| Gunnar Wik introduced Allen & Bradley HV-primary instantaneous-overcurrent-trip relays for the Litton 30 KV line. I call it smart-logic. Now if all else fails, my high-voltage trips off mechanically, whenever shorted. |
| We connect to buss-bars with suitably-rated breakers and fuses. Step-start resistors are bypassed after two-seconds. An overtemp-switch above each resistor-group signals trouble for such unlikely events. I did see one such switch interrupt a burn-in once. A loose conection left a step-resistor in circuit. The fault read: step-start overtemp. I'd absentmindedly forgotten about that sensing. One look at the step-start resistors revealed a loose screw, a burnt lug, and a wire with melted insulation. This is the beauty of computerization: we enjoyed newfound and robust data acquisition and logging capability. I used three CPU/relay interface boards in my relay-logic (we call them RC/Status-readback boards). Each board has 16 relays and 32 ea. DC-sense-inputs. Relays are remotely triggered, then pulling for one second. Water, oil, and air overtemp switches feed board inputs. Similarly, Deitz brand air-pressure switches signal internal fan-interlock clearance. Flow, pressure, and temp interlock switches keep tabs on floor-to-ceiling manifold ports, catering the many circulation loops. Raw industrial-chilled-water caters all but the beam pass tube where substantial anode-voltage necessitates DI water usage (to minimize corrosion and leakage current). I bypass my step-start time-delay relay-contacts at timing-cycle completion. I do so because I had a time-delay-relay dog me once. Whenever we arced our collector terminal (on SN 722) the regulated system-480V contactor mysteriously dropped out. Then the step-start-resistors started glowing. The transient made the time-delay relay latch up, to then require a power-down reset. At a loss, I installed a glass tube type (heated bimetal) time delay into the face of the formerly solid-state time-delay relay (where the knob once went). Bypassing of time-delay-relay contacts at main-contactor closure effectively sidesteps the above. |
| Allen & Bradley instantaneous overcurrent-trip-relays interrupt the high-voltage-transformer-primary contactor-coil for all crowbar faults. Step-start eliminates turn-on inrush surge. HV contactor-coil voltage originates through seriesed normally-closed instantaneous trip-relay contacts. Interruption also unlatches a Primary-AC-OC-fault relay. Resetting of the fault relay commences courtesy of isolated reset-relay contacts, poised to bypass OC-relay latch contacts. When latched, the OC fault relay contacts pass instantaneously-interruptable-120-VAC on to "beam-on" relay contacts, for use in energizing the high-voltage contactor coil. Chosen for AC-OC fault-relay service, the Potter & Brumfield KBP-series shaded-pole AC-relay unlatches in 2.8 Milliseconds. Dropping-out faster than the HV-contactor is what qualifies it for the job. It's ten times as fast in this case. Many relays are too slow. |
| Cookie-cutter AC-distribution had a cool feature worthy of preservation: the three-phase power came in through the first system contactor. Then it went through a second "high-voltage" contactor. So both had to pull for high voltage to appear. Application of high voltage "by accident" plagued ETM at Hughes in Torrance CA. ETM built five ea. 30 KV test-sets for Hughes without benefit of cookie-cutter contactor wiring. Then a Gould brand contactor lost one tiny spring (from above one of the contact-making plates). The contact-plate flopped around until finally landing sideways. Load-current spotwelded it down (and with it, the whole plunger). So the contactor delivered unconditionally. The operator turned the unit off and reached in. He got bit badly. Those units were dogs. Each subjected a single ignitron to 30 KV. Steve Warner got duped into testing those things. His name was mud again and again. Once his high-voltage compulsively went on when his mechanical crowbar jammed into the up position. This somehow pulled his HV contactor. Again someone got bit. Again, the unit appeared dark. Steve's auto-reset module also failed so as to blindly energize his cathode supply--more than once. This went on until Hughes scrapped those units. |
| The above 30 KV test set stood beside a 26-KV Stangenese test set. Magna made PFNs to match the electrodes of a high-peak-current coupled-cavity-TWT. The pulses looked square. PFNs are intrinsically soft: impedance limited. No crowbar is needed. None of us knew Magna even did test sets. |
| Al Johnson's first 100 KW-unit also had the high-voltage contactor come-on spontaneously at Hughes. A hose blew off at the collector pass tube anode. Those fittings dogged Al no end. This time it drenched the control-signal connector on an umbillical leading to the AC distribution bay. 24 VDC signal wires leading to solid-state relays in the AC bay got enough stray leakage current to turn on our main high-voltage contactor. The mechanical-crowbar dumping-resistors torched, along with adjacent red-glass. We were the laughing stock. Once again, an ETM unit looked like it was off--while seething with high-voltage; billowing smoke and steam; and gushing water. Flooding trumpeted our clout. We finally machined custom captive hose fittings for interfacing with those Eimac 100 KW tetrodes. The stock anode water-fittings compress a rubber ferrule around tubing we were to provide. This might have been designed to facilitate connection to captive glass tubing. Al's first adaptor wasn't designed to be captive. |
| Line filters are a must for the low-noise environment. I did a stint at Teledyne detailing a low-noise test set. The only major noise source was its line voltage. Solidly-bolted Line filters at the power-entry point took care of it. |
| The problem with line filters is they have big capacitors to ground inside them. I've seen five-amps of ground current. And this was only the inbalance-current (from a three-phase filter-bank). This makes it risky to simply pull a metal 480-VAC-plug from an outlet. If the ground connection disconnects first, you may get an arm-to-arm 273 VAC shock. This is deadly science, only to be used with great caution. |
| Caution: Line filters on floating-wye 480 VAC systems can retain the peak-voltage present at disconnection. Residual DC-Voltage lurks unannounced--well after de-energization. At 50 Microfarads per phase, the stored energy can pop loudly when shorted. It feels counter-intuitive but do it anyway. Wear protection, stand out of ionization-blast range, and short before you touch. |
| Ground is not a current carrying conductor. |
| I trust a relay to remember whether high-voltage should be off or on. TTL latches proved fickle handling this responsibility in similar designs. I liked the traditional 24-Volt beam-on relay. What the control-head operates in my system is an overglorified cookie-cutter "beam-on" relay. |
| +24 VDC passes through our door-switch interlock, body-current trip, crowbar-armed relay, HV-primary-OC, and our shorting-relay-upswitch to enable our remotely-operable beam-on relay. |
| True or False: |
| Glo-Cone heaters stood in to do internally- regulated-480-VAC system step starting. |
| Ohmite Cor-rib resistors stand in for cathode-supply step-starting. |
| All Variac wipers get fuses. |
|
| AC Distribution |