LAN/Ethernet cable for connecting to active loop aerial

Introduction

As loop aerials are highly directional, their orientation needs to be changed by either rotating the loop or by switching to another loop that is orthogonal to the first one (fig. 1). Additionally, the loop's resonance frequency may need to be remotely changed with either a motorized variable capacitor or tuning diodes.  Hence, a loop may require several wires for controlling purposes. 

Instead of using a coaxial cable to connect an active loop aerial to a receiver as is the convention, LZ1AQ proposed using a shielded LAN/Ethernet cable [1]. The LAN cable has 4 pairs of twisted wires (8 wires) and some of these extra wires can be advantageously re-purposed for controlling rotator, servo motor, tuning diodes and/or relay switches. Hence, one LAN cable can conveniently replace the coax and several control wires (fig. 1).

However, there are two technical hurdles to adopting the LAN cable. Firstly, the loss is unknown; i.e. LZ1AQ did not elucidate on this. Secondly, as each twisted wire pair has a 100 Ω characteristic impedance (Zo), hence, the SWR will be awfully ≥ 2 in 50 Ω system. Hence, the combination of unknown loss and mismatch makes many reluctant to embrace LAN cables. 

This article proposes a solution to the LAN cable's inherently poor matching /SWR and presents detailed measurement of the loss.

Fig. 1  Via the RJ-45 jack, one LAN cable can conveniently serve the loop aerial's signal and control needs, e.g.  for controlling aerial rotator, tuning diodes and/or relay switches (left). Photo of LZ1AQ's preamp with the LAN cable and RJ-45 jack visible at the bottom of the enclosure (right)

Measurement of the a twisted pair confirmed the impedance mismatch; I.e. the centre of the Zo locus is offset to the right of the Smith chart centre (fig. 2). To counter the high SWR, LZ1AQ proposed using a 2:1 balun, but this incurs extra cost and loss. Moreover, the 2:1 impedance ratio is uncommon.

Fig. 2:  Using a single twisted pair (Zo = 100Ω) per LZ1AQ will detrimentally offset the centre of the Zo locus to the right of the Smith chart centre, hence causing severe mismatch

Measurement of the single wire pair reveals unacceptably poor loss and matching in the 3-30 MHz HF range (fig. 3). The insertion and return losses are ≥ -6 dB and ≤ 3.8 dB (swr 4.6), respectively.

Fig. 3: The single twisted pair has poor loss (≥ -6 dB) and return loss (≤ 3.8 dB) in the 3-30 MHz HF range 

Material & method

In theory, connecting two twisted pairs in parallel can beneficially halve the rated 100Ω Zo to 50Ω. To enable the parallel connection, pins 1 & 3 on the RJ-45 jack are joined together and then connected to a SMA connector’s centre pin via a ~2 cm long hookup wire. Likewise, the RJ-45’s pins 2 and 6 are joined and then connected to the SMA ground tab (fig. 5). These SMA connectors enable the LAN cable to be connected to a network analyzer for measuring RF performances. 

Fig. 4: A PCB legend indicating the RJ-45's pin numbering

Fig. 5: To connect two twisted pairs in parallel, pins 1 & 3 on the RJ-45 jack are joined together and then connected to SMA connector’s centre pin via a ~2 cm long hookup wire. Likewise, the RJ-45’s pins 2 and 6 are joined and then connected to the SMA ground tab

A LAN cable length of 15m was chosen as representative of the typical home cabling (fig. 6). Two identical cables were tested; I.e. n = 2. These are the cheapest cables we can find on an online shopping platform. 

Fig. 6: Measurement setup

Results

The proposed intervention of connecting two twisted pairs in parallel beneficially shifts the measured Zo locus to the centre of the Smith chart (fig. 7). Consequently, the insertion loss and return loss are also significantly improved compared to the single twisted pair. 

Fig. 7: The proposed intervention of connecting two twisted pairs in parallel beneficially shifts the measured Zo locus to the centre of the Smith chart

The measured insertion loss is better than 2.6 dB over 3-30 MHz (fig. 8) - a value that can be easily compensated by the loop pre-amp. Hence, the worst-case loss has improved 3.6 dB over the single pair.  The  new loss value corresponds to a normalized loss of ~0.2 dB / meter. 

Fig. 8: The proposed intervention improved the  insertion loss to better than 2.6 dB over 3-30 MHz. Sample size, n = 2.

Two wire-pairs result in a measured return loss of better than 14 dB (swr 1.5) over 3-30 MHz (fig. 9). This represents >10 dB improvement over the single pair.  Therefore,  the original requirement for a  2:1 balun is obviated.

Fig. 9: The return loss is better than 14 dB over 3-30 MHz. Thus obviating the need for the 2:1 balun. Sample size, n = 2.

Discussion

On the flipside, the RJ-45 connector is not waterproof, and the cable jacket is not UV resistant. If used indoor, a shielded cable is necessary to prevent electrical noise pickup. 

 

Conclusion

The LAN/Ethernet cable is suitable for connecting active loop aerial to HF receiver. By using a two wire-pairs instead of the specified one, its performances can be greatly improved.


References

[1] C. Levkov, “Wideband active small magnetic loop antenna”, 2011. [Online] Available: http://www.lz1aq.signacor.com/docs/wsml/wideband-active-sm-loop-antenna.htm