Stefan's Tesla-Pages

My 4"-System

description of the parts and pictures of the photo session in August '97 and some new thoughts


[home]

Table of contents:

Specs
Results
Photos
Setup
Sparks
Arcs
Plasmaglobes
HV-tricks
Artificial lightning
Flame
Ring of fire
Pinwheel
Questions and new thoughts about an MMC buildt out of the WIMA FKP1 NEW! (March'02)


I buildt a 4"-system with an outer diameter of the primary of about 28.3" (72cm) which was designed to run with some of my neons (4 pieces of 63mA/8kV). In August '97 I fired my 4"-system up in the basement of my parents. The space I had there was 10' (3m) by 13' (4m) at a height of 8.5' (2.6m).

Specs of the 4"-system in August 1997:

Control board
The control board is the same as used in all my systems. I only changed one meter from 5A to 15A and installed an additional 6A-linefilter in parallel to the existing one. A detailed description is here.

Variac
A new one, because there are nearly no hamfests in Germany and when there is one, then it is very far away from my location. But this variac was a cheap one from Asia, it costs me about 100US$ (cheaper than from surplus here in Germany) and is rated for 10A continuous. It comes in a metal housing with cooling slits. This allows to cool it with an old hairdryer (switched to cold air of course) if the power level should increase sometime ;-)

Breaker-problem
One problem I had with my variac: nearly every time I turned the main switch on the control board into the ON-position and let the line voltage to the variac, the main breaker tripped. This happened due to the high current flowing during the switch on procedure. To overcome this situation, I buildt a small timer circuit with a relay which shorts some power resistors in series to the variac after the first second I turn the thing on. This is a kind of a 2-phase soft start. (was only a problem with breakers, I never had a fuse fail this way.)

Neons
Thanks the discussions on Chips list (especially the superb explanations from Richard Quick), I tried contacting our local neon design stores and by going into 15 or more shops, I collected one 50mA/6kV, one 80mA/6kV and four 63mA/8kV just for free (one shop offered me also to take some "short" pieces of neon cable for free, at home I measured that the total length of the pieces was 246' (75m) and all were longer than 26' (8m)). Asking again and again in the shops where I've been successful, I collected an equivalent of 5kW/8kV so far. For providing the power to the system in this session, I used 4 pieces of 63mA/8kV-neons. They were compensated with 172µF and the meter read 2600VA when I turned the variac full on. Click here to go learn more about neons on my neon page.

Filter board
The filter board on the HV side was the same as described above for my 2"-system. I only installed saltwater bypass caps (2x 0.56nF) and some high power resistors (2x 260Ohm@11W).
I don't recommend the use of chokes any longer, better use an RCR-network instead of the chokes!

"RQ"-Gap
My main system spark gap is 'the famous cylinder static gap'. It is a static gap made of 0.9" (22mm) dia copper pipes in a 4" PP drain pipe. The length of the electrodes is 4" (10cm), they are spaced 0.028" (0.7mm). I use a series/parallel combination of 2 times 4 gaps in series ("folded connections"). A muffin fan provides a steady air stream through the pipe. The feed line of the fan has a line filter to block the RF coming from the gap.

Capacitors
Saltwater caps again, because (1) they are cheap, (2) commercial caps are rare and NOT cheap and (3) I hadn't time enough yet to build some homemade poly caps (one of my current projects or click here to go to my capacitor page). I made the caps out of Schweppes bottles (1 liter). There are 8 of them combined in a cardboard box and I have 3 boxes. Box1 is 8.7nF, box2 is 8.1nF and box3 is 12.7nF for a total of 29.6nF.

Primary
The primary consists of 17 turns soft copper tubing with a diameter of 0.4" (10mm) with a spacing of 0.24" (6mm) between the conductor surfaces. It is wound in an inverted conical section of 30°. The spacing between primary and secondary is 1.2", the maximum inductance about 110µH.

Secondary
The 4"-secondary is wound with AWG22 (0.63mm) magnet wire on a 4.33" (11cm) dia form of PP. Winding length is 21.25" (54cm) to give a ratio of 1:4.9. About 820 turns give me an inductance of about 13.5mH, a self capacitance of about 8.85pF and therefore a natural frequency of around 460kHz, with various toroids the frequency drops down to around 200kHz. The winding starts 4.5cm from the bottom and the length of the form is 61.5cm. I sealed the coil with 4 coats of clear 2 part epoxy and sealed the form with pertinax (bottom) and plexiglass (top) which were also epoxied in place.

