‰  NOW 40 WPM  ‰  
CAPACITOR FILTER.  TWO RELAYS FORM THE SOFT START CIRCUIT.  THE FIRST RELAY
IS CLOSED BY THE HIGH VOLTAGE ON CONTROL.  THE SECOND RELAY IS CLOSED BY
FEEDBACK FROM THE SUPPLY ITSELF BY USING THE CURRENT THROUGH THE BLEEDER
RESISTOR.  60 KHZ SWITCHING THE 320 V DC IS APPLIED TO A PAIR OF SWITCHING
MOSFETS OPERATING AT A FREQUENCY OF 60 KHZ THROUGH A CENTER TAPPED
TRANSFORMER PRIMARY.  EACH OF THE THREE SECONDARY WINDINGS PRODUCES A PEAK
OUTPUT OF ABOUT 1150 V.  RECTIFIER/SUMME

DISPLAY A NON MONOTONIC TRANSFER CHARACTERISTIC IT WILL BE DOUBLE VALUED AT
SEVERAL POINTS PULSE WIDTHS.  HENCE, TO BE PART OF A CLOSED LOOP SYSTEM,
THE SYSTEM WOULD ALWAYS DEMONSTRATE AN INSTABILITY.  AFTER MANY BLOWN
SWITCHING TRANSISTORS, USING SPICE, I CAME TO AN AGREEABLE UNDERSTANDING A
TRUCE.  WHOLE CYCLE SKIPPING WOULD BE MUCH THE SAME, AND WAS DEEMED CLUMSY
AND IMPRACTICAL AS A CONTROL MEANS.  THERE WAS NO GUARANTEE OF STABILITY
WITH THIS METHOD, SO I DIDNT TRY IT.  THE REAL EFFORT HAD TO BE TO IMPROVE
UPON THE TRANSFORMER UNTIL IT WOULD BE GOOD ENOUGH TO PROVIDE SATISFACTORY
REGULATION IN AN OPEN LOOP APPLICATION, ACCEPTING THAT PERFECTION IN
COUPLING WAS IMPOSSIBLE.  THE HIGH VOLTAGE TRANSFORMER THE EXTREME
TRANSFORMATION RATIO REQUIRED FOR THE HIGH VOLTAGE OUTPUT RESULTS IN A
TRANSFORMER HAVING A POOR COUPLING COEFFICIENT, AND IN TURN, HIGH LEAKAGE
INDUCTANCE AND PARASITIC CAPACITANCE.  AS A RESULT OF THESE PARASITIC
CIRCUIT ELEMENTS, THE SYSTEM IS GREATLY UNDER DAMPED WITHOUT A LOAD ON THE
POWER SUPPLY, AND AN OSCILLATION TAKES PLACE AT THE END OF EVERY SWITCHING
HALF CYCLE.  THIS OSCILLATION, A DECAYING SINE WAVE, CAN CAUSE A VOLTAGE

HIGH VOLTAGE DESIGN THEORY  THE SIGNIFICANT DESIGN PROBLEMS WERE LIMITED TO
THE HIGH VOLTAGE PART OF THE TOTAL POWER SUPPLY.  THEY WERE, FIRST AND
FOREMOST, LIMITING OUTPUT VOLTAGE VARIATION DUE TO VARYING LOAD SECOND,
VOLTAGE BREAKDOWN OF SEMICONDUCTOR PARTS THIRD, CURRENT HANDLING CAPACITY
OF SEMICONDUCTOR PARTS, AND FOURTH, KEEPING HEAT TO A MINIMUM.  TAMING
OUTPUT VOLTAGE VARIATION LIMITING OUTPUT VOLTAGE VARIATION PROVED TO BE BY
FAR THE MOST DIFFICULT DESIGN CHALLENGE.  MY FIRST APPROACH WAS TO USE A
CLOSED LOOP FEEDBACK CONTROL OF THE OUTPUT BY REGULATING DUTY FACTOR PULSE
WIDTH DRIVE TO THE SWITCHING TRANSISTORS.  A TRANSFORMER IS A FAR MORE
CRITICAL MAGNETIC COMPONENT WHEN COMPARED WITH A SINGLE WINDING INDUCTOR
THAT MIGHT BE USED IN A BUCK OR BOOST TYPE OF SWITCH MODE SUPPLY.  MUCH
EFFORT WAS PUT INTO FEEDBACK CONTROL CIRCUITS AND CONTROLLER CHIPS.  WHAT
THAT WORK REVEALED WAS THAT NO FEEDBACK SYSTEM COULD GET AROUND THE
PROBLEMS OF A LESS THAN PERFECTLY IDEAL TRANSFORMER.  SPICE SIMULATED
PROGRAMMING WITH INTEGRATED CIRCUIT EMPHASIS MODELING SHOWED THAT WITH A
LESS THAN PERFECT TRANSFORMER COEFFICIENT OF COUPLING K LESS THAN 1 AND
WITH THE RESULTANT LEAKAGE REACTANCES IN ALL THE WINDINGS AND PARASITIC
CAPACITANCES BETWEEN WINDINGS, FEEDBACK CONTROL COULD NOT BE ACHIEVED WITH
DUTY FACTOR VARIATION AS THE MEANS OF CONTROL.  THE TRANSFORMER WILL ALWAYS
‰  END OF 40 WPM TEXT  ‰  QST DE W1AW  ƒ