Re: TV Flyback Voltage
Kolas, Sat May 13 2017, 05:31AM

A while back I played with light bulbs as plasma globes. I found only AC to work, And that was generally the primary requirement. I also did a little experimentation with a larger 6" globe.

Most modern day flyback transformers are internally rectified, but I bet you knew that.

25 kV at the proper frequency should be more then plenty for a plasma globe.

Hope it works out for you.

Re: TV Flyback Voltage
klugesmith, Sat May 13 2017, 02:08PM

When Alan says old color TV, he means really old color TV. Here is a video of someone fixing a flyback in similar (CTC28) set. The secondary winding termination is in the form of a tube anode cap, for the EHT rectifier.

Good luck with those, Alan. Please let us know what you learn about the transformer parameters. I presume it has a gapped ferrite or iron powder core. From the normal flyback application, we can get a secondary volt-seconds number that's comfortably away from saturation. Thus a kV per kHz limit for sinusoidal voltage. Probably about the same as other color TV flybacks not used with voltage multipliers.

Re: TV Flyback Voltage
alan sailer, Sun May 14 2017, 01:18AM

Klugesmit his right. It's an old flyback, no diode with a secondary coil over 3" in diameter potted in what looks like silicone rubber.

I'd like to pull apart the core but it is encased in a steel support that is glued to the ferrite with some sort of rubbery grey adhesive. I'd give a high probability of cracking the thing if I took it apart. So I can't easily tell if it is gapped. The core is not very beefy, about 1/2" diameter.

If you can educate me one what you are asking I can certainly consider making measurements. The terms volt-seconds and kV per kHz are not part of my technical lexicon.

I've powered it up using a push/pull MOSFET driver and gone up to 25kV pk-pk output so far. It has a resonance at ~35kHz driving as estimated ~20pF/2Mohm load (AKA electrode-less plasma tube).

Cheers.

Re: TV Flyback Voltage
klugesmith, Tue May 16 2017, 09:52PM

[edit]Bottom line first! The stresses on your transformer from the reported voltage and frequency are substantially lower than those in its original flyback application. That's without considering the output current and associated copper losses. [\edit]

Alan, here's one amateur's simplified review and conclusion about sine wave voltage limits. For now we'll ignore the winding's parasitic R and C values, and circuit resonances.

Take an old TV flyback (LOPT) secondary winding that's designed to power a color CRT anode, via one vacuum rectifier. Let's name some variables:

• Vfly, the nominal output voltage, 25 kV.

• T, the television horizontal scan period. 63 us in NTSC standard.

• Tf, the flyback pulse width. Say it's exactly T/9; that's 7 microseconds.

• Ts, the width of the rest of the cycle. 8T/9, or 56 microseconds.

• Fmax, the design value of magnetic core flux at highest point in cycle.

We don't need to know the values of Fmax or Ns, the number of turns in winding.

The induced EMF in secondary winding is, at every instant, proportional to dF/dt -- the rate that magnetic core flux is changing. (As discovered by Michael Faraday and Joseph Henry almost 200 years ago.) The ratio is fixed by design; in fact it's just the number of turns. The ratio doesn't change when we apply different frequencies and waveshapes to the same transformer.

In original application, the flux ramps from zero to Fmax during Ts, then back to zero during Tf. Slope dF/dt is -9 Fmax/T for the fast phase, which induces +Vfly. (The slope is 9/8 Fmax/T for the slow phase, which makes the winding voltage -Vfly/8.)

Consider a symmetrical square-wave voltage with the same old period, T. The flux can swing both ways (+/- Fmax) if the device is used as a regular transformer. Ramp from -Fmax to +Fmax for a half cycle, then ramps back to -Fmax in the other half cycle. Both slopes have magnitude dF/dt = 4 Fmax/T. So the voltage magnitude, both pos and neg, is 4/9 Vfly. About 44% of 25 kV, so 11.1 kV.

How about sinusoidal voltages and fluxes, with the same Fmax and T values? dF/dt reaches its extreme values at flux zero-crossing times. Peak dF/dt is 2 pi Fmax/T. So the voltage amplitude (peak value) is 2 pi / 9 Vfly. About 70% of 25 kV, so 17.5 kV.

RMS voltage is sqrt(2) pi / 9 Vfly. About 49% of Vfly, so 12.3 kV.

I'm not sure if peak-to-peak voltage of a sine wave is particularly relevant, except for being easy to measure on an oscilloscope screen. Its value, for the same frequency and flux level as original flyback circuit, is 4 pi / 9 Vfly. A ratio of 1.4, yielding 35 kV pk-pk.

All three examples have the same Fmax value, shown (by original design) to be reachable without saturating the core material. Core loss power, and temperature rise, might be as much as 4 times higher in the bipolar cases.

Without changing Fmax, all voltages can be increased in direct proportion to frequency. Not counting limitations due to insulation or core heating.

I hope that helps you feel more at home with your transformers, Alan. And that it's not seriously flawed.
-Rich

p.s. Here are a few SWAGs at actual values, which we didn't need for the comparison above.
Original application flyback pulse has a voltage-time product of 0.175 volt-seconds.
If Bmax (peak flux density) is 2760 gausses, in a 0.5 inch round ferrite bar, then Fmax = 35 microwebers. Fast slope dF/dt is 5.0 Wb/s, inducing 5.0 volts per turn. 5000 turns for 25 kV. Total flux linkage (max) is 0.175 weber-turns.

Re: TV Flyback Voltage
alan sailer, Wed May 17 2017, 11:26PM

That great. I'll look at it in detail later. I love diving into new (and possibly helpful) information.

I have been messing around with the flyback on my simple push/pull driver and have noticed that the core gets darn hot (>80C) at an input power level of about 30 watts. I'm guessing that the old core materials are lossy compared to modern ferrites.

I have done quite a few frankenstein transformers using flyback secondaries on a very beefy modern core with no noticeable heating.

I'll try looking into that question.

Cheers.