It was finally time to put a Mimosa A5-14 access point through its paces on the network. We wanted to test two different scenarios:
- Speed of the C5 subscriber module at three different distances from the AP
- Compare it to a Cambium Force 180 subscriber module in WiFi mode attached to same AP
The A5-14 was attached to a 250 x 250 fiber line via a 5GHz radio link. We were able to completely saturate that link at the Netonix switch feeding the A5-14 AP so we are pretty confident the speeds reported here are actual radio speeds, not limited by backhaul. We were also the only client on the A5-14 at the time of testing so this was absolute best case scenario.
For the speed tests, we used speedtest.net results through the Internet, not radio to radio or local speeds.
Now, I understand the A5-14 is designed for optimal performance at .5km and under distances. I get that. However, I also wanted to see what it did at distances further than that so we could get an idea of what to expect with these if we deploy any more.
- A5-14 mounted on top of a house - all tests were clear line of sight and fresnel
- 40 MHz channel at 5310 (no clean channels in U-NII 1 or U-NII 3)
- C5 and Force 180 were on a tripod ~5' above ground level
Here is the RF environment from the A5-14 prior to testing:
Test 1: 175 meters away and ~50' below AP:
Test 2: 540 meters away and even with AP:
Test 3: 1200 meters away and ~20' above AP:
I found a couple interesting things I am not easily able to explain:
- Mimosa C5 ping times went from 31 ms to 7 ms after the first test and stayed at 7 ms no matter how many times I repeated the speed test. Force 180 stayed a ~30 ms for every test.
- Mimosa C5 upload speed stayed very high throughout the tests and even improved at the 500 meter test from the 175 meter test. Maybe we were too low at 175 meters?
- Why did download speed drop off much faster than upload on the C5?
I get that a C5 to Force 180 is not a perfect comparison. Perhaps a Force 200 is a better apples to apples. The test was really for the C5 but since I happened to have a Force 180 laying around, I thought I'd give it a try since we have all Cambium gear deployed at this point.
I have now loaded a couple test clients on the A5-14 at distances less than 100 meters. They are reliably seeing speeds of 200 x 200 when testing. This is without GPS synchronization since that is not possible yet with Mimosa firmware and will be a requirement if we decide to deploy more of these.
For rough comparison, we are running a Cambium ePMP 1000 GPS AP on a 40 MHz channel in the U-NII-1 band at a 75/25 ratio. Speed tests are in he neighborhood of 120 x 35 best case but tend to average out around 100 x 30. We don't have data on that AP in flexible mode for comparison.
Left the test fairly impressed with the A5-14 but I can't draw any conclusions on usability for us until GPS sync is enabled.
For some additional reference, here are a couple UI screen shots from the C5 test at 1200 meters with the largest UL/DL difference.