A wireless bridge between buildings is a line of sight problem, not a distance problem

The path has to be clear, and clearer than it looks

Stand at each proposed mount point with binoculars and look at the other end. If you cannot see it, you have a height problem to solve before you discuss equipment.

Seeing the far end is not enough. The radio path is not a pencil line between the antennas. It is a fat shape, widest in the middle and tapering to a point at each end, a bit like a rugby ball lying on its side. That is the Fresnel zone, and anything clipping its edge steals signal even though it sits nowhere near the straight line you can see.

So something you can see over may still be in the way, because the zone dips below your sight line at the midpoint. The longer the link, the fatter that shape gets, which is why a hop across a yard is far more forgiving than a hop across a concession road.

Standard practice is to keep the great majority of that zone clear and add height rather than argue with it. Height is cheap at design time and expensive after the mast is up.

Trees are the obstruction that gets everyone

A brick wall is honest. It blocks the link on day one and you move the antenna. Trees are worse, because they are partial and they change.

Leaves are mostly water, water absorbs at these frequencies, and a canopy in the path loses signal in a way that varies with wind and rain. A link surveyed in March through bare branches works beautifully and degrades through May. Get well above the tree line, or find out who owns those trees and whether they can be topped.

Weather, and what to expect for throughput

Links in the 5 GHz range, which is what most yard-to-yard hops use, are largely indifferent to rain. Snow on the antenna face matters more than snow in the air, which argues for a slight downward tilt. The 60 GHz and millimetre wave radios that carry much more data over a short hop are a different story: heavy rain and wet snow attenuate them noticeably, so usable winter range is shorter than the clear-day figure. More throughput, less margin.

The advertised throughput on a radio pair is usually aggregate across both directions rather than each way. A second building needing internet, file access and a few cameras has room to spare on a modern pair. A building pushing a full backup across every night needs the arithmetic done first. Latency is not the thing to worry about.

The 5 GHz band is licence-exempt, so your neighbour may point a link across the same air. Scan the spectrum before choosing a channel.

The mount matters more than the radio

These antennas are narrow. A beam a few degrees wide has to stay pointed, and a mast that flexes in a February wind moves it off target on every gust.

• Mount to structure, not to a light standard, a vent stack or anything that vibrates. A ballasted roof frame or a bracket into masonry both work.
• Leave safe service access. Somebody has to realign or replace this eventually, in weather.
• Keep the run from switch to radio inside the 100 metre Ethernet limit, using outdoor rated shielded cable with a UV stable jacket.
• Put a drip loop before the wall penetration and seal it. Water tracking down a cable into a building is slower than lightning and more common.

Grounding is the part people skip and regret

You are putting metal on the highest point of a building and running copper from it into your switch. That is a path for surge.

Bond the mount to the building’s grounding system. Fit a surge protector on the Ethernet where it enters each building, grounded to that building. Use shielded outdoor cable and actually bond the shield rather than leaving it floating. A nearby strike that never touches the antenna can still take out a switch at both ends through that run, and usually the firewall with it.

Bonding is electrical work. Bring in an electrician.

When trenching fiber is the better answer anyway

The comparison people get wrong is cost. A short trench across your own gravel yard is often cheaper over ten years than a radio pair that needs a service call after an ice storm. A trench crossing a municipal road is not. Get a number for the civil work before you assume wireless wins.

What this means for your building

A common one around Bowmanville and Uxbridge: a shop or barn 150 metres behind the main office, same property, gravel yard between, and nobody wants to trench because trucks use that yard all day.

The link is easy. The problems are the row of mature maples along the fence line that will sit in the Fresnel zone by July, and a barn roof with no structure worth mounting to. Both cost money that does not appear on a radio datasheet, which is why a site visit happens before a quote.

What to do about it

1. Stand at both mount points and photograph the view towards the other, from the actual height if you can do it safely.
2. Measure the distance on a satellite map and note everything in the path: trees, poles, roof units, a neighbour’s parapet.
3. Check who owns anything in that path. A tree on someone else’s property is a negotiation, not an engineering problem.
4. Get a price for trenching the same route even if you are sure you do not want it.
5. If the far building is leased, confirm with the landlord what you may mount and penetrate.

The first four take an afternoon and no help. A wireless bridge installer earns the fee on height calculation, spectrum scanning, mount design and grounding. Where a bridge is the wrong answer, fiber optic cabling between the buildings is the conversation instead.