// POWER OVER ETHERNET · THE HIDDEN COST
Replacing access points is never just access points
The most expensive thing people miss when pricing a Wi-Fi refresh - and the five-minute check that tells you whether your existing access points are quietly running at half strength right now.
- ● Vendor sources linked
- ● Dates checked 27 July 2026
- ● Independent — we sell no vendor’s kit
Independent, and dated. 365 Techies is an independent IT firm. We are not an authorised partner, reseller or agent of RUCKUS Networks, Belden, CommScope, Cisco or Cisco Meraki, and we have no access to their support entitlements on your behalf. Every date on this page comes from the vendor’s own published notices, linked in the source column, and was checked on 27 July 2026. Where we could not verify something from a vendor source we say “check with vendor” rather than guess.
The quote says access points. The job is switches too.
This is the most expensive thing people miss when pricing a Wi-Fi refresh, and it is entirely knowable in advance.
Older business access points were designed to run on standard Power over Ethernet (802.3af). Their modern replacements almost universally need the higher-power standard (802.3at, often called PoE+). So an estate that has run happily for eight years on the existing switches cannot simply have new access points hung on the same cables.
What silently switches off when power is short
An underpowered access point does not fail and does not warn you. It reports as healthy in the controller while quietly running with fewer aerials, a dead second network port and, on several models, no USB and no IoT radio. Users experience it as “the Wi-Fi got worse and nobody changed anything”.
| Model | Streams | Headline rate (theoretical) | Ports | PoE for full function | What you lose on 802.3af | WPA3 |
|---|---|---|---|---|---|---|
| RUCKUS R310 | 2x2:2 SU-MIMO (no MU-MIMO listed) | 300 / 867 Mbps | 1 x 1GbE | 802.3af (11W max) | af is its normal mode | No |
| RUCKUS R500 | 2x2:2 SU-MIMO | 300 / 867 Mbps | 2 x 1GbE | 802.3af (10.5W max) | af is its normal mode | No |
| RUCKUS R600 | Check with vendor | Check with vendor | Check with vendor | Check with vendor | Check with vendor | Check with vendor |
| RUCKUS R510 | 2x2:2, SU-MIMO and MU-MIMO | 300 / 867 Mbps | 2 x 1GbE, 1 x USB 2.0 | 802.3af (12.6W max) — no separate 802.3at mode listed | af is its normal mode — this is why R510 estates sit on af-only switches | Yes Listed on the current datasheet; needs later firmware and a controller version that exposes it. |
| RUCKUS R610 | 3x3:3, SU-MIMO and MU-MIMO | 450 / 1300 Mbps | 2 x 1GbE with LACP, 1 x USB | 802.3at (21.87W) | 2.4GHz limited to 2x3 (2-chain TX at 21dBm aggregate); USB port off; one Ethernet port off | Yes Firmware and controller dependent. |
| RUCKUS R710 | 4x4:4 SU-MIMO; 3 streams for MU-MIMO | 600 / 1733 Mbps | 2 x 1GbE with LACP, 1 x USB | 802.3at (peak 25W) | 2.4GHz drops to 2x4 at 19dBm/chain; 2nd Ethernet disabled; USB disabled | Yes On the current (2020) datasheet. The 2019 revision listed no WPA3 - it arrived by firmware. |
| RUCKUS R720 | 4x4:4, SU-MIMO and MU-MIMO | 600 / 1733 Mbps | 1 x 2.5GbE, 1 x 1GbE, 1 x USB | PoH / UPoE (up to 33.5W) for all features | af (12.95W): both radios collapse to 1x4. at (25.5W): full 4x4 but the 2nd Ethernet port and USB are still disabled | Yes On the current (2020) datasheet. The 2019 revision listed no WPA3 - it arrived by firmware. |
| RUCKUS R320 | 2x2:2, SU-MIMO and MU-MIMO | 300 / 867 Mbps | 1 x 1GbE | 802.3af (12.18W max) | af is its normal mode - nothing is disabled | Yes On the current datasheet only. The 2019 revision listed no WPA3, so it arrived by firmware - old firmware will not have it. |
| RUCKUS R730 | 8x8:8 (5GHz) + 4x4:4 (2.4GHz) - 12 streams | 1148 Mbps / 4.8 Gbps | 1 x 1/2.5/5GbE, 1 x 1GbE | PoH / UPoE / DC (31.0W max) | RUCKUS marks 802.3af not recommended. On 802.3at (23.8W) the 5GHz radio runs 4x4 instead of 8x8 and the Zigbee/BLE radio is disabled - so an R730 on an ordinary PoE+ switch silently runs at half its 5GHz streams | Yes Datasheet states WPA3 and Wi-Fi Enhanced Open. |
| RUCKUS R550 | 2x2:2 both bands, SU-MIMO and MU-MIMO | 574 / 1200 Mbps | 2 x 1GbE, 1 x USB | 802.3at (18.71W) | 12.71W limited mode: 2.4GHz 2x2 at 19dBm/chain; 2nd Ethernet, onboard IoT radio and USB all disabled | Yes WPA3-Personal and WPA3-Enterprise both listed on the datasheet. |
