Wi-Fi Design and Access Point Placement
How many access points does a project actually need?
There is no square footage rule that survives contact with a real building. In homes, construction materials drive the count more than size. A 4,000 square foot house with plaster and wire lath walls can need more access points than a 6,000 square foot house framed in drywall.
In commercial and high-density spaces, the driver flips. Materials still matter, but the binding constraint becomes how many devices are active at once, since each access point serves a limited number of clients well. A restaurant or hotel lobby can need access points spaced for capacity in an open room where coverage alone would suggest fewer.
Rather than estimate either one, model it. Send floorplans through the intake form and the predictive design comes back with the count and the placement.
Should access points be ceiling mounted or wall mounted?
Ceiling mounting is the default in most spaces, because access point antennas are designed to radiate down and outward. Wall mounting is the right call in specific situations: very high ceilings, rooms that need directional coverage rather than broad coverage, and retrofits where there is no ceiling access.
Hotels, MDUs, and any building made of small enclosed rooms are the common exception. There, the goal is consistent signal inside each room rather than seamless coverage across an open floor, and wall-mounted units are often the better fit.
One placement to avoid in any environment is near an exterior wall, which sends a meaningful share of the signal outside the building and creates an unnecessary security exposure.
Can an access point be hidden above a ceiling or inside a cabinet?
Often yes, but it depends entirely on what surrounds it. Wood, drywall, and standard glass have minimal effect on Wi-Fi. Metal, foil-backed insulation, ductwork, and plaster with wire lath do not. An access point hidden above a ceiling with ductwork running above it loses coverage the design assumed was there.
- Wood cabinetry with no metal lining: usually fine
- Above a ceiling in a purpose-built mount that preserves the antenna pattern: usually fine
- Above a drop ceiling with pipes, ducts, or plenum equipment: expect losses
- Network closet with a metal rack and a solid door: almost never fine
- Soffits, recessed cavities, or behind metal fixtures: expect losses
Concealment is a legitimate design requirement, not a compromise. It just has to be in the model before install day. Paramount mounts exist for exactly this: hardware stays out of sight without putting the radio somewhere that costs coverage.
Why does Wi-Fi still drop after adding more access points?
Adding access points without a channel plan puts radios on overlapping channels at full power, where they compete for airtime instead of adding capacity. At the same time, clients hold onto a distant access point rather than roaming to the closer one. More hardware in the same space can make performance measurably worse.
Three causes account for most cases: co-channel interference, transmit power left too high, and devices that refuse to roam because minimum data rates were never set. All three are configuration problems, which is why adding another access point tends to deepen them. The effect is sharper in dense commercial spaces, where there is less room to separate channels.
Full breakdown in our AP placement guide.
Which building materials cause the most Wi-Fi problems?
Metal is the worst offender: metal studs, foil-backed insulation, ductwork, metal roofing, appliances, and mirrors. Concrete, brick, and plaster with wire lath come next. Drywall, wood, and standard glass pass signal with minimal loss. Water counts too, which is why aquariums matter more than people expect.
Two catch integrators out regularly. Low emissivity and solar control glass blocks signal about as well as a wall, while ordinary glass barely registers. Radiant barrier insulation behaves like a sheet of metal. In commercial buildings, add elevator shafts, fire-rated partitions, and poured concrete floors, which effectively split a building into separate RF environments.
None of this appears on a floor plan. That is why construction detail belongs in the design conversation and not just the drawing.
What is a predictive Wi-Fi model and when is it worth doing?
A predictive model does exactly what the name implies: it predicts wireless coverage from the floor plan before anything is installed. It accounts for wall materials, ceiling height, and access point placement, and answers three questions: how many access points, where they go, and what coverage the client should realistically expect.
The value shows up in the sales conversation as much as the install. A predictive model shows the client why an access point goes where it goes, which is a much easier discussion than defending a line item. On commercial bids, it does something else: it gives you a defensible design to submit rather than a count someone estimated.
How should access points be placed in a multi-story building?
Signal does not stop at the floor. Access points stacked directly above one another on the same channel will interfere, even when they look correctly spaced on each individual floor plan. Placement should be staggered between floors, with a single channel plan that treats the building as one three-dimensional space.
Vertical bleed is largely a power problem, so transmit power usually comes down as part of the fix. Wood-framed homes bleed heavily between floors. Concrete commercial floors bleed far less, which sounds easier but creates the opposite problem: coverage that stops abruptly at stairwells and elevator lobbies, where people still expect to stay connected.
Not sure how many access points a project needs? Send the floor plans through our intake form and get a predictive design back.
Segmentation, Multicast, and Discovery
Why do some devices disappear from their app when they are on a separate VLAN?
