Longer Term Tests

Most of our propagation tests have only been of a few hours duration so we have been anxious to run some longer term tests to look at things like the seasonal changes due to variation in vegetation, climate and variations in the noise floor.

For our first test we have select test 22 to the Mersey Tobeatic Research Institute, it is one of the longer links at 9km with at least 1/2 of the direct signal path through the trees a challenging link that none of the technologies currently available to unlicensed wireless internet service providers would be capable of doing.

More details of the link can be found at the test report page.

 

We were fortunate in that we were able to make the link to the Institute directly from our 2nd test sight because they were eager to try anything that could offer better internet service.  They are one of the larger employers in the area and their day to day activities have been negatively affected by inadequate internet access. They have tried a variety of options in the past including satellite and 900Mhz technology and whilst these were adequate for a while increasing data demands rapidly lead to problems. A maximum download speed of 1.5Mbs when it is working and a data cap doesn't meet today's business needs.

 

This is the TVWS equipment we are using a GVWS4000 subscriber module and 8dBi gain Yagi antenna from 6Harmonics, with any equipment the maximum speed we can achieve is depended on the signal strength and one of the purposes of this trial is to find out what kind of variation to expect. So far we are averaging a signal strength of -78dBm and with that our TCP download speed is 6Mbs greater than the 5Mbs DSL internet connection we are using, the fastest wired internet in the area. The are several measures such as increasing antenna height and gain that we could take to increase the signal strength and link speed that we haven't undertaken given the limitations of the DSL connections in the area.

We are confident if we did make these changes that we could get the link speed close to it's maximum speed of 16Mbs UDP or 11-12Mbs TCP and have already experienced this under certain climatic conditions.

The equipment we are using is capable of operating in dual channel mode when approved by Industry Science and Economic Development Canada this will enable us to nearly double the throughput to 20-24Mbs at no additional cost, unfortunately the is nowhere in the area we can economically connect to the internet at that speed. 

Technical Details of the Initial Installation

The specifications for the equipment we are using can be found here

More details of the link can be found at our initial link report here

After semi permanent installation of the equipment we are getting the following statistics

Base Station Antenna Height 14m

Mersey Tobeatic Research Instituted Antenna Height 7 meters

RSSI -78dBm, Noise Floor -103dbm SNR24

Data Rate MCS 4  Raw data rate 10Mbs UDP or ~6-7Mbs TCP

Earlier speed test were conducted using an LTE connection to the internet and with the weaker signal were 4Mbs TCP up and down in other words symmetrical

We are now connected to a DSL line which is limiting our speed to 5Mbs down and 0.5Mbs up

Tests show the latency between our equipment is less than 2ms but the available internet sources add 20-60ms.

Early Results from the Long Term Tests

This test was started in early December long after all the leafs where gone and we didn't expect much variation in signal strength, we were wrong.

Over a one week period we have seen a variation in signal strength of 12dBi over a 9km link fortunately higher than our initial test. The weather has been a significant factor we have seen the following conditions

Cold -15C windy blizzard conditions signal -70dBm

Heavy snow fall no wind -5C signal -78dBm this increased to -74dBm after a wind came up

Calm high humidity and rain temperature around freezing,  signal -82dBm

Rain windy humidity 97% +12C signal -81dBm

Our current theory is that unlike conventional links where most of the signal path is through free air a lot of the signal path is through vegetation specifically evergreen trees. A significant part of the signal path is skimming through the tree tops where a build up of moisture or snow would have a noticeable affect. We suspect that when we had temperatures around zero there was some ice forming which together with the high humidity/freezing fog lead to the highest attenuation

We will be watching this link with considerable interest, we had assumed we would experience maximum signal attenuation in early summer just after the leafs are fully out perhaps this is not the case with this particular link.

We now have longer term data that can be viewed on the Climatic Variations web page