The “Wrong Way” To Connect A Variac…..and why you might want to.

First off, to address confusion before I create it:  VARIAC is a trademark of General Radio and POWERSTAT is a trademark of Superior Electric.  For the most part, they are generically called “variac” by most people in much the same way that we know copying as “xerox” or soda as “coke.”  But they, and other similar devices of other makes, are all just a type of autotransformer with nearly infinite adjustability instead of fixed taps.

An auto-transformer is a nifty device.  Unlike a conventional transformer, it only has one winding.  This winding is constructed with taps at various points to allow accessing different numbers of turns in the coil….but in a variac, the construction is further enhanced by construction that allows a moveable brush contact to reach any turn in the coil with just a twist of the knob.  Similar in construction to a rheostat (a variable resistor) but instead of resistance, it is wound of typical transformer magnet wire and works with AC current only.  Transformers change the voltage by a change in the ratio of wire turns on the primary to that of the secondary.  For example, a 100 turn primary to a 10 turn secondary will result in 1/10 of the voltage at the output as from the input.  In the auto-transformer, however, there is only one winding and the ratio is achieved by selection of the points of contact in the overall winding.  (this also means the input and output are not isolated, more on this later)

The diagram above was photographed off of the data plate on one of my variacs.  As you can see, there are 4 fixed taps (#’s 1, 2, 4, and 5) and a moveable tap (#3)  In a typical application, they are wired to allow an input of, say 120V, and an output of 0-140V.  This is the normal, buck-boost application and it allows the user to compensate for high or low line voltage as well as apply lower voltages to items.  If you look closely at the diagram below, you’ll get the idea.

Let’s say, for sake of argument, that the winding contains 140 turns from #2 to #4 and each turn is worth 1 volt.  Let’s also say, that #1 is 20 turns from #2 and #5 is 20 turns from #4 (this makes it simple)   So, if we apply 120v to terminals #2 and #5 (which has 120 turns between them)  Then…terminals #2 and #1…as well as #5 and #4 will show 20v each.  (each being 20 turns from the ends)  #1 and #5 (having the left over 100 turns between them) will show 100v.  Just like the dataplate schematic says.  Now, terminal #3 is the moveable contact.  It has a brush and is connected to the control knob.  When you twist it, it glides across and makes contact with any point on the winding you choose.  If you take power out by way of terminals #2 and #3, you will find 0 volts when the contact is by point #2….passing by #1, you will have 20v…..when you move it to point #5 you will have the 120v input value at the output.  (at this tap, you have effectively connected the line input directly to the output)  As you have likely guessed, when that contact is moved over to point #4 you are now accessing MORE turns of the coil on the output side than you had at the input side.  At this point, in our example, you would be connected to 140 turns of wire.  The output would then be 140v.  So there you have it.  120 volts in….and 0-140 volts out.   Now, as you may have guessed–if you only wanted 0-120v output and had no desire to boost it above the input value, you could achieve this by moving the wire connection from #5 to #4….at this point, the other tap values would be different, but your output range would then be restricted to 0-120…..yep, these are versatile devices.

The big advantage of this, is that unlike a resistance based device, the heat dissipation is low.  Additionally, unlike electronic dimmer devices that use SCRs or Triacs, there is no distortion of the waveform.  The type of AC wave going in is the type that comes out.  Only the amplitude (voltage) is changed.  While these same types of result can be obtained with electronic devices, they are complex and expensive and not nearly as robust and reliable as a basic coil of wire that makes up the heart of a variac.  Variacs typically have found applications in radio gear, motor control, stage lighting, reforming of electrolytic capacitors, neon sign transformer matching, etc…They are also often used to drive saturable reactor chokes for control of larger current devices.  A typical variac is also rather tolerant of overloading when in the “buck” range (0 to input voltage range)….when operating into the “boost” range (input voltage to max output voltage range) they will still tolerate some overload but it is best not to exceed the ratings too much.  Keeping this in mind will insure your variac lives a long, cool, life.   And a comment about safety:  As I mentioned, the input and output–being on the same winding, are NOT isolated.  This means full current can be available on the output.  Usually this is what you want, but if it isn’t…or if a failure were to result somewhere, it could create a hazard.  If isolation between line input and your output is a requirement then you will want to add an isolation transformer to perform that function.  In either case, make sure to include proper fuse protection in the circuit.

 

Now…the “wrong way” to wire one…….and why you might want to.

Look at this diagram:

In this diagram, the unit is not even wired as a transformer.  It is more akin to the way you wire a rheostat.  What good is that?  Well.  Here’s the deal.  If you wire one in this fashion and apply 120v to the input, no matter where you set it, it will show 120v at the output–so long as there is NO load on it.  In this configuration, it has essentially become a variable inductance device.  Inductance is to AC current what resistance is to DC.  It basically becomes a big variable choke coil.  If you connect an inductive load to it, like a small fan or other transformer, you can effectively control its speed or output this way.  You are doing so by varying the available current to the device (this also effectively varies the voltage too)  It works similarly to the sliding core pull chokes used in neon bombarder circuits but it is much more pleasant than fighting a pull choke while running a 10kva bombarder.  I use this set up on my neon processing bench.  It works.  Arguably, a saturable reactor would be smoother–and offer more in terms of flexibility of installation layout (my BIG variac has to be where I can reach the knob) but those are small issues and I already had the monster sized variac on hand.  (hooray for parts boxes the size of a house!)

Now, this also works on a small scale too.  I once had a small neon flower sculpture (“planted” in a pot, of course)  It was too small for the conventional coil and core 2000v, 20ma neon transformers that were the smallest made at the time.  When run, it would get hot–heat kills things.  So, I wanted to dim it down a bit.  I discovered, through checking with my small variac on the bench, that I wanted to shave about 10 to 15 volts off of the input and this would subsequently lower the output voltage (and resultant current flow) by a value that allowed the tube to run cool.  So, I dug in the parts box and found a small 120v primary, 12.6vct tube filament transformer.  I taped off the 120v connections so as to prevent shorting anywhere (namely my fingers) and then connected the 12.6v side in SERIES with the 120v primary of the neon transformer.  This has the same effect as the “wrong way” to wire a variac and it did a great job of resolving the problem I had in this sculpture.  And, of course, it all fit into the flower pot quite nicely.

With modern electronic neon transformers, most applications will be dimmable within the specs of the transformer, but for conventional magnetic units, this is still a viable and easy way to achieve that result.  The variac and some good meters on the bench can help you sort out what you want to use.