Devices that find each other using multicast discovery go invisible across subnets because that traffic does not cross them on its own. Speakers, casting devices, printers, and some control systems rely on it. The device is online and reachable by IP. It simply never announces itself where the app is listening.
Not everything breaks, which is what makes this hard to predict. Devices that connect through a cloud service or are addressed directly by IP are unaffected, so a VLAN plan can look correct right up until the one system that depends on discovery gets commissioned.
Making it work takes three things configured together: a reflector to carry discovery traffic between VLANs, multicast handling that actually forwards it, and firewall rules permitting it in both directions. Miss any one, and it half works.
Does AV over IP require IGMP snooping, and what happens without it?
On a managed switch, yes. Without IGMP snooping, the switch treats multicast like broadcast and floods every video stream to every port. On a network carrying high-bitrate AV that saturates links quickly, and it produces exactly the symptoms integrators report as video instability and audio dropouts.
The part that trips people up is that snooping alone does nothing. The network also needs an IGMP querier, and every switch in the path needs snooping enabled, access points included. Enable it on one switch in a stack, and the problem moves rather than disappearing.
Should IoT and control devices be on their own VLAN?
Usually yes, for two unrelated reasons. Security, because IoT devices are the least maintained things on any network and should not share a segment with client data. And stability, because chatty discovery traffic from dozens of devices degrades performance for everything else on that segment.
The catch is the discovery problem above. A VLAN plan that isolates devices without planning how apps, control systems, and speakers will still see each other produces a network that is more secure and less functional.
Drawing the VLANs is the easy part. Making the system work across them is the engineering.
Which switch features actually matter for AV over IP?
Four, and port count is not one of them. IGMP snooping with a querier, enough multicast group capacity for the number of streams, a non-blocking backplane that can carry those streams at once, and a PoE budget that holds up with every port populated rather than the number on the datasheet.
Multicast group capacity is the one that gets missed. Every switch has a ceiling on how many groups it can track, and a large AV deployment can exceed it on hardware that looked fine during specification. The system works in commissioning with a handful of streams and fails when the full install comes online.
What causes a network to work at handoff and fail three months later?
Nothing about the network changes. The environment around it does. Devices get added, the client swaps the ISP router, firmware auto-updates, a DHCP pool sized for the original device count fills up, and neighboring wireless networks appear on channels that were clear during commissioning.
The pattern is consistent enough to plan for. A network validated once at handoff is only validated for the conditions of that day.
Proactive monitoring is built for this. When a device goes offline or starts behaving abnormally, a ticket opens automatically, and diagnosis starts before the client notices and calls you.
Hardware
Does WhyReboot sell hardware, or only design and configure the network?
Both, but the hardware is not the product. We supply, configure, stage, and warranty the gear, and the reason to work with us is the engineering around it: the design, the documentation, and the support after handoff. Buying a switch from us and buying a designed network from us are different things.
Can you work with equipment that has already been purchased?
Yes. We will work with part of a network or all of it, and we do not require a rip and replace to get involved. Retrofits, budget changes, and mystery hardware behind a television are normal conditions, not exceptions.
We prefer our own gear because we know exactly how it behaves and can stand behind it. But projects have history, and refusing to work with what is already installed helps nobody finish the job. Though there are limitations to the type of gear we can fully support and configure, if a third-party device is in the network, we will support the network up to where that device connects.
When is enterprise-grade networking overkill?
When the client's actual demands do not justify it. A small home with a dozen devices, no AV over IP, no whole home audio, and no expectation of coverage outdoors does not need an enterprise stack. Specifying one there raises the price and gives the client nothing they will notice.
What actually moves a project into enterprise territory is not size or budget. It is device count, roaming requirements, AV traffic, segmentation needs, and how much the client cares when something stops working. A 2,000 square foot condo with heavy automation can need more engineering than a large house with a television and a laptop. A homeowner who cannot be down for even 5 seconds could gain real benefit from an enterprise-grade network.
What ongoing licensing or renewals does enterprise networking gear require?
Terms vary by manufacturer, but the general principle is consistent: buying the hardware and maintaining entitlement to support, firmware, and replacements are separate things. Renewals typically cover support access, warranty service, and in some cases continued access to firmware and management features.
This is worth confirming per product line before quoting, because it is a recurring cost the client will encounter after you have left.
On gear purchased through us, most indoor hardware carries a lifetime warranty for as long as the manufacturer still supports it, with next-business-day advanced replacement for the first year and ongoing while the contract stays active. Configuration changes are included for that same contract period.
Working with WhyReboot
What does WhyReboot actually do for an integrator?
We are the network engineering department you do not have to hire. We design the network, configure and stage the gear, document everything, and support it 24/7 after it is installed. You keep the client relationship, and we stay behind you.
