TP-Link Omada is good kit. It is capable, well priced and genuinely enterprise-grade. Yet we are called out to Omada networks most weeks, and almost none of the problems turn out to be the hardware. They are design and configuration issues: access points in the wrong places, roaming never set up, devices on the wrong part of the network, and a controller nobody has logged into since the day it was installed. Here are the faults we see most, what actually causes them, and what fixing them properly involves.
The 30-second version
- Most Omada problems are configuration, not hardware. The kit is rarely the reason your Wi-Fi is misbehaving.
- Three faults account for most call-outs: roaming that was never configured, IP and VLAN conflicts, and too many access points all running at full power.
- Adding more access points usually makes it worse, not better, unless somebody surveys the building first.
- A survey costs less than more hardware. We regularly fix networks by moving and reconfiguring what is already on the wall.
- We are certified on TP-Link and Ubiquiti and cover Merseyside and the North West with our own engineers.
On this page
- The Wi-Fi is fast, but Teams keeps freezing
- Some devices work, others just will not connect
- Wi-Fi everywhere, but there are still dead spots
- Access points that will not adopt, or keep going offline
- Access points that reboot on their own
- Mesh used where a cable should have been
- Fast roaming that breaks the card machine
- A guest network that is not actually separate
- The controller nobody thought about
- Frequently asked questions
The Wi-Fi is fast, but Teams keeps freezing
This is the single most common complaint we hear, and it is the one that confuses people most. You run a speed test, it comes back at several hundred megabits, and yet Teams drops, the cloud system hangs and someone loses a call every time they walk to the meeting room. The broadband is fine. The network is not.
What is actually happening
A speed test measures a moment, standing still, on one device. A video call is a continuous stream that notices every interruption. The usual causes are all things a speed test will never show you:
- Access points in the wrong places, chosen for where there happened to be a cable rather than where people work
- Overloaded or overlapping channels, with your own access points and the neighbours all competing for the same airspace
- Interference from other equipment, or from access points set to a channel width too wide for a busy environment
- Devices hanging onto the wrong access point, clinging to a weak signal from across the building instead of handing over to the one above their head
That last one is the biggest and the least understood. Your laptop decides when to switch access points, not the network, and left to itself it holds on far longer than it should. Omada has the settings to fix this. They are simply not enabled by default.
How we fix it
We survey the network properly, check coverage and interference across the building, optimise channels and transmit power, reposition or add access points where the survey says they are needed, and configure Omada roaming so devices move smoothly rather than dragging a dying connection around with them. In most cases the hardware stays exactly where it is and the network starts behaving.
Some devices work, others just will not connect
Printers that vanish. A card machine that works until Saturday. CCTV that appears offline in the app but is plainly recording. Laptops that connect fine while the tablet next to them will not. When some things work and others do not, the problem is almost never the Wi-Fi signal.
What we normally find underneath it:
- Duplicate IP addresses, usually because something was given a static address inside the range the router is handing out automatically
- DHCP problems, most often a scope too small for the number of devices, so the network runs out of addresses at the busiest time of day
- Incorrect VLANs, with a device physically connected but sitting on a part of the network that cannot reach what it needs
- Old settings left behind from a previous installer, a previous router, or a device that was moved and never reconfigured
How we fix it
We look at the complete network rather than rebooting the Wi-Fi and hoping. We identify IP conflicts, tidy up the DHCP and VLAN configuration, reserve addresses for the equipment that matters so it always gets the same one, and make sure every device is sitting on the right part of the network. Card machines and cameras in particular should not be wherever they happened to land.
If you have Omada already, three things worth checking today
- Log into the controller and look at how many access points show as connected versus disconnected. If you cannot log in at all, that is the first problem.
- Check whether any access point is running over a wireless uplink rather than a cable. Omada shows this clearly, and it halves throughput for everything behind it.
- Check whether your guest Wi-Fi is a genuinely separate network or just a second password onto the same one. If guests can see your printer, it is the latter.
Wi-Fi everywhere, but there are still dead spots
The instinct when Wi-Fi is patchy is to add another access point. It is also the thing most likely to make it worse.
Access points share a limited number of usable channels. On 2.4GHz there are only three that do not overlap at all. Add units without planning, leave them on automatic settings at full transmit power, and they stop covering the building and start competing with each other. The symptom looks identical to not having enough coverage, which is why the usual response is to add a fifth access point to a problem caused by the fourth.
We have improved plenty of networks by turning power down and moving units rather than buying anything. Coverage is not the same as capacity, and more Wi-Fi is not the same as better Wi-Fi.
How we fix it
We optimise the Omada installation as a complete system: coverage, access point positioning, transmit power, channel selection and channel width, and roaming behaviour. The result is consistent Wi-Fi across the building rather than simply more of it in the places that already had some.
Omada installed but not performing properly?
We troubleshoot, optimise and manage business Wi-Fi networks across Merseyside and the North West, from annoying dead spots to more complicated VLAN, roaming and connectivity problems. Support is included, not billed by the incident.
