Re: DANGER! Strange transformer problem

Tesla List wrote:
> Original Poster: "Reinhard Walter Buchner" <rw.buchner-at-verbund-dot-net>

> If I read it correctly you have the pig, a choke and the variac all
> connected in SERIES. Is this true? If so, and if the variac is
> unmodified (slotted), this won´t work.

That is the correct connection.  

> Think of how a variac is normally hooked up: The top and bottom
> of the main winding are connected across the mains. Manufacturers
> want to invest the least amount of money possible. The inductance
> of the variac has got to (just) be high enough, so that the core does
> not saturate at mains input voltage. This is the nominal, no load
> current flow, ranging from a few 100mA to a few amps, depending
> on variac size, actually voltage applied, etc. Normally, you apply the
> load from the wiper to N, so the TOTAL core, variac wire, etc
> (i.e: resulting flux strength) never sees the complete load, you are
> presenting (which is one of the reasons you can overload variacs).
> The way you hooked it up, your variac´s core ALWAYS sees the
> full current flow (i.e: the full magnetizing current flows through the
> winding, core, etc). Try this: measure the no load current draw
> of the variac itself with the variac connected normally across the
> mains. Compare this value with the current value (variac connected
> your way) where your variac starts to growl. I´ll bet they are similar
> (with some leaway built in due to manufacturer´s "overbuild").
> The only way you can use the variac as a variable inductor (and not
> as a variable transformer) is to slot the core. That way, the leakage
> inductance is high enough to prevent saturation and the only limit will
> be the ampacity of the wire gauge used to wind the variac. The way
> you are using it now, is limited to the saturation current of the core.
> My question to you: Why did you chose this route and not the "normal"
> way? I.E: variac across the mains, choke and pig in series, etc.
> Another question: Is your variac (sorry if you already posted this) a
> 220V or a 120V model? If it is a 120V model, then you are really
> driving the core "nuts" connecting the variac the way you did.

This is a 240 volt, 25 amp variac.

Reason:  I intend to move my welder-transformer-based choke to my
budding tesla coil and replace it with the variac on my neon
bombarding bench.  In case you're not familiar with neon tube
processing, the tube is evacuated and then a high voltage, high
current is passed through it to heat the glass for outgassing and to
heat the electrodes red hot to convert the electron-emissive
coating.  This is done with a pole pig (or similar purpose-built dry
transformer) connected in series with a variable inductance.  The
traditional inductance is a choke with a slidable core.  The core
has to be blocked at the setpoint because of the solenoid action. 
More modern bombarders use a saturable reactor or a solid-state
controller.  Cheap SOBs like me use welding trannys with the
secondary shorted.  Works very smoothly but makes a lot of noise.  I
figured the rattling of the welding transformer core would be
drowned out by the noise of the tesla coil so I decided to move it
and use the variac.  Several other neonists have told me that they
use variacs as chokes to regulate their bombarders.

The reason it is connected as a choke instead of an autotransformer
is that the bombarder needs to act as a constant-current source.  It
needs full open circuit voltage to initiate the discharge in the
neon tube and then  regulate to the set current after the discharge
is started.

The way I designed the circuit is I computed the choke inductance to
limit the pig primary current to 60 amps with the pig primary
shorted.  This translates to the maximum current I'll ever need from
the primary to process the largest tube.  Then the variac is used to
turn the current down to the normal operating range, typically 20
amps for a 14.4kv pig.

I know just enough about autotransformers to be dangerous.  (I have
a transformer design handbook but it's big on tables and small on
theory.)  I contemplated saturation but then imagined the variac set
at mid-scale delivering its full rated current.  That would seem to
subject the core to as many amper-turns as my choke would set at
midscale with 25 amps flowing through it.  Then there was the fact
that several other neonists are using this same setup WITHOUT the
fixed choke and claim to be satisfied with it.  I tried that but the
control range was only about 20 deg of rotation, as one would

The problem with straightforward saturation is that I'm seeing very
odd behavior, not what i'd expect to see.  The really anomolous
behavior is with the pig primary OPEN.  That is, only charging
current flowing.  With the pig loaded either with a short or a neon
tube, the behavior is much more normal.  But the control is still
rough.  It is better at higher current which is consistent with
saturation but the low current behavior is very odd.  Plus I'm
seeing VERY strange waveforms.  I don't have a web page so I've sent
Terry some scope images*.  He's going to post them to his site.  I'd
love some expert opinions about these.

I don't mind cutting this variac because once it works correctly, it
will be a permanent fixture on my bench.  But I don't want to ruin a
great old variac if there is a chance it won't work.  I need a high
voltage current range between about 250 ma and 2 amps or a turndown
range of 4:1.  Think it can do that?  It would seem feasable to cut
the core a bit at a time until the variac behaved properly as a
choke.  Correct approach?

Thanks in Advance,

John De Armond
Neon John's Custom Neon
Cleveland, TN
"Bendin' Glass 'n Passin' Gas"

*  <<<< I posted the following six pictures from John on the web - Terry >>>>