• Print
• .

KakPowerKNO

Kilo Nalu Power

Edited 06/25/2014 2:00pm

Contact: Dave Harris cell 779-5436

The AC power for the Hawaii Oceanography Lab radar and the Kilo Nalu Observatory is on an isolation transformer with 3kW total capacity. There are separate secondary windings for each system. There is a 30 amp RF filter on the input of the transformer and and 20 amp filters on each secondary. There are separate circuit breakers and separate outlets. This provides multiple-stage filtering plus isolation for RF coming in on the main power lines.

Fig. 1. Schematic drawing of filters.

Beyond this point, each system will be responsible for their own local filtering, grounding and ground loops.

Phil's group was able to drive one additional ground rod through the floor, which will be connected to the base of the rack with a heavy braided strap to minimize inductance. Given that this conductor carries the sum of all of the ground loop currents from all sources -- originating from a multiple high power sources (frequencies) -- this inductance should be kept as LOW as possible.

The RF coming directly into the room by radiation (and picked up by local wiring and hardware) is a separate issue. There may also be RF coming in on water pipes. An obvious possible problem is the incoming Kilo Nalu cable and the RF coaxial cables. These may all need clamp-on ferrite chokes and solid grounds at the entry points. (See Fig. 2)

Fig. 2. Schematic system grounding.

Note the fundamental scheme for reducing RF ground-loop problems. Placing multiple clamp-on ferites on the antenna cables and the sea cable introduces a relatively high impedance for stray RF currents from the 25kW of AM broadcast power. This has no effect on the signals on the inner conductors. The sum of all of those currents must flow into the system ground, so they must individually be minimized and the inductive and resistive impedance of the ground strap should be kept to an absolute minimum.

It is also interesting to note that the clamp-on ferrite beads used for RFI suppression are actually very "poor quality" material. With rising frequency the inductive reactance of course rises. But before long the eddy current core losses catch up and at all higher frequencies they appear simply resistive. These losses dissipate the RF energy within the core itself. A "good quality" inductance could allow large RF voltages to develop across the component, but these cores become more of a "black hole for RF energy". ;>)

I have a clamp-on current probe for measuring these RF loop currents.