For aluminum wire, conductor grade, the conductivity is taken at 60% of copper wire, conductor grade (98%?). Thus the equivalent aluminum wire has to be two AWG numbers larger. Example: #12 aluminum is equivalent to #14 copper, altho it might be hard to find as I believe it is not permitted in 120v house wiring in these sizes.
We did use aluminum wiring in heavy 400 cycle feeders, because the increased surface area helped deal with the skin effect problem. Due to size these were stranded, but I think that added little or no advantage against skin effect. I don't know for sure as I always used tables, and there are no tables for large solid wire (ie, MCM sizes, say 200,000 up).
Ah, tables for skin effect-- they are not common for telephone style comm wires, such at #s 26, 24, 20 because it was not much problem in such small wire. #s 18 and 16 were used for longer lines also, for some of the cable pairs, but still it wasn't too much problem; low capacitance insulation is therefore available in all these sizes, and I would recommend its use on the #16 wires, with DCS signal. (See following,) Note: Larger wire above #16 does begin to get uneconomical of copper due to skin effect, as least in communication work, so you have no tables for it in house wiring sizes #14, 12, and 10.
So I made my own skin effect calculations for #14,12, and 10 relative to the DCS signal. It is really a painful calculation to do by hand, because it must be done in incremental layers. I found that although the skin effect increases, it is more than offset by the added surface area== so that larger wiring does not reduce the distance the DCS signal can travel. Thus the original designers of my club layout had been wrong to limit themselves to #16 wiring, when currents (10-12A) and distance (100' each way in each track output pair) to the TIUs central mounting board (followed by a star pattern).
Here the DCS signal worked fine. In general, I had concluded years before that the signal strength of the DCS signal would be okay to about the length of a football field. The real problem is capacitance in the wiring, between the pair or rails and to common surfaces (foil-backed foam). You see, although inductance is fairly constant with frequency, the capacitance is not. The DCS signal consists of 31 discrete frequencies starting at about 121 kHz, and with 3rd harmonics covering about 10 mHz. this is a range of 100 to 1. Thus the 31 frequencies travel at different speeds, and wih distance they cannot be all properly received. This is due to relative timing drift at the receiving locations.
The longest distances of receiveable DCS signal, fed from a TIU adjacent to the two blocks, that I know of, was 135 feet. One was center fed, and the other end fed. A yard of 9 tracks and comparable total length was also workable, but here programming required a separate block. The track was Gargraves, with DCS on one outside rail only, thus the interrail capacitance of about 40 feet of track could be offset by adding a lightbulb (never installed on a 175 foot yard lead with a dead signal area at one end). Transformers were 4 feet upstream of each TIU (8' max), and distribution boards about the same. The distribution pairs were #14s (IIRC) in multiple, to progressively further points on the block. The progressive connections were spaced at about what is recommended today (IIRC), with the individual pairs not twisted, but draped separately about 1 to 3 inches apart and up to a foot below the table (I'll discuss this and PS3 later). This layout was the test bed for DCS development.
--Frank
