ALOHA Cabled Observatory
|
ALOHA Cabled Observatory
|
|
|
Aloha Cabled Observatory Engineering Documentation A/D Patch 2 Board
Power System Collapse
We are accustomed to working with voltage sources. As you increase your load, the voltage drops slightly. At full load there may be a moderate voltage drop, depending on the source. Under overload conditions the current continues to rise and the voltage drops further. Under short-circuit conditions, the current may become extreme and cause catastrophic damage. Most circuits include a fuse or breaker to handle short circuit faults. Things are profoundly different with a current source. In order to operate devices designed for voltage source operation, we must first add a shunt regulator to locally create a low impedance voltage source. In this case as we increase our load the voltage drops only slightly, depending on the shunt regulator. The total system current remains constant and power remains constant. Excess power (or current) is dissipated by the regulator. As load current approaches maximum, shunt current goes to zero. At this point there is an abrupt discontinuity in the system operating characteristics: the voltage is suddenly unregulated and begins to drop precipitously. The switching converter sees this drop in its input voltage and attempts to compensate for it by increasing its current demand. This instantly exacerbates the situation and causes the voltage to drop even faster. Any other converters on the bus do likewise. The result is that it appears that source gets into trouble, all of the users “gang up” on the supply. Unfortunately, this transition is abrupt and without warning. Once initiated, such a collapse is likely to go all the way down instantly. What is needed is some sort of “ESP” that can predict when a collapse is about to happen and stop it before it starts. The best way we found was to monitor the Shunt Regulator current. If it drops below some preset level then we know that the system is headed for a potential collapse. The action which we selected is to immediately drop the Users’ B-Bus. We are aware that some users may consider that action extreme and it will disrupt the experiments. The answer is that the bus was headed for an uncontrolled collapse anyway. This approach leaves the A-bus and the communication to shore intact to keep the Observatory itself under shore control. The implementation of this rule is on the ADpatch2 board. Note that if either bus is about to collapse, the B-bus will be tripped. Again, the rationale is that a “functioning” User B-bus would not be useful without a functioning communication channel. The actual trip levels are yet to be set pending system testing and further discussion.
Rev: 08/10/2007
|
|
|