In practice, that covers routing and firewalls, switching, and wireless, designed for what AV, control, and automation systems actually demand rather than for office traffic. The work happens across your whole project lifecycle, from predictive design during the bid through support calls two years after handoff.
Do you work directly with my client, or behind me?
Behind you, always. We work only with our partners, never directly with homeowners or end users. Your client relationship stays yours, and nothing we do puts a second company in front of the person who hired you.
What happens when something breaks at 9 pm on a Sunday?
You get a senior engineer. The first question will not be for a serial number to check warranty status, because a technician standing on site at 9 pm on a Sunday needs help, not an eligibility interview. We help first and sort out the contract afterward, unless there have been repeated calls out of contract and we’ve already helped out, gratis.
If it turns out a system was outside its service contract, we will contact your renewals department later that week to let them know we assisted. That conversation happens after the problem is solved, and it happens with your office, not with your technician in the middle of a job.
How do I reach an engineer, and how fast?
Email support@whyreboot.com is the fastest, or call 877-307-0052 and choose option 2. There is no tiered call center and no script. Our internal metric is that nobody sits on hold and every request reaches a certified engineer immediately.
One requirement: requests need to come from your company domain so we can link them to the right company and project. And we cannot give out passwords or documentation over the phone, since there is no way to verify who is calling. Those come through the documentation portal.
Do you charge for configuration changes after the install?
No. Configuration changes are included for a full year and continue as long as the contract is renewed. A new VLAN, a second WAN, a policy adjustment: if it can be configured, it is covered, and you will not receive an invoice for one more change. Did you delete the config by accident, or make a change that killed the entire network? That’s what we’re here for as your partner; we’ll get it working again for you.
What does the lifetime warranty cover?
Most indoor gear purchased through us, for as long as the manufacturer still supports the product and it has not reached end of life. For the first year, you also get next-business-day advanced replacement, which continues while the contract stays active.
The practical difference is that a failed switch gets replaced before the old one ships back, so the project keeps moving instead of waiting in RMA limbo.
Do you monitor networks after deployment, and what does that include?
Yes, through our Pro-Active Monitoring service. When a managed device such as an access point, switch, or firewall goes offline or starts behaving abnormally, a ticket opens automatically, and our team begins diagnosing immediately rather than waiting for a client to report a problem.
It also covers automatic configuration backups with version tracking, so a failed device can be restored to its last working configuration quickly. Remote access works even on sites without a public IP, including cellular, satellite, and CG-NAT connections, which removes a lot of unnecessary truck rolls.
What do you need from me to start a project?
Floor plans and a completed intake form. That is the whole starting requirement, and the form takes about five minutes.
From there:
- We build a predictive Wi-Fi design from the floor plans
- You get a quote with hardware and datasheets included
- Once approved, gear is configured, firmware updated, and tested in-house
- Everything ships staged, with installer-ready diagrams and a printed booklet for the client that contains no passwords
- Documentation and credentials live in your Bibliotek portal, with a full audit trail
Ready to try us on a project? Send floor plans through the intake form and see what comes back.
For Homeowners
My integrator mentioned WhyReboot. Who are you?
We are the network engineering company behind your integrator. We design and support the network your smart home systems run on. You hired your integrator; your integrator brought us in for the networking, and we work through them rather than with you directly.
Do you work with homeowners directly?
No. We work only with integration companies. If you have a question about your system, your integrator is the right call, and they have direct access to our engineers whenever they need us.
Why does a smart home need a designed network instead of a router from the store?
A consumer router is built for a few laptops and phones. A connected home runs dozens of devices that talk constantly: speakers, cameras, thermostats, lighting, televisions, control processors. The failures show up as things that mostly work, which is harder to live with than something that plainly does not.
A designed network handles device count, keeps coverage consistent throughout the house and outside it, separates devices that should not share a segment, and can be diagnosed remotely when something goes wrong. Most of what people describe as a smart home being unreliable is a network problem underneath.
If my network is monitored, what can you see?
Monitoring watches the health of the network equipment itself: whether an access point, switch, or firewall is online and working normally. It exists to catch a failing device before it interrupts your home, not to observe what happens on your network.
Your integrator can walk you through exactly what is monitored on your system and turn it off if you prefer.
My Wi-Fi stopped working. Who do I call?
Your integrator. They know your system, they hold the relationship, and they can reach our engineers 24/7 if the problem is on the network side.
Answers reviewed by WhyReboot’s engineering team. WhyReboot has authored and coauthored CEDIA's Advanced Networking courses and certification exams and contributed to the Advanced Residential Electronic Systems textbook used in colleges nationally. Our founder, Bjørn Jensen, was an inaugural member of the CEDIA IT Task Force.
