Its nominal resistance is 20 ohms, and as temperature rises, it drops to very low resistance.  

Now you might say  "Great, just put this baby in series with my AC input line and I am done".  In some cases you might be able to do this but the power amplifier I am designing is a 200W per channel set up and, though the pyrotechnics might be fun to some, being nominally averse to smoke and crispy electronic components, as well as slightly paranoid, I decided to add a relay to the circuit.

In addition, when someone turns off a device that has been on for a while, the thermistor will be very hot.  If they then turn on the device immediately after shutting it off, the thermistor will dutifully say "I'm hot and tired, and I am not providing you any resistance".  Your transformer will say "Ow" once again.

Step 3:

Here you see the thermistor in parallel with the relay (minus the relay control circuitry). The relay I used is the Omron G5Q-1A4 DC12, available from Digikey (Z223-ND).

When power turns on, all current flows through the thermistor.  But if we put a timing circuit in that closes the relay a few seconds after turn-on, the thermistor would do its job limiting current initially.  But the relay would then take it out of the circuit by by providing an essentially zero resistance path through itself, and around the thermistor.  The thermistor will then say "Thank You", cool down or just stay cool,and thus be ready for the next power turn on.

So how do we create that timed delay?

Step 4:

Here you see the timing circuit.  The relay will be driven by transistors Q1 and Q2.  These transistors are configured in what is called a "Darlington pair" to be a simple transistor switch.  When they are off, the relay cannot turn on because no current flows.  When they are on, the relay closes.

Two transistors were used so that a very small amount of current at the base of Q1 would be guaranteed to drive transistor Q2 into saturation, switching it on.  A side benefit of having 2 transistors is that the normal .65VDC from base to emitter of the transistors becomes 1.3VDC now that there are two transistor base to emitter junctions in the circuit.

For my purposes, I have 12 volts DC available in my amplifier, sourced through a simple regulator circuit supplied from an independent low power transformer.

Resistor R1 and  Capacitor C1 provide the timing delay. When power is turned on, and the 12 VDC comes up, current flows through R1 to charge capacitor C1. Over time, the voltage on the capacitor reaches the 1.3V or so needed for transistor Q1 and Q2 to turn on.

When power is turned off, capacitor C1 drains to zero through R2 so the circuit is ready for the next turn on.  Care must be taken calculating the value of R2 - the voltage divider of R1 and R2 must always allow for greater than 1.3VDC or else the transistors will not turn on.

The DC voltage driving the timing circuit can be any voltage, including a DC source generated from the output of the power transformer that we are controlling the inrush current to. For example, if the output of the transformer was 36VAC, you could construct a simple zener diode DC source to drive the relay and a timing circuit with different component values. You might want to simplify your life by going to a 24VDC coil relay. 

Step 5:

Here is the complete circuit.  Once power is turned on, the 12 VDC supply enables the timing circuit to act, eventually turning on transistor Q2, enabling current to flow from the 12VDC source through the relay and down through the transistor to ground, shorting out the thermistor.  (The diode on the relay is there simply to prevent large voltages that can happen when a relay drops out from damaging any devices).

The delay this circuit injects is 2-5 seconds, depending heavily on the value of C1.  Electrolytic capacitors have very wide tolerances (such as +10% and minus 40%) so they are not useful for precise timing circuits but certainly useful here.  You should always just experiment with components to insure you get the right delay.

Hopefully this helps your large transformer have a happy day!
<p>Ive been checking my broken chargers and most have NTC connect in series with the negative and then a 471kd07 Varistor beetwen pos and neg. Isnt that enougt? The NTC con will work if when temp is high (all currents) and Varistor will work when current is high and NTC is Hot and lower the resitansa is current is high? Only 2 component ans seems to be standard in most chargers.</p>
<p>Is there ANY commersial products i can buy instead of messing around with high voltage and current? Trying to run a spot welder but i trip all breakers under 20A.</p>
This is a really cool and ingenious instructable, but does it have to use a resistive load + relay?<br>could it not be that you just wait till the zero crossing of the AC line, then switch it full on, considering that because it's an AC wave, it will ramp up the voltage in a sine wave pattern from 0? Or does the ramp have to be relatively slow?<br>-Thanks. Most transformers i work with are air core resonant.
I think the problem there is it may take the big coil longer than the cycle time to build up to its full magnetic flux. So for that first fraction of a second you're good, but what about the next wave? Inductance is reluctance. A coil this size has big feet to drag.<br><br>Now I'm wondering if one couldn't build a circuit of cascading chokes to start a big coil. Kind of like fighting fire with fire? Or would that just aggravate the problem? Inductance is one of my weaker points of understanding. I just never seem to get the stuff.
Maybe a motor starter
If you switch a transformer on at the zero crossing the current is theoretically infinite! Switching at top or bottom and current is zero I learned at school. So do not use SSRs or SCRs with zero-crossing circuitry on transformers ever!
As I understand it, if you switch at the first zero crossing, that is the time that it will draw the 100+ amps - on that first cycle.<br><br>Most of the books I have read said to give the transformer limited inrush current for 1-2 seconds. I may have gone overboard with 2-5. <br>
This seems very useful. Can this method be used with motors? Forgive my ignorance.
Forgive me as I must plead ignorance as well - I do not know much about motors. <br><br> I googled &quot;motor inrush current&quot; and got many sites that seemed to have thoughts on it though.
Haha! What a coincidence!<br><br>I ask that because some time ago I bought a generator, and the vendor said that for using it to feed the fridge and/or the freezer, I need 7 times more power than the nominal intake of the motors, due to initial pulse. I think the case is like that you show, of tranformers. <br>
I think the motors need that current just to overcome their own inertia. Inductance motors have start windings in them that draw more current just to get them starting to spin. I think there is also a phase shift involved with the start windings that aids in motors beginning to revolve, but I could be wrong about that.<br><br>Really, it is amazing electric motors run at all. Motors and transformers have coils in common, but after that the similarities end.<br><br>In order to reduce starting current with electric motors I think they have to be designed differently before they're built. I've heard of large industrial motors with special coil arrangements to bring them up to speed and limit the current they need in order to start.<br><br>To sum up, your vendor was right, and I do not know about anything you can do to reduce the current draw of an electric motor from what it is. Well, maybe if you hooked up a pull starter off a lawnmower ... it is the initial pulse that is the problem, once motors get moving they go within their rating.
Thanks, Fred!<br><br>I asked advice from a manufacturer of generators, and they said that maybe I could use a small generator if I change my refrigerator and my freezer for a new ones, because currently there are not made with the motor fixed to the compressor but through a centrifugal or electric clutch to prevent the inrush pulse is so high. Clever, isn't?
&quot;Ow&quot;, you confuse thyristor and thermistor.<br>You&acute;re using a thermistor or NTC.
Argghhh ... never do these things at 7AM. Yes to rammstein and unmitigatedaudacity .... it is a thermistor, not a thyristor (if you go to Ametherm's site, it is quite plainly stated) . I will edit and fix. Thanks.
are you sure its a thyristor?? :?<br>thyristors usually have three terminals.<br>The &quot;resistance varying with temperature&quot; fits a thermistor better.

About This Instructable




More by saratogadude:Controlling the Inrush Current Required by Large Transformers Generating Gcode for ShopBot from AutoCAD file using Vectric Cut 2D Make a 1-2 uH Output Air Core Inductor for Audio Amplifier 
Add instructable to: