DePIN

Geodnet | A challenging GPS antenna install

GPS antenna installed on top of a roof
~ Andrea ~ Leave a comment

In December 2023 I installed a GPS antenna at a new site which is somewhat distant from where I live and, as I discovered in the following months, it has has some peculiarities which are negatively influencing the performance of the antenna.

At this link a quick video which allows you to understand the site situation.

This article describes all the steps that I have undertaken with the aim of improving the performance of the antenna, the results and the conclusions.

Since its first installation the antenna token reward rate was below my expectations. The green curve represented in the graph below represents the token reward rate which is directly bound to the performance of the antenna.

My expectation was to be completely flat curve showing a 100% token reward rate all the time since the antenna is located on the roof of a quite tall building.

The first attempt: go even higher

The first thoughts were related to the crane which you see in the background of the picture above. Maybe it was reflecting the GPS signal while moving. Or maybe it was a rather tall tree which is nearby. I knew that I could not overcome the presence of the crane but I could try to minimize the effect of the tree. Therefore I decided to install the antenna on a mast. Before proceeding I asked help to support and they suggested that I should position the antenna higher by a few centimeters. I went overkill and ordered a telescopic mast which would get 3 meters high.

The GPS antenna installed on the new mast

Once the installation was completed I started to monitor constantly the token reward rate curve as I was hoping to see the desired improvements. This was the first time that I realized that something in the graph was not exact: the curve token reward rate showed a deep fall in performance and this was expected, because during the works the antenna had to be moved around. What struck me was that the loss in performance in the graph had been registered rather late at night, between 21:00 and 22:00. At that time the electrician for sure was not working on the roof

After some weeks I had to realize that the antenna was performing slightly worse than before. The mystery was becoming even more mysterious. How was it possible that having installed the antenna on a much higher position I would get the same or even worse results?

The token reward rate (green curve) showed greater instability.

The wind

It was until one night during which a strong wind invested the whole region that I realized that I might have missed an aspect: I had never taken into consideration that the antenna might be moving with the wind and the fact that the token reward rate worsened during the storm was for me proof enough that all of my issues were caused by the wind. 😀

Soon enough I had some guy wires installed.

I had a lot of discussions with friends, electricians and with myself about the new setup. Were the guy wires pulling enough to avoid the mast to swing back and forth during windy days? Or did I need more weight as the ones I had the guy wires anchored to were too light (15Kg each)? I would have loved to anchor the guy wires in the floor but I had been strongly advised against it as the insulation would not be water tight anymore.

Once again by looking at the graph I could not find the answers I was looking for: the token reward rate seemed to show some improvement but after some time, the drops started to increase in depth and frequency. Moreover also the Rolling Reward Rate (RRR) decreased from 98.2% down to 94.4%.

There was also the suspicion that the antenna cable which I had left tight rolled in the box together with the miner could cause some signal loss. Since I had all those questions open it was evident that another intervention was needed, which I scheduled during one of the first possible free days. The objective was to further stabilize the installation by using a better weight distribution at the base, use a thicker mast (50 mm diameter) and distribute the antenna cable on the roof.

Here below some of the pictures I collected while doing renewing the install.

The antenna with 2 arms on the sides
The base is now much heavier
Beautiful insertion for the antenna
Antenna cable is now extended and running under the stones
Miner is now bound to the box and in general tidier

The mountains

Now it is already a few days that I have completed the last improvement and the graph looks like this:

Apparently starting at 09:00 of the 26/04/2024 the number of satellites was high enough to ensure a stable performance until 21:00. After that the signal started to bounce again for the whole night until 09:00 of the next day, time when the signal started to improve again.

So even after the third intervention I still was not able to reach the maximum performance and stability. At this point, having tried everything, only one thing was left for me to do: accept the fact that the location is not that good and that the mountains being so tall and near are covering some of the satellites.

Missing an analytical tool

By constantly watching the graph I realized that some of the events which happen on the antenna are either delayed or not tracked at all. During the 26th of April for example I was able to monitor the token reward rate and witness a drop from 30 to 27 satellites which I could not find on the graph later. The fact that the graph is rendered with a few hours of delay does not help troubleshooting, because if you change something to the installation you will get the results a few hours later. But even worse is the fact that since I could not see the drop in the performance which I had witnessed I started to wonder which data the graph is recording and/or if it is discarding some data. This is the reason why I started to wish for a detailed report which showing the number of effective satellites connected to the antenna every 15 seconds.

