Bandwidth of neon xfmr

[Kevin Tate]

I know neon transformers are designed to operate from 60Hz. But what
range of input frequencies can they handle and still produce enough
juice to light a small neon tube?

Kev

 

[Don Klipstein]

In article <levgv0d1uicff41n0kf6gvjavfkmfq8998@4ax.com>, Kevin Tate wrote:

I know neon transformers are designed to operate from 60Hz. But what
range of input frequencies can they handle and still produce enough
juice to light a small neon tube?

For frequencies below 60 Hz: Input voltage must be reduced
proportionately with frequency once you go only a few percent below 60 Hz.
Output short circuit current should not change much in such a case but
will decrease more significantly once the frequency gets below about 20-30
Hz or so, and will decrease nearly proportionately with input voltage
(proportionate with frequency) once the frequency is low enough for output
current limitation to be due mostly to winding resistance rather than
"leakage inductance" - and I guesstimate this point to be roughly 15 Hz or
somewhere in that ballpark.
But more importantly I suspect that when using a transformer recommended
for a specific length or set of tubing, 1/3 to 1/2 voltage may be
insufficient to spark through at all - which means a low frequency limit
of around 20-30 Hz to get any results at all unless the input vopltage is
kore than the transformer should get at the frequency in question.

As for frequencies higher than the transformer is designed for:

As frequency increases, there is a trend for the maximum permissible
input voltage to increase, but less than proportionately since one
significant loss mechanism producing heat is eddy current loss, which does
not decrease much when frequency deviates significantly upward from
"proper" and "proper" input voltage is maintained.
But do expect the "leakage inductance" to cause output current to
decrease as frequency increases!

Operate a neon sign transformer rated for 60 Hz at 400 Hz, and I would
guesstimate it would have its usual temperature rise when the input RMS
voltage is about 150-160 volts. Output current to a neon tubing load at
400 Hz instead of 60 with normal input voltage I guesstimate to be about
16-20% of "normal", and somewhat more than proportionately more (possibly
30% of "normal") with input voltage of 160 volts.

Open circuit output voltage at 400 Hz: With 120V input, most likely
slightly higher than normal, but possibly very different and even possibly
much more. Please have a load connected when operating a neon sign
transformer at frequency much above rated!

--------------------------

Neon sign transformers have a resonance mode at some higher frequency,
where the "leakage inductance" and the stray/interlayer capacitance of the
secondary form a series resonant circuit. In my limited experience with a
small sample size this has happened at a frequency around 1-1.5 KHz as
best as I can remember.
Such a resonance mode can be a serious problem with no load - output
voltage can be a few times that otherwise expected, and can be a major
stress on or even causing outright quick failure of the insulation within
the transformer! Please have your "neon" tubing connected!
-------------------------------------------

Once frequency increases to much above that of the "resonance mode",
expect open circuit output voltage to be a little to somewhat more than
"expected" times the square of the ratio of "resonant frequency" to
applied frequency. As in mostly below usual and decreasing bigtime as
frequency increases. Additional "resonance modes" may at some frequencies
cause some significant deviations from this, but expect mostly output
voltage awfully low once the frequency exceeds a few KHz.
As for output current - mostly keeps going down as frequency increases.
(Exceptions possible but not the "norm".)

--------------------------------------------------------------

This is for the "traditional" iron core transformers as opposed to the
newer "electronic ballast" type devices. The latter should have
performance varying less with input frequency and hardly changing at all
when input frequency makes a moderate change, although probably have
nowhere to go but down in any performance perameter (including
reliability) if the input frequency is changed greatly.

- Don Klipstein (don@misty.com, klipstei@netaxs.com)