RF-ground
Instead of my basement in my old home, I had the chance to built a good RF ground by myself: I took some old copper pipes and "wired" them down to the garden (the basement of my parents is on the first floor...). I ended up with a length of 41' (12.5m) of copper pipe (0.63" = 16mm outer diameter) containing 11 connections. I hammered 4 pipes (3' = 1m long) 33" (85cm) deep into the ground. They were lined up with a distance of 26" (70cm) between them. This ground was not heavy enough for my 2.6kW-coil !!! Though I watered it with 10 liters of water every morning, I increased the length of my arcs from 110cm (43") to over 143cm (56") one day when I decided to water the RF-gnd with over 50 liters. I did this because the sparks were unsteady and much shorter than the days before. Dismantling the main spark gap and cleaning it with fine sand paper brought no improvement, also I couldn't see any other failure than a bad ground. Next time I look for longer ground rods and an area with more moisture...

Counterpoise
Due to the long way down to the garden where the RF-gnd was located, I decided to build a counterpoise for providing more current to the bottom of my secondary. It consists of several sheets of aluminium foil laid down on the concrete floor and connected to the coil above. The sheets themselves simply made contact by an overlap of 2"(5cm). The covered area on the floor was about 6.7'x7.2'(2m by 2.2m).

Top-load
I used two different top capacitances for the 4"-coil:
Version_1: T8 (65x10cm, ca. 26pF, made of flexible aluminium vent tubing with styro plate in the middle) for max. spark length (=> f0 drops down to about 230kHz).
Version_2: conical section with different smaller toroids e.g. for "ring-of-fire display" and "pin wheel".

back to top of page


Results:
To make it short, after some days of experimentation, I ended up again with space problems, because the produced arcs exceeded the length of 56" (1.43m). The typical spark length (length of most of the sparks) was around 31/2 feet. Now I know that my 4"-system can produce arcs longer than 41/2 feet, but I have no chance to measure them :-( .This makes me even more sad as I recently found six 50mA@8kV-neons for free which makes it possible to use all of the power the line gives me at 230V (16A cont. or up to 30A for short time runs). Well, this should be the goal during the next years as well as finding a place to fire those bigger and bigger TCs up (until I eventually decide to make a 15kW DC-TC with all 3 phases of the line... ;-)

But perhaps I switch to magnifiers some day and use the power more efficient. I made my first 3-coil-system also during the August'97 experimentations, but it was not a real magnifier because it only consists of what I had at hand: two secondaries at the right frequency, but not a tight coupling primary, a high current secondary, a short tertiary or a really fast quenching gap. So the result was as expected, the sparks were of the same length as they were in the system where I used the 2"-secondary originally. But it was my first 3-coil system :-)!

Another personal record I made during these days was 28" (70cm) arcs with only 635W (?? have to measure this again with 30nF, this is the value for 9nF) input power to my 4"-system. This was done with all of my caps (around 30nF at this time) and lead to "resonant charging". Lets do some math here to explain this: Power from one xfmr is 635W at 8kV. So the current is 80mA. The impedance Zx of the xfmr is therefore 8000V/0.08A=100kOhms. For resonant charging, the impedance of the cap (Zc=1/2.Pi.C) must be equal to the impedance of the xfmr. So the right value of the cap would be C=1/2.Pi.Zx=32nF. This resonance condition leads to a rise in the voltage. I observed this by adding more identical xfmrs as I wondered about the voltage where my main spark gap breaks through (all time with 30nF): one xfmr -> 50V, two xfmrs -> 100V, three xfmrs -> 150V, four xfmrs -> 200V. So by adding more xfmrs, the 50Hz resonance circuit comes more and more out of tune. From this it is obvious that the right capacitance for 4 of these neons would be around 128nF but up to now I have no more caps.

One more thing I played with was a pickup-coil made from my old secondary of the school project and T7 for tuning it. A neon indicator bulb was placed between the coil and T7. I connected the base of the pickup coil directly to the ground rods in the garden. I was able to light the bulb by turning my TC on at the first floor.

back to top of page


Images from the August '97 photo shooting:

All spark lengths are measured straight line point to point!

