Cap choice sims for static gaps
After posting my 200BPS rotary gap simulation results a couple of days
ago I am now presenting the results of my static gap capacitor size
simulations. I have been following the 2 Questions on Resonance thread
and am very interested.
Here are some simulation results which show the effect of using different
sized tank capacitors with a STATIC SPARK GAP. First the gap is set
to a low firing voltage which would minimise resonant voltage rise.
"Neon sign transformer -
10kV Static gap, capacitor selection"
Actual measured Neon Xfmr parameters: 6Kv / 125mA
50Hz resonant cap: 66.3nF
C = Capacitor value in nF
V = Capacitor firing voltage
P = Power throughput (= 200 * 0.5 * C * V * V)
VA= Mains supply VA
pf= Power factor (P/VA, kind of like a measure of efficiency)
C [nF] BPS V [v] P [W] VA pf
16 583 10000 466 809 0.576
20 468 10000 468 832 0.563
22 452 10000 497 872 0.570
24 424 10000 509 886 0.574
28 366 10000 512 957 0.535
30 342 10000 513 976 0.526
32 318 10000 509 966 0.527
34 302 10000 513 1011 0.507
36 298 10000 536 1009 0.531
38 268 10000 509 1025 0.496
40 256 10000 512 1044 0.490
42 240 10000 504 1038 0.486
44 230 10000 506 1074 0.471
48 210 10000 504 1113 0.453
52 190 10000 494 1063 0.465
56 198 10000 554 1262 0.439 <----- Only fired on neg
60 160 10000 480 1120 0.429
64 142 10000 454 1193 0.381
68 132 10000 449 1180 0.381
72 130 10000 468 1293 0.362
76 106 10000 403 1241 0.325
80 104 10000 416 1081 0.385
90 96 10000 432 1235 0.350 <----- Only fired on pos
100 68 10000 340 1298 0.262 <----- Missed several half
120 49 10000 294 1224 0.240 <----- Missed several full
130 - - 0 1506 0 <----- Gap would not fire
Conclusions can be drawn from these results more easily if each parameter
is plotted on a seperate graph against capacitor value on the X-axis.
It is easy to see why people beleive that a large capacitor allows more
power to be got from a neon transformer. Compare the results for both
a 20nF capacitor and a 72nF capacitor. Both capacitors charge to the
same voltage (determined by the static gap,) but the firing rates are
obviously different. The point of interest is that the simulation
indicates identical power throughput (Bang energy*BPS) however the
supply VA is 55% higher for the 72nF capacitor !
Some people will ask "Where did the extra 55% of power go ?"
Well, it is not real power but reactive power. Although there would be a
55% increase in current from the "wall socket" the power delivered to
the sparks remains the same. It is merely the power factor which is
made worse. (The system is "less efficient".)
Also for those who beleive it is not possible to exceed the face plate
rating with a static gap, I would say YES as far as real power
throughput is concerned provided the gap is set small. Nowhere in this
wide range of capacitor values does a point occur where the full 750W
faceplate rating of power is processed. HOWEVER, if you only measure
the current in the power cord then the VA product will almost always
exceed the faceplate rating !
During my simulations of the static gap, I noticed several things:-
1. The gap always seems to fire quite chaotically. Sometimes a half
cycle will be skipped, and it will fire more times in the next half
cycle. This must be down to the behaviour of the circuit and not
randomness in the spark gap breakdown voltage as I have always
believed. (This spark gap is ideal and fires at exactly 10,000 Volts.)
2. With certain capacitor values I was able to get the simulated gap to
fire totally on positive cycles, (see comments to right of results
above.) I believe this is down to the "perfect world" nature of a
simulation, and a bit of real world randomness would probably stop
this from occuring. (But maybe not ?)
3. The randomness of the gap firing over many cycles makes the supply
cord current fluctuate by a small amount over time.
4. The power factor is always much lower than I found during my 200BPS
sync rotary simulations. This goes along with peoples comments about
supply current dropping when a rotary is installed.
I then repeated these tests with a higher firing voltage for the static
Vg= 25.3 kV
C [nF] BPS V [v] P [W] VA pf
32 105 25300 1075 1497 0.718
I have omitted the rest of the results to keep this already too long
post down to size. However it can be seen that the firing rate for
this 32nF capacitor is lower than before but it charges to a much higher
voltage. The power processed now does exceed the faceplate rating by
over 40% but the transformer is destined for a quick demise due to the
300% overvoltage and 200% over current of its windings !!!
(Notice the power factor has actually gone up too, but it never comes
close to that of the sync rotary.)
So what does this prove. Well, that matched capacitors are not needed
for best performance on static gap systems, and that wider gap settings
produce better performance at the expense of irregular firing and risk
of overvolting components. Big capacitors draw more current but the
simulation suggests that no more real power is processed. Maybe choose
a small cap and get power from the higher break rate.
It seems to me that static gaps and rotary gaps are indeed two totally
different ball games.
PS. If this post is too offbeat, not appropriate or just plain WRONG
then don't hesitate to shut me up.
- More simulation results from
Richie in sunny Newcastle.