# Can I power my neon sign transformer with a higher frequency?

My neon sign transformer (NST) input says 120 volts 60 hertz. Its output is 15 KV, 30 mA, 60 hertz. I was wondering if I could get a light ballast which I’ve seen drive fly back transformers at 30 kilohertz. Can I get one of those to drive my NST? I’m really wanting the high frequency instead of the 60 hertz. Would it work or not work?

• Probably the inductance will prevent any significant current from flowing. May 16 at 12:59

Your existing 60 Hz Neon Sign Transformer (NST) must be fed from a low frequency of approximately 60 Hz.

A lower frequency will reduce the voltage handling, as the core has a constant Vs. You could run it at 50 Hz for instance with a corresponding proportionate de-rating of the voltage.

A higher frequency will increase the core losses, and increase the voltage rating of the magnetics, though not of course the insulation, so there's a limit to how much increase would be useful. If you had a suitable inverter, you could maybe run it up to 100 Hz with a corresponding increase in core heating, and probably not too much effect on the magnetic efficiency. Going all the way to up 400 Hz, a common aircraft frequency used to reduce the amount of iron carried on an aircraft, would probably not work as 400 Hz transformers are built with much thinner laminations than 50/60 Hz transformers.

Going up to 30 kHz would be completely infeasible. You would be better off building a ferrite-cored transformer to be driven by your inverter.

A mains frequency transformer uses an iron core of high permeability for achieving high magnetic flux in response to the exciting current. The high magnetic flux is what makes reasonable energy transfer through a changing magnetic field possible. There are remagnetisation losses with every polarity change: the core is dimensioned in a way where those losses are comparatively insignificant regarding the operating efficiency and related heat-up of the transformer.

Increasing the operating frequency by a factor of 500 is not going to do your core any favors. Something will die comparatively fast from the developing temperatures. Power transformers are designed for a narrow range of operating frequencies.

20kHz NST are common and becoming the new industry standard. Much easier to use. I don't know if any issues develop if you jump to 30kHz. Searching online for 20kHz neon sign transformer will yield many results.

• Yeah I already own a 60 hertz NST. I just plug it into the wall socket. Now I’m wondering if I can get a high frequency ballast to drive my NST or will it not work and something go wrong? May 16 at 12:10
• No, your on-board transformer must be fed by mains frequency and mains AC voltage. May 16 at 12:14

No. The frequency response of such a transformer is quite low; disclaimer, I haven't measured one myself, but I would assume it hardly extends into audio frequencies, a few 100 Hz perhaps.

The limitation is due to the design of the transformer, and its high ratio.

The design uses magnetic shunts to increase leakage inductance. This reduces output voltage some, but importantly limits output current under short circuit conditions. This is necessary because, before the neon tube is ignited, it requires a very high voltage to break down; and once ignited, its voltage drop is rather low (a few kV?), practically a short circuit in comparison.

The high turns ratio is needed to achieve enough open-circuit voltage, and doubly so for the voltage lost due to the use of magnetic shunts. So the secondary might tens of thousands of turns. This is a lot of wire, and anywhere two wires lay adjacent there is capacitance; so there is a lot of capacitance as well. The combination of leakage inductance and secondary capacitance creates a low-pass filter, and this likely rolls off at some 100s of Hz, maybe kHz at most. (I have measured standard power transformers: the kind with the primary and secondary "bank" wound side-by-side on a bobbin, but no magnetic shunts between them, typically have a cutoff of some ~10kHz. Microwave oven transformers are designed somewhat inbetween, with magnetic shunts, but not enough for full short-circuit current limiting; they have a cutoff around ~4kHz.)

To generate high voltages at ~30kHz, not only are much tighter transformer designs required (for lower leakage inductance and stray capacitance), but the bandwidth is still not very wide, typically requiring resonant operation. A random lamp ballast circuit will likely not be suited to such a task. Basically, solutions converge on a Tesla coil style of design, above some, to 10s of, kHz.