| RUCKUS R650 | 4x4:4 (5GHz) + 2x2:2 (2.4GHz) | 574 / 2400 Mbps (2974 combined) | 1 x 2.5GbE, 1 x 1GbE, USB, BLE + Zigbee | 802.3at (21.59W) | 12.25W: 5GHz drops 4x4 to 2x4; 2nd Ethernet, IoT radio and USB disabled | Yes |
| RUCKUS R750 | 4x4:4 both bands | 1148 / 2400 Mbps | 1 x 2.5GbE, 1 x 1GbE, USB, BLE + Zigbee | 802.3at (22.34W) | 12.54W: both radios 2x4; 2nd Ethernet, IoT radio and USB disabled | Yes |
| RUCKUS R850 | 8x8:8 (5GHz) + 4x4:4 (2.4GHz) - 12 streams | 1148 Mbps / 4.8 Gbps | 1 x 1/2.5/5GbE, 1 x 1GbE | 802.3bt-class supply (PoH / UPoE / DC) | On 802.3at the 5GHz radio runs 4x4 instead of 8x8 | Yes |
| RUCKUS R670 | 2x2:2 / 2x2:2 / 2x2:2 tri-band (6 streams), or 2x2:2 + 4x4:4 dual-band | 689 / 5765 / 5765 Mbps (9.34 Gbps combined) | 1 x 1/2.5/5GbE, 1 x 1GbE | Check with vendor | Check with vendor | Check with vendor |
| RUCKUS R770 | 2x2:2 (2.4GHz) / 4x4:4 (5GHz) / 2x2:2 (6GHz) | 689 / 5765 / 5765 Mbps (12.22 Gbps combined) | 1 x 1/2.5/5/10GbE, 1 x 1GbE | Check with vendor | Check with vendor | Check with vendor |
| Cisco Aironet 1700 (AIR-AP1702I) | 3x3:2 | 867 Mbps (80MHz) / 300 Mbps (40MHz) | 2 x 1GbE + console | 802.3af / 802.3at / Cisco Enhanced PoE; 15W draw | — | No |
| Cisco Aironet 2700 (AIR-AP2702I/E) | 3x4:3 | 1.3 Gbps | 2 x 1GbE + console (plus AUX for a downstream device) | 802.3at; 15W draw | Cisco's wording: the access point will dynamically shift from 3x4 to 3x3 | No |
| Cisco Aironet 3700 (AIR-AP3702) | 4x4:3 | 1.3 Gbps | Check with vendor | 802.3at, Enhanced PoE or PWRINJ4 (16.1W for 4x4:3) | Degrades to 3x3:3 | No Inferred from sibling Wave 1 datasheets — the 3700 datasheet has been withdrawn from Cisco's site. |
| Cisco Aironet 1830 (AIR-AP1832I) | 3x3:2, SU-MIMO and MU-MIMO | 867 / 300 Mbps (vendor aggregate: up to 1 Gbps dual-radio) | 1 x 1GbE only, plus console and USB | 802.3af / 802.3at; 15.4W draw | Cisco's wording: if 802.3af PoE is the source of power, the USB port is disabled | Yes Delivered by the controller (AireOS 8.10.x or Catalyst 9800), not by the AP. |
| Cisco Aironet 1850 (AIR-AP1852I/E) | 4x4 MIMO, 4 streams (SU-MIMO); 3 streams (MU-MIMO) | 216.7 Mbps / 1.73 Gbps | 2 x 1GbE (no multi-gig) | 802.3at or Cisco Enhanced PoE (20.9W draw) | 802.3af cannot supply its 20.9W draw | Check with vendor |
Vendor headline figures are theoretical PHY rates. Real throughput per client, on a shared access point, is a fraction of them — Cisco’s own datasheets use the word “theoretical”.
Three traps worth checking today
The silent degrade
An access point negotiated at 802.3af when it wanted 802.3at looks perfect on the dashboard. Check the negotiated class on the switch, per port. Highest-yield check in this entire guide, and it costs nothing.
The upgrade cliff
Budgeting for new access points without checking the switch is how a Wi-Fi project goes over budget in week one. Count the ports, check the standard, then price.
Per-port class is not the budget
A switch can support the higher standard on some ports and run out of total power for the rest. The result is a coverage complaint that looks random and moves around - maddening to diagnose if you do not know to look.
Long cable runs and outdoor sites change this arithmetic again — see rural and multi-building Wi-Fi, where injector siting and run length matter as much as the standard on the switch. If your “access points” are really a mesh kit, the mesh guide is the better starting point.
// BEFORE YOU RING ANYONE
Walk the site and write down what people actually get
Our free Wi-Fi signal test runs in a browser, needs no sign-up, and scores every room. Ten minutes with it turns “the Wi-Fi is rubbish upstairs” into numbers — which is the difference between a guess and a diagnosis, whoever ends up doing the work. See also our room-by-room method.
Read next
The full guide
End-of-support dates for every model we could verify, and how to tell whether yours is actually the problem.
R510 field playbook
Diagnose it properly, and the 802.3af trap that sends an R550 upgrade backwards.
Your controller is the real deadline
Not the access points. This is the one most people miss.
Not sure what you're looking at?
Send us a photo of the label on one access point and the make of your controller. We'll tell you what you've got, what its dates are, and whether it's actually your problem — no charge for the answer.
01202 775566 · help@365techies.co.uk · MON–FRI 9AM–5PM