| ELECTRIC SCHEMA (the original schema of Mazzilli Vladimiro, april 2003) |
| Between the schema of Mazzilli and my mountage there are a few differences because the higher inductance secondary I have built. See below: F1 = fuse 250 volts @ 10 A variac = 220 vac input @ 9 amp, output 0-250 volts, I use 0-140 volts input. If you want a real OLTC (without transformers) you can build a resistive divider, but a soft increment of the voltage input from 0 to 140 volts is better. Do not use more than 140 volts for IGBTs health. D8 = diode bridge 800- 1000 V @ 35 A K = switch for choosing the envelope of the current (half or full- wave rectifyed) C1 = 10 uF @ 400 V non- polarised R9, R10 = 820 ohms @ 5 watts D1, D2 = MUR1100E (ultrafast rectifyer 1000 V @ 1 A, recovery time = 75 nsec) or BYV26E D4, D5 = zenner diodes 12 V @ 1 watt. I used BZX85C 13 R1, R8 = 10 kohms @ 1/2 watt Q1, Q2 = IGBTs : BUP 314 (1200 V @ 52 A), mounted on a heatsink D6, D7 = RHRP15120 (hiperfast diodes 1200 V @ 15 A, soft recovery < 65 nsec), on a heatsink L6 = charging inductor: 74 turns with enamel copper wire, 1 mm diameter, wound on a toroidal core (two layers) made by iron powder (height = 2 cm, outer diameter = 5 cm, inner diameter = 2 cm) 300 microH, 5 amp cooler fan 220 V for heatsink cooling C (primary tank cap) = 134.6 nF @ 1600 V metalised polipropilene. This is more than in the original schema because my resonant frequency of the secondary is 76.2 kHz, while the freq of Vladi's secondary is around 130 kHz. Because the primary has a fixed number of turns you have to tune the circuit varying the primary capacitance. The connections between components have to be kept shortest as possible to prevent parasitic inductances (eg: the capacitor C have to be direcly coupled with the collectors of the IGBTs). Primary coil: 4 + 4 turns with PVC insulated multistranded copper wire with average diameter = 2.5 mm, wound on a plastic basket with average radius = 10 cm (3.93 inches), wound lenght = 3 cm (1.18 inches), at 10 cm above the bottom of the secondary. L calculated for all 8 turns = 20.95 microH, Lreal (with the coils coupled ) = 32.36 microH. The coils AB and CD are wound in the same direction, one turn of the coil AB alternate with one turn of the coil CD, like this: N1(AB)...N1(CD)...N2(AB)...N2(CD)....N3(AB)...N3(CD)....N4(AB)....N4(CD) The middle point (M) connects N1(CD) with N4(AB) N1(AB) and N4(CD) connect each of them with the collectors of the IGBTs The wiring is very compact with no lossy turns. My secondary = N = 3760 turns of magnet wire with 0.125 mm diameter, wound on a white PVC pipe, diameter = 14 cm (2.75 inches), wound lenght = 47 cm ( 18.5 inches), with acrilic end caps glued, not varnished. Lsec = 509728.72 microH, Cself = 8.5494 pF, Fres = 76.2 kHz. Topload: big iron needle fixed on the top of the coil with plasticine Ground: home water pipe |
| NOTE: THE DIMMER HAVE TO BE MOUNTED BEFORE VARIAC NOT AFTER. |
 |