core saturation in variac

Hello Malcolm, Ed and all,

I am struggling with the idea of using variacs as current 
limiting reactors.

Firstly, a variac rated as, say, a 10A variac is one which 
can supply a _load_ current of 10A I would say. If you 
consider a variac to be an autotransformer, this 10A is 
made, if you like, of a 5A current flowing down the winding, 
and a 5A current flowing up the winding in the opposite 
sense. The sum of these two flows out of the tap into the 
load ( this assumes that the variac is set mid-way, for ease 
of discussion. If it was set at 25% then the current from 
the top would be lower and the current from the bottom to 
the tap would be proportionately higher).. remember the 
phase shift that occurs. The ampere-turn product for each 
half of the autotransformer is made to balance.

If you use a variac in series mode i.e. as a choke, you will 
try to drag, say, 10A through the whole of the winding. 
Unless you set the variac brush (tap point) close to the top 
end, the ampere-turns product will be very big and the core 
will saturate. I know that the magnetic path length is very 
long in a toroidal core, but even so the core will saturate 
if the current is pulled though all of the windings.

If you now cut a slot in the variac, and "gap" it, the core 
will not saturate until the ampere-turn product is much 
higher, but now the reactance of the variac has dropped to, 
say, a third of its previous ungapped value.

The point is, if you saturate the core, the current through 
the LV winding of your pole transformer will become 
decidedly non sinusoidal and the harmonic power that you end 
up creating as a result of the non-linear B-H curve of the 
core will be wasted. You can't charge the capacitors with 
the harmonic currents, in other words. Not to mention the 
increased EMI problems, Joule losses, iron losses in the 
pole transformers.

Today, I gapped all three of my variacs and I will remove 
the 250 turns of 22swg wire, and replace them with 100 turns 
approx. of 16swg (1.6mm diameter). I will thus end up with a 
compact reactor that won't saturate and will be good for 40A 
( I can then switch the three in series/parallel 
combinations to progressively wreck the regulation of my 
pole transformer).

The thing that bothers me is that my understanding of 
magnetic circuits is limited, and I know that lots of people 
have successfully use variac limiters. Where have I gone 
wrong?. Have I gone wrong? Theory and explanations would be 

As a final remark, I realise that the RWS approach avoids 
this by using an air-core and lots of turns, so this is my 
viable alternative. I might just go and buy a reel of 2.5mm 
squared single core cable and measure its inductance on the 
reel. This would be a nice lazy way of building up inductor 
banks. Also you don't add to the magnetising current that 
you already need for the pole transformer.


Richard Craven