Access points that will not adopt, or keep going offline
An access point stuck on Adopting, showing Disconnected, or reporting Managed by other is one of the most common Omada support calls there is, and it is almost always a discovery problem rather than a faulty unit.
Omada access points find their controller automatically only when they are on the same subnet. The moment you introduce VLANs, put the controller somewhere sensible, or move management onto a different network from the access points, that automatic discovery stops working. The APs carry on broadcasting Wi-Fi perfectly well using the last configuration they were given, which is why the Wi-Fi appears fine while the controller insists everything is offline.
The fix is to give them a way to find the controller across networks, using a controller inform address, a DHCP option pointing at it, or the discovery utility. It takes minutes once you know that is the problem, and it causes days of confusion when you do not.
Access points that reboot on their own
If an access point drops off and comes back every so often, and it is always the same one or two units, check the power before you check anything else.
Omada access points are powered over the network cable, and every switch has a total power budget shared across all its ports. Modern Wi-Fi 6 access points draw considerably more than the older units they replaced. Add one more AP to a switch that was already near its limit and the switch starts shedding power, which shows up as access points rebooting apparently at random, usually the ones furthest down the list.
It is an easy problem to diagnose from the switch and an expensive one to guess at, because from the Wi-Fi side it looks exactly like a coverage fault.
Mesh used where a cable should have been
Omada supports wireless uplink, so an access point can connect back to the network over Wi-Fi instead of a cable. It is genuinely useful where running a cable is impossible, and it is frequently used where running a cable was merely inconvenient.
The cost is significant. An access point on a wireless uplink is using its radio to talk to the network and to your devices, so the available throughput for everything connected to it is roughly halved before anyone has done anything. Chain two together and it halves again.
If a building has a mesh access point covering an area people actually work in, that is usually where the complaints are coming from. The honest fix is normally a cable.
Fast roaming that breaks the card machine
The features that make roaming work well for laptops and phones can break older equipment entirely.
Fast roaming and band steering are worth enabling for modern devices, but they are frequently switched on for the whole network without checking what else is connected. Older card terminals, some printers, handheld scanners and a lot of IoT equipment either do not support them or handle them badly, and the result is a device that connects, works for a while, then disconnects for no visible reason.
The answer is not to abandon the features. It is to put the older equipment on its own SSID with those options off, which is straightforward in Omada once somebody has established which devices actually need it.
A guest network that is not actually separate
Creating a second SSID called Guest does not, on its own, separate anything. Without guest isolation and a VLAN behind it, customer devices sit on the same network as your tills, your files and your cameras, with a password that half the building knows.
This matters more than it used to. If you take card payments, keeping payment equipment on its own network segment away from public traffic is standard practice and what the card schemes expect. If you work with clients who audit their supply chain, it is something you will be asked to evidence.
Omada does this properly. It needs configuring rather than assuming.
If you want to build a marketing list from your guest Wi-Fi as well, that is a genuinely useful thing to do, but it has to be opt-in. Under UK GDPR and PECR you need consent before sending marketing, the sign-in page has to make clear what someone is agreeing to, and every message needs a working unsubscribe.
The controller nobody thought about
The controller is the brain of an Omada network, and it is the part most often treated as an afterthought.
We regularly find software controllers installed on a PC in the corner that gets switched off at night, or on a machine belonging to somebody who left two years ago. When the controller is unavailable the access points keep broadcasting, so nothing appears wrong, but nobody can change a password, add a device, see what is connected or be told when something fails.
The other half of the problem is backups. A controller holds the entire configuration of your network. Without an export, rebuilding after a failure means reconfiguring everything from memory.
A hardware controller, or a properly hosted one, with scheduled backups and someone actually monitoring it, turns the network from something you hope is fine into something you know is fine.
Where Omada networks actually give businesses trouble
Ranked by how often we are called out to each, across offices, warehouses, practices and hospitality venues:
| What people report | How common | What it usually turns out to be |
|---|---|---|
| Calls and video freeze | Very common | Roaming never configured, so devices cling to a distant access point instead of handing over |
| One device will not connect | Very common | IP conflict, DHCP scope exhausted, or the device sitting on the wrong VLAN |
| Dead spots despite plenty of APs | Common | Too many access points at full transmit power, interfering with each other |
| AP shows offline or will not adopt | Common | Controller on a different subnet with no discovery method configured |
| Access points rebooting | Occasional | PoE budget on the switch exceeded once the last AP was added |
| Half the expected speed | Occasional | An access point running over a wireless uplink rather than a cable |
| Card machine or printer drops | Occasional | Fast roaming or band steering enabled globally on kit too old to support it |
| Everything works, nothing is separated | Very common | A guest SSID created without guest isolation or a VLAN behind it |
Tell us what your network is doing
If any of the above sounds like your building, send us the details and we will tell you honestly whether it is a configuration fix, a design problem or genuinely a hardware limitation. We will also tell you when the answer is cheaper than you expect, which it frequently is.