|
Radlab /
KakPowrKAKAAKO POWER PICTURES Edited 6/28/2014  Fig. 1. The main box is on plastic conduits, so it is apparently grounded only by the mains neutral. To isolate ourselves from this ground, we need to insert a PVC fitting (or a whole PVC conduit) between our enclosure and the main breaker.  Fig. 2. Here,as at many other places, some clamp-on ferrite filters are in order.  Fig. 3. All of the "green grounds" and neutrals are tied to this box. They are isolated from our grounds because we do not use the neutral into our isolation transformer, and our neutrals derive from the transformer secondary and are connected to our ground rods.  Fig. 4. This filter at prevents RF noise from getting into the room on the two hot phases. The neutral noise needs to be addressed. This neutral is filtered, but it is referenced to a ground at some remote unknown location. These neutrals should thus NOT be connected to the local ground. Being "RF-clean" is enough.  Fig. 5. Note that unlike the "normal" wiring the neutral is not grounded to the box. The conduits in our system serve only as shields and "green" safety grounds. The transformer secondary is 120 volts, and both sides are interrupted simultaneously. The neutral will be connected to earth ground within our system, but that will carry no current since the transformer secondary is floating. The conduits should also be grounded at only one point.  Fig. 6. The upper boxes are isolated and filtered power for science and communication. The air conditioning and lighting cconduits are isolated from building ground by the plastic conduit outside of the room. It may be convenient to replace this plastic box with a quad outlet box since that provides convenient screw terminals for branching the wiring. The conduits and boxes should be metal up to the lighting conduit that goes into the ceiling.  Fig. 7. The separate outlets for Kilo Nalu and HFDR should be mounted fairly far back to minimize the exposed wiring going into the back of the rack.  Fig. 8. The air conditioning and lighting are filtered to a lesser level that the science power. Mounting the box sideways allows for two conduits out of the top. If we install LED lighting, we should ensure that it is RF-quiet since some use HF switching blalasts.  NOTE: the power rail should be connected to the KN UPS, not to the wall outlets. Fig. 9. It will probably be best to cut the cord and put a clamp greenfield clamp on the box to hold the cable. It can be screwed onto the outlet terminals. The plastic conduit that runs all of the way up to the ceiling would be difficult to change. It is also a large (and high) RF "loop antenna" that would put a lot of RF on the outlet grounds on the left side. A shorter (lower) route over the door (Fig. 8) would be less troublesome since it cuts the cross-sectional area of the loop nearly in half. Eliminating the vertical conduit on the other wall (Fig. 6) also simplifies things there.  Fig. 10. Schematic drawing of filters.  Fig. 11 System Grounding Phil was only able to drive one ground rod. It is inside to the left of the door, and the ground wire is far too long.  Fig. 12 There are only a few wires I could not get connected. All nuts need to be tightened with a ratchet.  Fig. 13  Fig. 14 I erroneously wired the filter in the upper right to the isolated circuit instead of the utility circuit. |