[Prev][Next][Index][Thread]

More real experiments



Well gang,

I have been busy over the weekend with my Pearson current monitor and 
capacitive HV voltage divider.  I physically measured the voltge across 
the gap and current through it while the coil was in operation.  In an 
effort to find the power consumed by the gap in operation.

I used the small 15VA coil of mine in this test.  The gap was placed 
across the 4KV neon sign transformer secondary.  (turned out to be a bad 
move)  This allowed ground referenceing one side of the gap. The CT was 
placed in the gap to cap connection.

Readings were taken as a trial balloon on the old, bulletproof, but 
marvelous Tek analog 535 space heater scope unit.  The voltage waveform 
looked like a bunch of noodles boppin' around on the screen.  Line sync 
caused a more or less stable display, mostly less.  It seems the gap was 
firing at odd, totally non- repetive, parts of the sine and also a 
totally random points on the sine and also a totally random number of 
times on the half cycles.  Thus the sync stability was of only a mixed 
blessing.  This showed that the system could be safely observed with 
normal scopes with little problem.

I then used a much more sophisitcated and expensive Tek 5000 series 
analog storage scope.  The waveform was locked in and captured.  It was 
noted that the above characteristics of the gap firing in a willy nilly 
fashion was why a repetivie scope, properly sysnced, was just no good for 
detailed analysis.

Now, confident of my divider and CT, I commited my new Tek TDS 340 
digital storage scope to the effort. Waveforms were isolated and captured 
and direct printouts were obtained from the scope's memory via its 
centronics parallel port to a 24 pin dot matrix printer.

Analysis:

Due to my using a variac and only allowing the AC line voltage to reach 
28 volts in to the neon's primary, I measured a peak voltage out from the 
transformeers secondary of 2640 volts.

This showed up as the max voltage across the non-firing gap.  Firings at 
the gap showed a rather immediate fall to ~200 volts.  The  current 
transformer indicated a peak current in the system at this point of ~80 
amps.  Thus under the optimal conditions the gap had a lowest possible 
resisitance of 2.5 ohms. (The gaps were hardly making any noise!) For the 
instant of max turn on, the gap consumed 16,000 watts of peak energy.  
The peak tank energy in our little 15VA system would have been on the 
order of 160,000 watts. based on a 2KV firing point and 80 amp tank 
current.  10% losses figured this way.  

The scope was set up to yeild a mathed third trace as A X B  yielding 
volt amps.  This was intergrated by hand with time to yield a total sine 
consuption of energy on the order of 1.6VA.  With the power factor 
corrected primary hooked to a watt meter we read 15.1 VA while the system 
was on (auto integrating).  This shows that we lost about 10% of the 
input energy in the gap on average.  Another cross confirmation.

Caveats:

The gap shunted the transformer out to such a degree that the transformer 
voltage followed that of the gap!  There is good news and bad news. 
(ain't that the way!)  The good news is that the power drain on the 
primary side was not nearly as bad as if a non-shunted transformer had 
been used.  This allows for a simple watt meter to be used with little 
error due to non-upset of the input voltage. (no odd waveshapes or 
spiking)

The bad news is that the high impedance of the secondary also caused the 
voltage to stay low for a protracted period (as much as 3ms-.003sec!!!). 
 This was due to cap shorting of the secondaary as it continued to bring 
the little transformer to its knees while it charged.  The charge time 
reflected the RC time constant of the high impedance secondary.  With a 
.008ufd capacitor this indicated a TC of about 600us  (.003s/5).  Thus, a 
rough idea of the charging impedance was on the order of 75,000 ohms!!  
This effect also aided in quenching the gap very well.  I detected only 5 
cycles in the current ring extending over only 10us! Again, peak current 
in the primary tank was on the order of 80 amps.

The gap consisted of 6 series static gaps (tungsten).  The operation is 
so quiet as to be virtually inaudible. (to these old grizzled ears.)  
Spark output at this level to a grounded rod is under 4".  At not time 
during measurement was a spark taken to ground.

I am sure that had a non-shunted transformer been used the voltage would 
not have hung around the zero axis so long after gap quench.  The 
capacitor TC would have improved, the quench might have actually 
worsened, system power consumption at the same peak cap voltage would 
increase due to increased gap losses and additional firings per half 
cycle (cap could come back on line more rapidly than in shunted system).

Basically, I got about 4 pops per cycle.  It varied wildly too, from 3 
per cycle to 5 per cycle. (~240BPS).

Experimenting will continue to obtain real data.  Note, I would not read 
a whole hell of a lot into this crap!!  These results are for one system 
with very specific size, frequency, gapping, top loading, operator skill, 
etc.

The neat thing is that it showed the old venerble neon transformer to be 
a good quenchin' thing, helping us out a bit and limiting the number of 
pops per cycle.  It further points up the old maxim I coined a few years 
back.  "Rotary gaps on a neon system are valuless"!  In general, they 
just are not needed nor desired in most any scenario involving one or two 
neon transformer systems.

Richard Hull,  TCBOR

P.S. good to get away from the theoretical bullshit once in a while.  
Back to some good ole "rolled up sleeve" engineering!