Setup:

F04B10
"Setup"

Setup of the session performed with the 4"-secondary in August '97 in the basement of my parents.

F04B14a
"Neons"

4 Neons 63mA/8kV (driven up to 2600W) with PFC-caps (174µF).

F04B16a
"HVFilter"

Filterboard (labeled image) with bypass caps, safety gap, damping resistors and chokes (click here to get the unlabeled image).

I don't recommend the use of chokes any longer, better use an RCR-network instead of the chokes!

F10B04
"Caps"

2 of 3 cardboard boxes, each containing 8 salt water caps (1 liter Schweppes bottles) for a total capacitance of around 30nF. Click here to go to my capacitor page.

 F10B03
"SystemV1"

Setup for max. spark length (single sparks from toroid T8). Max. arcs with this 4"-secondary and 2600W input power: 143cm (4'8").

F10B18
"SystemV2"

Setup for other beautiful displays, e.g. "ring of fire" and "pinwheel". The T8 was replaced by several other toroids and a conical section.

back to top of page
Sparks:

F10B16
"Spark1"

85cm (33.5") long vertical spark from T8. The faint spark to the left is hitting the curtain rail at a distance of somewhere between 125cm and 140cm. 400ASA, f2.8, 1/60s

F10B07
"Trident"

A spark which looks like the trident of Mr. Neptune. 400ASA, f2.8, 1/60s

F10B08
"Spark8"

Another nice spark from T8, spark length 92cm (36"). 400ASA, f2.8, 1/60s

F05B20a
"Flame1"

Long time exposure, give a flame-like display of many sparks. 200ASA, f5.6, 1s

F07B10
"6String"

Hot spark, showing the banjo effect. 200ASA, f5.6, 1/15s

F07B13
"forked1"

Nice forked spark from T8. 200ASA, f5.6, 1/4s

F07B16
"forked2"

Another nice forked spark from T8. 200ASA, f5.6, 1/8s

back to top of page
Arcs:

F05B09a
"Snake1"

Short time exposure of a 60cm (24") long arc from T8. 200ASA, f16, 1/60s

F05B03
"Twister"

HOT 60cm (24") long arc from T8, looks like twisted. 200ASA, f16, 1/8s

F07B01a
"climbing"

A climbing 93cm (37") long arc from T8. SystemV1, input power 2600W. 200ASA, f16, 1/4s

F07B06a
"rubber"

85cm (33") long arc from T8, looks like a vibrating rubber band. SystemV1, input power 2600W. 200ASA, f16, 1/8s

F07B09a
"Snake2"

Short time exposure of a 94cm (37") long arc from T8 SystemV1, input power 2600W. 200ASA, f11, 1/60s

F07B21a
"25steps"

25 steps to the ceiling: 113cm (44") long arc from a pencil on top of T8 (systemV1) to the ceiling. Input power 2600W. 200ASA, f5.6, 1/4s

F07B30
"FireBall"

Flame like display with many power arcs (max. 125cm / 49") from foil covered styro ball on top of T8 (systemV1) to the ceiling 115cm (45") above. Input power 2600W. 200ASA, f5.6, 8s

F08B06
"coolbulb"

A cool way to light up an incandescent bulb: the current is flowing through the filament of the 60W-bulb, the arc (107cm = 42") arised from a wire contacted to the thread of the bulb. The bulb is glowing with the brightness of an 35W-bulb (@230V/50Hz => 160mA RMS)! SystemV1, input power 2600W.
200ASA, f5.6, 1/8s

F05B21
"Oops!"

Expanding 122cm (48") long arc from a pencil placed on top of T8 to the ceiling. SystemV1, input power 2600W. 200ASA, f8, 2s, distance 2.5m, focal length F40mm

F06B12
"manyarcs"

Lots of power arcs @2600W input power. The length of every arc exceeded 75cm (30"), max. arc length exceeded 84cm (33"). 200ASA, f16, 8s

back to top of page
Plasmaglobes:

F07B33a
"gapglobe"

Plasmaglobe (12cm = 5" dia) on top of T8 on the spark gap excited TC (systemV1, 2600W), a nearby placed pencil reduced the electrical stress on the globe (and prevented the poor globe from puncturing by the wild 45" long sparks). 400ASA, f5.6, 1/15s

