Designing a Tesla Magnifier
k12 L3
C1
P. supply C3
L1 L2 C2
The idea is to start from this primary coil:
L1: 859.1nH (calculated, 50mm x 3 mm copper strip)
This is the top load::
Toroid with:
D: 600mm (measured)
d: 170mm (measured)
With these dimensions:
C3 = 26.33 pF (not 40.38 pF)
And this is the third coil:
389mH (measured Meterman 37XR)
Diameter: 200mm
Length: 860mm
Medhurst self-capacitance: 13.89 pF
Toroid 100mm above top turn.
Probable primary capacitor:
C1: 15.51µF (measured)
Initial voltage in C1: 1000 V
Maximum primary current: 4500 A
First, some verifications:
C1 and L1 resonate at 43.6 kHz
Since L3 and the topload were used in a Tesla coil, L3 and C3 shall also resonate at 43.6 kHz.
1
f =
2π LC
1
C= = 34.25pF
(2πf )2 L
Consistent with the calculations above.
The system has 8 parameters to determine: L1, C1, L2, C2, L3, C3, k12, and w, from the equations:
1) w2L1C1 = (2m2k2+(m2-l2)(l2-k2))/(2k2l2m2)
2) w2L2C2 = l2/(k2m2)
3) w2L3C3 = 1/l2
4) L2/L3 = ((l2-m2)(k2-l2))/(2*k2m2)
5) k122 = ((k2-l2)(l2-m2))/(k2(l2+m2)-l2(l2-m2))
The ratio k:l:m determines the mode of operation. With 8 unknowns and 5 equations, 3 elements can be chosen
at will. The other are consequences.
Let’s fix L1, L3, and C3.
L1 = 859.1µH
L3 = 389 mH
C3 = 34.25 pF
Using the program Mrn6:
Mode 3:4:5:
C1=17680.0000000000 nF
L1= 0.8590783371 µH
C2= 278.3492063492 pF
L2= 54.4600000000 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.3504383220
A simulation results in:
Maximum output voltage: 718 kV
Maximum input current: 4601 A
Maximum voltage over C2: 164 kV
The input current is excessive, C2 is quite large, and the voltage over it is large too.
This will be the reference design.
Trying the next mode, 3:4:7:
C1=19570.0000000000 nF
L1= 0.8591043750 µH
C2= 75.9134199134 pF
L2= 101.8809523810 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.4555734516
A simulation results in:
The voltage over C2 is even greater, but C2 is smaller. The input current is the same.
Next mode: 3:4:9:
C1=20340.0000000000 nF
L1= 0.8594422004 µH
C2= 38.5406593407 pF
L2= 121.3957475995 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.4876948992
This mode has something very interesting. C2 is a bit larger than C3. For reasons of symmetry, the capacitance
“seen” at the bottom of L3 must be similar to the Medhurst capacitance of L3. With this design a distributed
capacitor probably can be used for C2. Modes with higher “m” will be unrealizable because C2 will be too small.
The input current is also reasonable. But k12 is too high.
Trying then the next series:
Mode 4:5:6:
C1=16840.0000000000 nF
L1= 0.8591577670 µH
C2= 432.4494949495 pF
L2= 33.4296875000 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.2813124431
C2 is too big. The input current is within the limit. Note that the output is inverted.
Mode 4:5:8:
C1=18160.0000000000 nF
L1= 0.8593985738 µH
C2= 121.9729344729 pF
L2= 66.6694335938 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.3825061606
C2 is smaller, k12 is quite high.
Mode 4:5:10:
C1=18780.0000000000 nF
L1= 0.8590853593 µH
C2= 63.4259259259 pF
L2= 82.0546875000 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.4173650062
Too high k12, too high voltage on C2. C2 may be distributed.
Mode 4:5:12:
C1=19110.0000000000 nF
L1= 0.8592289244 µH
C2= 39.9743230626 pF
L2= 90.4121093750 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.4342689724
C2 is possibly too small. k12 is too high.
Trying the next series, that has the output with the original polarity:
Mode 5:6:7:
C1=16410.0000000000 nF
L1= 0.8592865786 µH
C2= 620.8111888112 pF
L2= 22.7048979592 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.2348371280
Huge C2, smaller voltage, good input current, good k12.
Mode 5:6:9:
C1=17400.0000000000 nF
L1= 0.8592900702 µH
C2= 179.3454545455 pF
L2= 47.5444444444 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.3300165012
Not too bad values.
Mode 5:6:11:
C1=17900.0000000000 nF
L1= 0.8593462417 µH
C2= 94.9475935829 pF
L2= 60.1181818182 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.3658664616
Smaller C2, greater voltage, k12 increased a bit.
Jumping to mode 5:6:15, to explore the nice symmetry on the voltage over C2:
C1=18370.0000000000 nF
L1= 0.8593024823 µH
C2= 42.7012987013 pF
L2= 71.8872000000 mH
C3= 34.2500000000 pF
L3= 389.0000000000 mH
k12= 0.3949375800
C2 can be distributed, the voltage over it is quite high, the input current is ok, k12 is probably at the limit.
It’s clear what will happen in higher-order designs. The next with this regular waveform over C2 is mode 7:8:21,
or 6:7:18 with inversion. These modes are about the last that can be constructed with a positive extra
capacitance at the “transmission line” between L2 and L3. These low-order designs exploit better the fast energy
transfer possible with a magnifier. Higher modes will be progressively less difficult to build, but in terms of
number of cycles required for the energy transfer will not be very different from a Tesla coil.
So far, the modes of the type k:k+1:3k look the most interesting.
There are other possibilities, as increasing the difference between “k” and “l” instead of between “l” and “m”.
21/2/2010 Antonio Carlos M. de Queiroz
