Neon sign transformer power: a new look

This forum is for specialized infomation important to the construction and safe operation of the high voltage electrical supplies and related circuitry needed for fusor operation.
User avatar
Rich Feldman
Posts: 1007
Joined: Mon Dec 21, 2009 11:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Neon sign transformer power: a new look

Post by Rich Feldman » Sun Nov 18, 2018 10:22 pm

So far so good. With nominal primary voltage, I got:
425 watts into a matched resistor.
222 watts with 10 R's in series (14.4 kV x 15.5 mA). Still thinking of using 2 of these NST's in series for fusion someday.
159 watts into a string of 25 Christmas lights, last year converted from parallel to series by snipping wire in 24 places.
DSCN0931.JPG
The LED meters were hard to read in daylight, until some shades were added. Green optical filters would probably also help.
Here are the IV curves for 70, 140, 210, 277, and 290 V on primary. Plus an iso-power line at 425 watts.
nst_iv_1118.JPG
Richard Feldman

User avatar
Rich Feldman
Posts: 1007
Joined: Mon Dec 21, 2009 11:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Neon sign transformer power: a new look

Post by Rich Feldman » Sun Dec 09, 2018 8:37 pm

Last night I measured NST primary current. It's a key design requirement for high-voltage isolating transformers to go upstream of the NST's.
Same setup as before, except with a true-RMS ammeter in series with primary. Primary voltage was set to nominal and many different resistive loads were measured. Also I tried loads of 6, 10, and 20 fluorescent lamps in series.
nst_iv_1208.JPG
.
Primary current ranged from 0.305 A (no load except kV meter) to 3.43 A (short circuit except secondary mA meter). It was 2.3 A at the peak power point (11 kV, 426 W).
I expected it to behave as the sum of two components 90 degrees apart in phase: a fixed magnetizing current and N times the secondary current. Not bad, with N of about 56.
Got a better fit with a mixed model for primary current: (quadrature sum of Ia and N * Ipri) + Ib, where N=53.6, Ia=0.1, Ib=0.19. THis is more precision than necessary, and more than is justified by the metering accuracy.

Right before inserting the primary current meter, I had fired up a test fixture that connects 20 fluorescent lamps in series. They are T5's about 45 inches long (1.2 m).
DSCN0993.JPG
That was an opportunity to do some plasma wrangling, and get in the mood to make special holiday light displays.
nst_iv_1208_c.JPG
First run, in blue, began at full blast (290 V on NST primary). 287 watts, if we ignore errors from metering non-sinusoidal kV and mA. As variac was dialed down, HV current went down and voltage went up as expected. Eventually the lamp array got flickery and almost went out. Then I dialed the variac back up in small steps. Currents lower and voltages higher than before, I guess because the cathodes were now all pretty cold. At a couple of points, there was a progression to higher current and lower voltage with no change to variac setting, I expect as the cathodes warmed up to a new equilibrium. Eventually the rising-Vpri sweep curve merged with the falling-Vpri sweep curve.

Second run, in orange, started with variac at zero. Got to over 8 kV before any significant current or flickers of light. Eventually, tubes were all warm and bright. From 35 mA to max, the voltage followed familiar "close to normal operation" curve. I think operation below that point is hard on the cathodes. Good thing the lamps were all free. :-)

How about that apparently bistable lamp characteristic, hypothetically with different cathode temperatures? I think experiments like this could see stable operation with different subsets of the lamps (or individual cathodes) in hot or cold state. Novel data storage technology?

[edit]Updated the lamps IV chart. Added scaled copies of the "Resistors w/ Vp=nom." curve, to represent variac knob effect. Removed one point in blue curve that was missing in notebook, WAG'd in spreadsheet, and obviously wrong.[\edit]
Last edited by Rich Feldman on Mon Dec 10, 2018 5:33 pm, edited 1 time in total.
Richard Feldman

User avatar
Richard Hull
Site Admin
Posts: 10997
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: Neon sign transformer power: a new look

Post by Richard Hull » Sun Dec 09, 2018 9:59 pm

The lamps are a non-linear gas load and only a bit like a fusor. To bad the tests weren't run on a real fusor which is what folks here might be interested in. Resistive loads are smooth and remain constant. The transformer is designed for a gas load. Neon gas loads are really just diacs, (voltage regulators), It takes only so much voltage to light a tube of highly variable length and keep it lit. The transformer is made to crumple to its knees on a very specific ionized gas "arc" load and supply only enough current to keep it lit based on it magnetic flux shunting ability under varying load situations.
The varying load for neon signage makers is the total tube length of the signage. Ne-2 lamps have been used as low current voltage regulators of many years.

The moment you take it out of a neon gas load, a transformer's performance can vary based on what is demanded of it in a total different load scenario than it was designed for. Fluorescent tubes are a mercury metal vapor load. Clear glass blue signage is often nothing but mercury vapor with neon electrodes which are quite different from fluorescent tube filaments which can heat to incandescence if over driven. Expect interesting and often surprising results.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
Retired now...Doing only what I want and not what I should...every day is a saturday.

Post Reply