```
Original Poster: "J. B. Weazle McCreath" <weazle-at-hurontel.on.ca>
> Original Poster: "Michael Novak" <Acmnovak-at-execpc-dot-com>
> >     I can't afford to buy a variac but I need some way to vary the line
> >voltage... So, I decided to build one. I'm pretty sure I can get my hands on
> >some 10 guage magnet wire, however, I still need to find a suitable core.
> Isn't
> >a variac basically a bifliar-wound 1:1 turns ratio transformer? If so, does

No, it does not have two separate windings. Just one winding.

You connect input phase (L) and neutral (N) to ends to winding, and
take output with N (same as input) and brush. In order to get
other ratios (output:input) than 0-1:1, it is possible to add
windings past L to get higher output voltages.

> >anyone have some sort of formulas on the subject?

Go to your nearest library and find a book on transformer design.
For example, Smith: magnetic components, design and applications.
Besides basic theory, good books (like the one above) have practical
information on actually winding the coils, insulation etc..

Anyway, the basics:

Volts/turn = 2.865 B f A / 1000
Where
B = Flux density in kiloGauss (something like 12-15)
f = frequency in Hertz (50/60)
A = core cross-section area in square-inch

Power handling capability is _roughly_
A = sqrt(P)
Where
A = core cross-section area in square-centimeter (1 inch = 2.54 cm)
P = power in Watts.

> > What would be the
> >disadvantages of having a core which is not laminated such as the kind
> > found in transformers?

HUGE losses. Does not work.

> >Any thoughts?
> > -Michael
>
> I think you'd have a pretty hard time home brewing a variac if for no
> other reason that finding the circular core material.  However, there

There is NO reason to use a circular core. In fact, I'd say that
is is MUCH easier to build a linear variac. Just wind the coil on
one leg of core and add linearly moving carbon. This is how 3-phase
variacs are built - just three linear coils on three legs of EI core,
and I have seen (commercial) linear 1-phase variacs as well (in old
magnetic particle testing devices).

Cores are readily available and quite cheap, so just buy a new one (or
laminations to create a long UI core of your own dimensions), and wind
the coil on one leg, epoxy it in place, grind flat for carbon brush, and
add carbon on some linear-motion fixture.

In fact, at the same effort wind primary at other leg and separate
secondary (with slider) at other leg to get isolation as a bonus.

Idea:
The most brave would add a magnetic bypass for primary to make the
thing current-limited - ie. something like this:

SSSSSSSSSSSSSSS
XXXXXXXXXXXXXXXXXXX
X SSSSSSSSSSSSSSS X
X                 X
X PPPPPPPPPPPPPPP X
XXXXXXXXXXXXXXXXXXX
Z PPPPPPPPPPPPPPP Z
Z                 Z
ZZZZZZZZZZZZZZZZZZZ

Where S secondary, P primary, X and Z core. Now, the XX core is
the normal case with isolated variac. By adding Z one will add
a magnetic shunt and thus current limitation like in a neon - just
a little bit different core config. Neon-like config would be:

SSSSSSSSSSSSSSS
XXXXXXXXXXXXXXXXXXX
X SSSSSSSSSSSSSSS X
X                 X
XZZZZZZZZZZZZZZZZZX
X                 X
X PPPPPPPPPPPPPPP X
XXXXXXXXXXXXXXXXXXX
PPPPPPPPPPPPPPP

It is as well nice that as Z is "separate", one can easily alter
(or remove) it to get desired current limit. One could use a standard
3-phase EI core for this, and just leave off laminations from the Z
leg. Now, I leave it as an excercise to calculate the flux densities
for primary and secondary (and volts/turn:s as they change), taking
into consideration how much flux goes through Z leg, to get proper
output voltage. :)

Just an idea.. This _would_ work, right?

Kristian Ukkonen.

```