F09B07a
"ssglobe"

Plasmaglobe (6cm dia standard light bulb) driven by a fly-back circuit with two 2N3055. 400ASA, f5.6, 1/4s

back to top of page
HV-tricks:

F08B31
"LongTube"

This is me standing on the electrode of the fly-back circuit, holding a long fluorescent bulb in my hand. 200ASA, f5.6, 8s

F08B23
"trick1"

Me and my TC. DO NOT ATTEMPT THIS! This photo is taken by trick photography J. NEVER TOUCH THE PRIMARY OF A RUNNING TC! NEVER GO AS NEAR TO A RUNNING TC! Input power 2600W. Max. spark length from the fast spinning pinwheel is about 57cm (22"). 200ASA, f3.5, 4s (additional photo taken by light with automatic)

F08B21
"trick2"

DO NOT ATTEMPT THIS! This photo is taken by trick photography J. NEVER TOUCH A RUNNING NEON XFMR! NEVER GO AS NEAR TO A RUNNING TC! Input power 2600W. Max. spark length 68cm (27"). 200ASA, f5.6, 4s (additional photo taken by light with automatic)

back to top of page
Lightning bolts (all pictures shown upside down):

F07B25
"Strike1"

115cm (45") long power arc to the ceiling, upside down it looks like a lightning bolt striking into a lake. Input power 2600W. SystemV1. To make the picture look 'more realistic', I removed the TESA-strips from the 'lake'. If you want to see the original photo, click here.  200ASA, f5.6, 1/8s

F07B24a
"Strike2"

Another 115cm (45") long power arc from the styro ball on top of T8 to the ceiling, upside down it looks again like a lightning bolt striking into a lake. Input power 2600W. SystemV1. 200ASA, f5.6, 1/15s

F07B14
"1Bolt"

Sparc from T8 (systemV1), upside down it looks like a lightning bolt. Input power 2600W. 200ASA, f3.5, 1/30s

F10B14
"2Bolts"
Sparks from the styro ball on top of T8 (systemV1) , upside down it looks like two lightning bolts. Input power 2600W (current background picture). 400ASA, f2.8, 1/60s

back to top of page
Flame:

F08B11
"Flame2"
Long time exposure of many sparks, flame like display. SystemV2. Input power 2600W. 200ASA, f5.6, 4s

back to top of page
Ring of Fire:

F08B15
"RoF1"

Many sparks at the same time (!) from the (undersized) toroid T7 (systemV2) in different directions. Input power 2600W. 200ASA, f5.6, 1s

F08B12
"RoF2"

Many sparks at the same time (!) from the (undersized) toroid T7 (systemV2) in different directions. Input power 2600W. 200ASA, f5.6, 4s

back to top of page
Pinwheel:

F10B15
"Pinwhl1"

Low speed pinwheel, the sparks are twisted due to the movement of the ends. Input power 2600W. 400ASA, f2.8, 1/60s

F08B16
"Pinwhl2"

High speed pinwheel, short time exposure, note the red glow of the neon indicator lamps. Input power 2600W. 200ASA, f5.6, 1/4s

F08B20
"Pinwhl3"

High speed pinwheel, long time exposure. Input power 2600W. 200ASA, f8, 4s
see also the trick photography of the pinwheel

back to top of page


Some really hard questions:
The energy stored in my caps (bang size) is E=1/2.C.Û2=.5.30nF.1.4.8000V.1.4.8000V=1.92J.

Some more math around my 4"-system (as configured in 1997): 

Tank circuit:
Cp=30nF
L=16uH
f=230kHz
U=8kV
Ep=1/2.C.Û2=1.92J=1/2.L.Î2
=> Tank circuit peak current Î= 490A, reactive power  per pulse was about U.I= 2.8MVA.
Secondary:
Es= up to 0.5 times Ep (no measurements on the efficiency up to now)
Cs=30pF
Ûs=sqr(2.Ep/Cs)=
up to 253kV.
The sparking secondary rings down in about 5 cycles (guess). From the light bulb experiment we get Is=0.16A RMS. So we can calculate an RMS (per pulse) current of 64A for this power arc. Being conservative, the reactive RMS power per pulse therefore is in the order of 7MVA.
According to John Freaus formula (spark length = 5.8 * sqrt (power input) / 4th root ( BPS)), the spark length with 2500W wallpulg power and 1200BPS should be 125cm, which is right what I got.

back to top of page



NEW (March'2002): NEW!

