Re: DANGER! Strange transformer problem
Tesla List wrote:
> Original Poster: "R.E.Burnett" <R.E.Burnett-at-newcastle.ac.uk>
> However, it could be resonating at an odd harmonic of the
> supply frequency, such as 180Hz or even 300Hz. This can happen
> because these frequencies are quite strong in the line current,
> and little stray capacitance is required.
> This effect is well known to power companies, and is a hazard
> of running large transformers with no load. The losses in such
> equipment are so low that there is nothing to damp these
> spurious resonances. (A little resistance in the line may help ?)
> Also, as the voltage builds up, the transformer will become
> more non-linear as its core saturates, and will add too these
> 3rd and 5th harmonics ! Does the growl build up slowly ?
> You said you had observed voltage waveforms. Did they contain
> LARGE quantities of 180Hz or 300Hz, at the "bad-spot" on the
> variac ?
Bingo! I think you nailed it. Look at
This sure looks like a bunch of harmonics to me. Then look at:
This shows the waveform across the pig during the first 30 seconds
after power up. It takes 30 seconds for the thing to stabilize.
I've built this whole system using utility (high quality, low loss)
components. With the pig primary horn wrapped in self vulcanizing
tape, there's not even any corona to dissipate energy. We even
baked the pig under vacuum at the transformer shop to get the oil
super-dry. Everything is conspiring to make this thing a high Q
oscillator. Hot darn!!!!
The waveform during the 30 second powerup period is so asymmetrical
it looks like an inductor with DC saturation. I was so deep into
this that I never thought of harmonic saturation; I was off chasing
the source of the DC.
I'm going to do some more testing to confirm this but I'm almost
sure you've nailed it. The "machine gun" sound I heard would be
consistent with this too. Each thump would likely result in enough
corona to damp the harmonic oscillation, then it would build up
again and repeat irregularly. I'm going to try to digitize enough
of this waveform to do an FFT on it but just walking dividers across
the waveform at the bottom of pigscope4.gif indicates lots of 5th
Thanks a zillion! I was about to start obsessing over this.
John De Armond
Neon John's Custom Neon
"Bendin' Glass 'n Passin' Gas"