We'll come back and run standard routing throughput, storage performance and wireless performance tests on the EA8500 in Part 2. I wanted to get MU-MIMO testing done since that's the main reason you'd want to buy this or any other second generation 4x4 router. We won't be doing much for feature review since the EA8500 has Linksys' standard Smart Wi-Fi feature set. No sign of the promised OpenVPN support yet.
Since MU-MIMO testing involves multiple clients, our test process uses open-air testing. Linksys' reviewer's guide suggested the "triangle" arrangement below, with at least 6 feet between the EA8500 and each test laptop. The reason for this arrangement is to maximize the router's ability to get optimum beamforming for each client by maximizing client-to-client distance at short range. Simply put, the better the beamforming, the better throughput to each client can be optimized.
Suggested MU-MIMO test configuration
Because laptops and tablets with MU-MIMO enabled radios won't be out until July, Linksys and QCA had to gin up MU-MIMO enabled laptops and include them in the reviewer's kit. The box from Linksys contained three Dell Inspiron 13 7000 Series 2-in-1 laptops (7348, Early 2015) with QCA-sourced 1x1 AC MU-MIMO enabled wireless card installed. The laptop configuration ordered had a 2.2 GHz Intel Core i5-5200 CPU with 8 GB RAM running Windows 8.1 Pro 64 bit.
These laptops usually come with a 2x2 Intel Dual Band Wireless-AC 7265 + Bluetooth 4.0 card. Device properties for the substitute QCA adapter identified it as Qualcomm Atheros QCA9377 Wireless Network Adapter using a 18.104.22.1686 driver. The adapter properties didn't include any references to MU-MIMO, beamforming, band selection or bandwidth selection. This card is not and will not be commercially available.
The photo below shows it installed in the Dell laptop. Note two antenna wires are connected, which QCA says support antenna diversity. Note that each client showed a full 433.3 Mbps link rate with 5 full signal bars in Windows Wireless Connection properties, no matter where clients were located. So I can only conclude the Windows driver couldn't access the correct information from the adapter.
Dell laptop with QCA MU-MIMO WLAN card
As explained earlier, MU-MIMO's claim to fame is improved bandwidth utilization by effectively splitting a single 4x4 radio into three 1x1's. So the test is pretty simple:
- Disable MU-MIMO, connect three 1x1 clients, run a test to all three clients simultaneously, record each client throughput and calculate total throughput
- Repeate the test with MU-MIMO enabled
- Calculate the throughput differences
All tests used the IxChariot throughput.scr script with the test file size set to 5,000,000 Bytes, TCP/IP protocol, one connection per client, 90 second test length. Each test was run once. The in-house 5 GHz network was not shut off during the test, but it was idle and only beaconing (no traffic). Channel was set to 153, bandwidth mode was in its default
80 MHz Auto mode and everything else was set to defaults.
I was able to easily enable/disable MU-MIMO and you'll be able to also, using the hidden advanced wireless screen accessed via the URL shown in the screenshot below. The Qualcomm screen also has beamforming controls for both bands.
Advanced wireless screen - Qualcomm tab
The 2.4 and 5 GHz advanced wireless screens have the same controls and default settings. The only difference is the 5 GHz radio Sideband Selection control is set to Lower.
Advanced wireless screen - 2.4 GHz tab
In the sections to follow, all plots show downlink throughput only, since that is the only direction that MU-MIMO works in. To check, I also ran each test configuration in uplink with MU-MIMO enabled and disabled and saw total throughput change from -3 to +5%—which I chalk up to measurement variation.
The first test shows Linksys' suggested triangle configuration. Due to limitations of my test space, I had to position one client in front, one in back and one on the right side (facing router front). All were about 6 feet from the router. The accomplished the same goal of optimizing beamforming as the suggested triangle configuration.
The plot shows throughput difference between MU-MIMO disabled and enabled for each client and for all three as a total. Throughput change for each client varied greatly from -6% worst case to 167% best. In all this test configuration showed a total throughput gain of 31%.
MU-MIMO Throughput change - Triangle
The IxChariot plot below shows the test run throughput plot with MU-MIMO disabled.
MU-MIMO Throughput - Triangle configuration - MU-MIMO disabled
And with MU-MIMO enabled. It's obvious that throughput is more equally distributed among the three clients and has reduced variation. Remember, this is an ideal test configuration.