Lessons learned

Site check

In order to avoid all the trouble I went through it is very important that the choice of the location is attentively made. Go to the spot in person and check the following:

  1. Presence of mountains nearby
  2. Presence of trees nearby
  3. Use the mobile app for android GNSS view and use it to verify if there are objects/buildings which cover satellites
  4. Take pictures/videos while being right on top of the antenna position as well as from away: this way you may print the picture and measure with the help of a protractor if possible objects could interfere with the antenna sight or not
  5. Ask/check for wifi availability and/or for mobile coverage
  6. Ask/check for wind behaviour
  7. Verify possible fastening points

Analysis of previous months

Trying to understand the behavior of the antenna by looking at past performance it is not that useful as satellites keep moving as GNSS satellites are not geostationary. Furthermore the number of effective satellites is influenced by space weather as well, it makes no sense trying to understand what has happened in previous months. For this reason it is even more important that the location you choose allows you to have more than the bare minimum of 30 satellites in order to have some buffer in case the effective number of satellites drops. Having more than what is strictly required will allow you to keep a good performance 24/7.

Cranes

Cranes do not impact the satellite number. This was confirmed to me by some users on Discord as well.

Geodnet: description and requirements

~ Andrea ~ Leave a comment

Brief description

The aim of Geodnet is to make high precision GPS signal available for everyone at really competitive prices. As per today the technology to have 1 cm precise signal is already available, but it is expensive and in some cases not very convenient to deploy.

Geodnet and its community are working towards making high precision positioning available at very competitive pricing and convenient to use. In order to do that Geodnet needs to have at least one antenna every 25 km. A second antenna, at least 100 meter distant from the first one, is useful as backup in case the first one goes offline: more than two antennas are redundant.

If you would like to double check if the area you live still requires an antenna you may double check this map.

Requirements

In order to make the miner work you will need:

  1. Location with a clear view of the sky at 360°
  2. A stable 2.4 Ghz wifi connection
  3. A stable internet connection
  4. Electrical power
  5. A smart socket

Location

Geodnet has shared a few indications on how you should proceed in order to achieve an optimal installation. Please remember that everyone should try to get the best possible installation as it is of paramount importance that the quality of the data that we create and share is near to perfection. Otherwise the project is not useful for the end user.

The antenna should have a clear view of the sky in all directions around it. Please find here below some infographics which I have taken from the Geodnet Discord channel.

This picture shows that the antenna needs a clear view of the sky. It has 10 degrees of tolerance.
This pictures shows an example of a bad installation. The antenna is surrounded by tall buildings and trees which stop the signal coming from the satellites. Such an installation is to be avoided as it damages the project and its reputation.
The antenna circled in orange is a half way situation: there are tall buildings and trees around it but they are not totally covering the sight of the sky like in the previous example. The best situation is the second one highlighted by the green circle, where the antenna has no obstacles around it.

Before buying the antenna and miner I suggest to have an idea of good spot where to install it: even better would be to personally go to the site and verify the situation. Remember: you are looking for a position where the antenna will have a clear view of the sky. An application like GNSS View by NEC corporation (Android) may help you understanding the amount of satellites that your miner would be able to connect to: consider that in order to give a perfect service you miner will need to connect to at least 30 satellites at all times.

Stable networking

As stated above a stable internet connectivity is also mandatory: the miner does not require much bandwidth or speed, an ADSL with 25/2 mbits/s is more than enough to upload the data. More important is the fact that the internet connectivity should be stable.

The miner sadly does not dispose of ethernet connectivity: this is bad, because we know that wifi signal is not that reliable but at the same time it is good because you do not need to pull an dedicated ethernet cable for it to work. You just need a stable wifi network working on the 2.4 Ghz frequency. 5 Ghz is not supported.

The amount of data uploaded is estimated at 15 GB per month. But this is a scenario that I have never seen on my install which sends 5GB worth of data every month. 15GB/month is the figure which is shared by Geodnet considering a triple band miner, most probably to be considered the “worst case scenario”.

Electrical power

In order to work the miner needs electricity around 1 Kwh/month. On such topic I might suggest you buy a stable power supply. It does not need to be particularly powerful, 3A on the 5V are more than enough.

I suggest you buy also a smart socket: every now and then it might happen that the miner needs to be turned off and on again. If this is the case, having a smart socket that you can manage remotely from your phone is very convenient.