Somebody in the back office says the Wi-Fi is terrible. You walk back there, look at your phone, see four bars, and figure it seems fine. So you order an extender, plug it into a hallway outlet, and for about a week things feel better. Then the complaints come back, and now they come from people who never had a problem before.
We see this constantly in Denton County offices, warehouses, and clinics. Bad Wi-Fi is almost never a product problem. It is a design problem. The gear you can buy today is genuinely good, and it will still perform badly if it sits in the wrong place, on the wrong channel, doing a job it was never built for. Design problems have real fixes, and those fixes usually cost less than the pile of gadgets people buy while guessing.
More Bars Is Not More Speed
The bars on your screen measure one thing: how loudly your device hears the access point. They say nothing about speed, about how many devices are competing for airtime, or about whether the access point can hear your device answer.
That last part matters more than people expect. Cisco’s wireless RF reference documentation notes that client maximum power is 6 dB below the access point’s maximum power. Picture someone with a megaphone at the far end of a warehouse. You hear them perfectly. They cannot hear you answer, because you do not have a megaphone. Your laptop is the one without a megaphone. Full bars and a dead connection can both be true at once. That same documentation points to -66 dBm as a standard value to survey for when checking client coverage. That is an engineering target, not a vibe.
What Walls and Metal Actually Do to a Signal
Signal loss is measured in decibels, and decibels are not a straight line. Roughly every 3 dB of loss cuts the power in half. That is why two walls flatten a strong signal faster than your intuition expects. Cisco’s RF reference documentation puts approximate numbers on common materials:
- Drywall: typically about 3 dB. One interior office wall is a modest hit.
- Brick: around 10 dB. A serious wall, and older buildings around the county have plenty of them.
- Concrete: around 12 dB, and Cisco notes concrete tends to contain more metal, which makes it worse.
- Everything else: Cisco is clear that attenuation depends on metal, moisture, thickness, and conductivity, so no chart replaces measuring your building.
Metal behaves differently. It does not politely absorb signal, it bounces it, and those reflections reach your device slightly out of step with the original. That is multipath distortion. Cisco’s wireless site survey guidance names multipath distortion, hidden node problems, and near/far issues as exactly the things a survey exists to find. In practice: racking, metal studs, elevator shafts, walk-in coolers, and metallic window film cause failures that look random until somebody measures them.
Why Another Extender Usually Makes It Worse
A wireless extender has one radio doing two jobs. It listens to your router, then repeats what it heard to your laptop, over the same airwaves. Think of a meeting where an interpreter repeats every sentence before the next person can speak. The meeting still happens. It takes about twice as long.
The second problem is less obvious and more damaging. Cisco’s RF guidance says access points should not hear one another on the same channel above -82 dBm, because past that they take turns instead of working in parallel. An extender only functions if it hears your router clearly, which means you just installed a second device parked in the middle of the first one’s channel. You did not add capacity. You added a competitor.
People also reach for 2.4 GHz because it travels farther, and Cisco explains why: the radio wave is physically bigger, so it takes less power to cover the same distance. Reach and capacity are not the same thing. The band that goes farthest is also where every neighbor and every microwave is already shouting. This is the hidden cost we wrote about in why saving time takes priority over saving money. A fix that eats twenty minutes a day from eight people was never cheap.
Coverage and Capacity Are Two Different Problems
Coverage is whether a signal reaches a spot. Capacity is how many devices can get real work done in that spot at the same time. Wi-Fi is a shared medium, like a conference call: one device talks at a time on a given channel and everyone else waits. Adding talkers does not add line.
Cisco’s general density guidance lands around 1,200 to 2,000 square feet per access point, roughly one every 36 to 44 feet. For voice traffic, Cisco’s site survey guidance recommends no more than about 10 calls per access point and no more than 15 to 20 percent overlap in cell coverage. Notice that the voice number counts devices, not square footage.
A warehouse aisle and a training room holding 40 devices are opposite problems. The warehouse needs reach. The training room needs several smaller, quieter cells with tightly controlled power. One access point in the hallway between them solves neither.
When a Site Survey Is Worth Paying For
A site survey is not a sales ritual. It is somebody walking your building with measurement tools, mapping what the radio environment actually looks like. Here is where it earns its keep:
- Warehouses and shops: a full rack and an empty rack are different buildings as far as radio is concerned.
- Multi-floor offices: signal does not stop at the ceiling. Floor two interferes with floor one, and nobody notices until both are slow.
- Rough environments: Cisco’s guidance notes access points in extreme moisture, temperature, or dust may need a sealed enclosure.
- Before you spend real money: a survey costs a fraction of the hardware and tells you how much you actually need. Sometimes that is less than you were about to buy.
The Wired Backhaul Most Consumer Setups Skip
Here is the biggest structural difference between a business wireless network and a good home setup. In a business design, every access point has its own Ethernet cable running back to a switch, and that cable usually carries power too. No wall wart, no extension cord, and the access point goes where the radio design says rather than where an outlet happens to be.
Mesh and extenders are a bucket brigade. Every hop hands the water to the next person, and every hop costs you. Wired backhaul gives each access point its own pipe.
Cabling has rules worth knowing before you promise a mounting spot. The ANSI/TIA-568 structured cabling standard allows a maximum of 90 meters of installed horizontal twisted-pair cable, with 100 meters total including patch cords. Your network closet location is a design input, not a detail. If the far corner of the warehouse sits past that limit, the answer is another closet or a fiber run.
The Bottom Line
Wi-Fi that works is designed, not purchased. The questions that matter are where the access points go, how much power they use, which channels they sit on, how much they overlap, how many devices land in each cell, and whether each one has a wired connection back to the switch. None of those get answered by a box on a shelf.
If your network today is a router, two extenders, and a standing complaint, you do not have a hardware shortage. You have a layout nobody ever mapped. That is the same logic behind bringing in an outside IT partner: you are paying for measurement and judgment, not just parts.
We design and install business wireless networks across Denton County, from single-suite offices to multi-floor buildings and warehouses. If your Wi-Fi has a spot where work stops happening, we will survey it, tell you what is actually causing it, and fix the design instead of selling you another extender. Contact us today.
Sources:
Comments are closed