Computer Network Switch Selection for the Outbuildings

I’m using Ubiquiti UniFi network equipment in the House – notably for the PoE Network Switches and the Wireless Access Points. I find the UniFi gear provides a reasonable balance between Enterprise-grade functionality and Consumer-grade pricing.

The CCTV cameras are also from the UniFi product range – although those operate as part of the UniFi Protect sub-system and are largely independent of the underlying network equipment. I don’t currently use a UniFi network ‘Security Gateway’ since I prefer the flexibility of a more extensible router and firewall solution that I can add my own code to (I’ve recently moved to OPNsense) – though when the time comes to sell the house I might swap to one of the more user friendly UniFi gateways (and also replace the self-hosted UniFi Network Controller application, which currently runs in a Docker container on an Gen8 HP MicroServer).

It therefore makes sense to continue to use UniFi network equipment in the Outbuildings, managed via the existing UniFi Network Controller dashboard. The question is: which model from the extensive UniFi switch portfolio to choose?

The basic requirement is for at least 22 wired Ethernet ports, at least 13 of which need Power over Ethernet (for CCTV cameras, Wireless Access Points etc). That requirement maps nicely to a 24-port switch – although two 16-port switches would provide some redundancy and more expansion capacity.

  1. Second-hand US 24 250W PoE?
    • That’s just a scaled-down version of the US 48 500W PoE switches in the house – which are OK but they’re relatively deep, and heavy – and the cooling fans are quite noisy (and would be far from ideal in the more dusty / gritty environment of the Outbuildings)
    • These seem to go for around £175 on eBay – and might have been running 24×7 for many years so are likely to need new fans, if nothing else
  2. Two Gen2 USW-16-POE switches?
    • These are fanless units (good) and have a 42W PoE budget (more than adequate) – but they only have 8 PoE-capable ports, so only16 from a pair – which is barely enough
    • These also seem expensive compared to the 24-port variants – about £275 (inc VAT) each, brand new, so £550 for a pair
  3. One Gen2 USW-24-POE switch?
    • These are fanless and have 16 PoE-capable ports (a much better proportion)
    • They’re about £350 (inc VAT) each
    • The main limitation is they only have regular SFP uplink ports – i.e. 1Gb/s max
  4. One Gen2 USW-24-Pro-POE switch?
    • These have SFP+ uplink ports – i.e. 10Gb/s max – and all 24 ports are PoE+ (or PoE++) capable
    • They have a 400W PoE power budget (much more than required) and they provide Layer 3 capabilities (e.g. a DHCP Server and inter-VLAN Routing) which can also be provided by the Firewall (which is required anyway, for other reasons)
    • They’re over £600 (inc VAT) – which is a big premium to pay for features that aren’t currently necessary for this ‘domestic’ installation

On balance, the USW-24-POE seems the least-bad choice. While the 1Gb/s uplink is a limitation, the network link back to the house will be via two firewalls which also impose a 1Gb/s limit – until those are moved to more modern hardware. Second-hand (but relatively young) examples of the USW-24-POE occasionally appear on eBay – though typically for more than half the ‘new’ price.

If additional ports are required, a second (smaller?) switch can be added later – potentially in a ‘satellite’ location (avoiding the need to run all the structured cables back to the Plant Room).

One mitigation for the 1Gb/s uplink limitation would be to ‘aggregate’ both SFP ports for (a slightly better) 2Gb/s back to the adjacent firewall. Unfortunately, it seems that one of the three 1Gb/s ports on the PC Engines APU2 device that runs this firewall is currently refusing to run at more than 100Mb/s – due to some sort of hardware issue – so it’s stuck with one Gigabit connection to the House and one Gigabit connection to the Switch.

Brick and Tile Choices for the Outbuildings

The concept for the Outbuildings has always been to construct the outside using materials as close as possible to what they would have originally used 100 years ago. This means the buildings will look ‘new’ (as opposed to using reclaimed materials to try to make them look ‘old’) but they will still appear ‘traditional’ – and hopefully not ‘fake’. The contrast with the modern-looking House is completely intentional.

Roof Tiles

For the roof tiles the choice was fairly straightforward: from old photos and examples of broken tiles in the demolition rubble it’s clear the buildings originally had small (10.5″ x 6.5″) clay roof tiles. Some of the examples found are red and marked ‘Rosemary’ whereas others are blue. The old photos kindly shared by a previous owner clearly illustrate that most of the original tiles were Staffordshire Blue – as can be seen in the photo below.

North-West corner of Barn, sometime in 1980s

Using ‘blue’ tiles (rather than ‘red’) also means that the solar panels integrated into part of the roof will be less obvious – especially since the tiles with have a smooth finish, so will start their life looking slightly shiny.

The appearance of the roof is important because: a) there is a lot of it, and b) it will be very visible from the East-facing windows of the House – especially since the Outbuildings are (mostly) single-storey, with a floor level over a metre lower than the ground floor of the House.

There are some concrete roof tiles available which are finished to look like clay tiles; some of these are double-width with a fake ‘join’ in the middle and offer savings on both materials and (especially) on installation labour. However, those firmly fall into the ‘fake’ category – and are definitely not what would have been used originally.

The selected tiles are therefore machine-made plain, smooth clay tiles in Staffordshire Blue from Dreadnought Tiles in Brierley Hill (about 50 miles away).

Facing Bricks

For the bricks the choice was less clear-cut. Again there are old photos and plenty of examples of old bricks around the site which show the original bricks were an orange-red with quite a bit of variation in colour. It’s quite possible the originals were made at the old Chellaston Brickworks (that’s a PDF file link to a pamphlet from the Chellaston History Group) which was less than a mile away.

Based on the walls that were still standing in 2015, it’s evident many of the original bricks had weathered very badly over the years, which makes it difficult to tell what sort of finish they originally had.

Very badly weathered original brickwork on derelict barns in 2015

However, in a few places, there were examples of bricks which were much less badly weathered (next to some that were still quite badly eroded) – which gives a better indication of what the original bricks might have looked like.

Less badly weathered original brickwork around one of the barn ventilation slots in 2015

One brick manufacturer that is very visible within the self build community – especially since they always seem to have a presence at the self build shows – is Furness Brick, based in Cumbria. They still make bricks using traditional coal-fired kilns and hence naturally get colour variations based on where the bricks are located within the kiln. Their website is also very good at listing examples of buildings built using particular blends of their bricks, in case customers want to see a real-world example of what they look like.

The Furness bricks are expensive though – and not always readily available. They also have an even higher carbon footprint than bricks produced in a gas-fired kiln. For matching bricks when extending an original building they’re probably a good option, but for a new, standalone building there are alternatives which offer better value for money.

The main contractor identified the Clayburn Red Civic brick as being a suitable candidate and built a small test panel, using a couple of different styles of mortar pointing. This confirmed these bricks looked like a good match to the ‘best’ of the original bricks visible in 2015. While they might not suit a domestic building, they seem a very good choice for a semi-industrial agricultural building.

Test panel of Michelmersh Clayburn Red Civic Brickwork

The original full planning Application didn’t specify the brick or tile materials and the resultant Approval included the expected Condition that these would need to be approved later. A formal Application to discharge this planning Condition was made in mid-February 2025 and formal Approval was finally granted on 22 April – after a bit of discussion and the sharing of photos of the original brickwork. The planning Approval confirms the choice of ‘bucket handle’ pointing.