Capacitor:

I got some Maxwell caps for my BIGcoil, so I can use the WIMA caps to build an MMC for my 4"-TC. Further I will build a new primary coil which will take the new cap into consideration. Of course, the secondary will get the new 6"-toroid.

Hopefully, my 4"-TC can be driven up to 2m spark length (nearly 4 times the winding length) which equals about 1.85kW input power at 100BPS according to John Freaus formula
   
spark length = 5.8 * sqrt (power input) / 4th root ( BPS)

At 9.4kV (8kV-neons with some additional voltage due to the variac step up at full scale) and 2000W input power, the resonant value is 72nF. Since all the components neede for this hobby are really heavy, I'll use no variac if the system is brought roughly into tune (=>recalculate for 8kV). For a static gap, best size (LTR) is 120% (160%) of this value which will result in 87nF. (For an SRSG, 260% (320%) of resocap size will be best, this will result in 188nF).

I will use two or three 63mA-neons with their shunts removed (the four unmodified neons I used in 1997 didn't get warm even after prolongued use).

Thanks to Kurt and his great EXCEL-Sheet (http://home.datacomm.ch/k.schraner/MMCcalcWIMA33b.xls), I calculated some variants of a suitable MMC out of the WIMA FKP1 (6kVdc, 33nF, 700Vac):

3 caps in series:
static gap: 8 strings in parallel 88nF 7.8J per bang 24 caps needed
SRSG (100 BPS): 17 strings in parallel 187nF 16.5J per bang 51 caps needed

4 caps in series:
static gap: 11 strings in parallel 91nF 8J per bang 44 caps needed
SRSG (100 BPS): 23 strings in parallel 190nF 16.8J per bang 92 caps needed

5 caps in series:
static gap: 13 strings in parallel 86nF 7.6J per bang 65 caps needed
SRSG (100 BPS): 28 strings in parallel 185nF 16.3J per bang 140 caps needed

6 caps in series:
static gap: 16 strings in parallel 88nF 7.8J per bang 96caps needed
SRSG (100 BPS): 34strings in parallel 187nF 16.5J per bang 204 caps needed

So, which one will be realised? Though the experience of others say that 3 caps in series will be sufficient, I will take more to be on the really safe side (so I'll have some margin if anything goes wrong). I bought 189 caps, so 6 caps in series is not an option if I ever will decide to run the 4"-System with an SRSG. So the correct answer is (at least up to now ;-) 5 caps in series. I will leave some caps for replacement purposes (hopefully I will not need to do this). A reasonable arrangement seems 2 panels of max. 16 strings each (or 4 panels of max. 8 strings each or even 8 panels of max. 4 strings each).

At the "toom" hardware store, I found some U-profiles (19x10x1mm) made from plastics (dont't know what kind of, perhaps PVC or PP) here my WIMA FKP1 (6kVdc, 33nF, 700Vac) caps will fit perfectly in. This will allow some space between the caps in each row (good feature if one will blow) and an easy mounting of the strings. The mounting process will be as follows: cut the profile to the desired length, mark the position of the caps, dispense a dab of hot glue inside the profile and place the first cap. After 10 seconds or so, dispense the next dab and place the next cap and so on.

Each cap is 40mm long, so with 5mm spacing between the caps and 30mm at the ends (for mounting purposes), I get 3 strings out of one length (1m) of the U-profile. With a lenght of 28cm, the strings will fit in a standard box for storage purposes (important, because my basement is VERY small).

I'll use a special high voltage bleeder resistors across each cap. Other people have ruined their MMC due to overvolting the bleeding resistors which resulted in a shorting out the cap on which they are mounted - which led to a higher voltage on the other components of the string resulting in more failed resistors or caps... So it's really worth to be on the save side here with such a big MMC! For a low dissipation I'll use 18(22/27)MOhm resistors (18-27W total losses) rated 10kVpeak each.

( ... )

Primary:

I'll build a new (smaller) primary out of flat copper strap and use a separate tuning coil (perhaps motor driven so that I can tune the coil remotely while it is running).

( ... )

back to top of page

[home]