Wireless Performance - 5 GHz
Whatever pixie dust Belkin put into the 2.4 GHz radio, it forgot to put into the 5. While I was running the 5 GHz tests, I almost thought someone had swapped in a different product! The 5 GHz Wireless Performance Table in Figure 8 shows the N750 DB as the lowest performer among the three, pretty much across the board!
Highest throughput of 62.0 Mbps was obtained running upnlink in 40 MHz mode in Location A. I was able to squeeze out only 85 Mbps running simultaneous up and downlink tests in Location A and 40 MHz mode.
Figure 8: Belkin N750 DB Wireless Performance summary - 5 GHz
You can see in Figure 9 that hroughput stability in 5 GHz was pretty bad compared to 2.4 GHz. Connection was stable enough, even in Location D. But it didn't break through into my dead-zone Locations E and F.
Figure 9: Belkin N750 DB IxChariot plot summary - 5 GHz, 20 MHz mode, downlink
Here are links to the other 5 GHz plots if you'd like to check them out.
- 5 GHz / 20 MHz uplink
- 5 GHz / 20 MHz up and downlink
- 5 GHz / 40 MHz downlink
- 5 GHz / 40 MHz uplink
- 5 GHz / 40 MHz up and downlink
Closing Thoughts
Belkin has been trumpeting the results of the testing that it paid Tolly to do. You can download the report here for the price of giving up some contact information. The report is full of lots of impressive big numbers, like > 900 Mbps of total combined throughput, which was obtained by adding together bi-directional wired routing throughput and best-case 2.4 and 5 GHz wireless throughput.
The one important thing the report doesn't mention is whether they used TCP/IP or UDP for the test streams. I suspect UDP, since it usually gooses the results a bit due to lower protocol overhead. (I use TCP/IP for all routing and wireless testing.)
I can say that the N750 DB is one of the fastest routers I've seen to date for wired routing throughput. And it's also one of the best 2.4 GHz routers I've tested for mid to low signal strength performance.
But when it comes to the 5 GHz band, the N750 DB seems to morph into an entirely different beast, with anemic throughput and less stable throughput. Maybe it'll look different when I come back and run the three-stream tests on it, once I get a properly-manufactured test machine in house.