Salient Soft

I am running 2 x CAT6 SWA cables to a garden office, along with a 6mm SWA cable for mains power.

I checked widely online about the wisdom of going with CAT6 SWA UTP vs FTP or S/FTP, and opinions as to how well this will all work vary widely – it was very hard to get a clear consensus on performance, particularly re the issue of how close you can run the CAT6 cable to a mains cable without compromising performance, although a number of posts cited not having any problems in practice. I had already opted for SWA UTP for simplicity, having previously reckoned that CAT5e might even be enough. The plan was to put a gigabit switch at each end, which should give far more performance than actually needed – The need to shovel large amounts of data between the house and the garden office would be minimal – possibly occasional copies to/from the TV pc, and some RDP use from the house to an office PC. The main need is to facilitate connection to the internet, which is a moderate speed FTTC connection to Zen Internet, typically giving about 30Mb/s down and 6-7Mb/s up, after significant effort by Zen and OpenReach to solve some line/connection problems. Slower internet is the price for living in a quiet rural area!

Having already obtained the cable, I was reluctant to switch and also have the additional size and larger bend radius of S/FTP, plus the additional complexity of the correct earthing and screened keystones etc.. I planned to just ask the installer to try to maximise the distance between the network and mains runs of cable, but was concerned enough at a late stage to try a simple test.

I ran an old cat5 UTP cable of around 15m in length as close as possible to a long mains extension, such that for around 7-8metres the 2 cables were tightly adjacent all the way along. I then connected one PC to a switch via this cable, and then ran a lan speed test from this PC to another target PC on the same switch, using LanBench on both PCs, a simple open source lan tester. I ran various tests both with the mains lead unplugged, plugged in, and actively powering a 2KW fan heater, all while the test was running. Whilst initially I did get some significant variation in performance, after experimenting I concluded that this was not due to the mains cable – some other factor was at play. In the end I ran with 2048 byte packets, and 4 simultaneous connections – it was upping the connection count to 4 from 1 which gave better and more consistent performance all round.

The results gave a send rate of approx 900-1000Mb/sec, and a receive rate of 250-300Kb/sec. Crucially, this level of performance was the same whether or not the mains was connected or in use during the test, and over a significant number of repeated tests.

This was not a definitive test, and it was over a much shorter distance than the SWA cable. The latter will likely be around 40m, but hopefully with minimal closeness to the mains cable compared to this test. The test did however rebut a significant number of posts online which said on no account must the cables be closer than a few inches and even if perpendicular you could get issues. Whilst the official rules might in theory mandate tighter restrictions to ensure all edge cases are covered, this simple test did give reassurance. In addition, the SWA cable would be CAT6 rather than CAT5, and the earthing of the SWA shield should go some way to help with screening interference at mains frequencies as well, in addition to the balanced line working of the twisted pairs which would eliminate common mode interference.

Comments Off on Running a network cable to a garden office