Conduit Field Deployment

[Szymon Zieba]

Hi,

I have a field deployment set up for the Conduit prepared however I have found a problem when I attempt to use it to do range testing. It seems that at random (sometimes it works, sometimes it doesn’t) the Conduit will decide not to run Node-RED meaning that my program on the Conduit to record the results locally is not being run. This is problematic as it makes range testing difficult. In case you were wondering, the LoRa server appears to be turning on as the mDot is able to connect to the LoRa server, but the Conduit is not running any scripting as I noticed this while running tests with simple send/respond scripting. Are there any suggestions on what may be the source of this problem?

Also, I was wondering if there was clear documentation on the range limits for the mDot/Conduit combination with the different data rates factored in. At the moment I have found maximums of 1.27km within urban setting and 3.56km with rural testing which are both obviously below the expected rating from the advertising. It would really help out to know what conditions were tested to get the advertised distances.

[Steve Kovarik]

Hello Szymon

To determine if Node Red is running, you can check the process and view the log.

To do that SSH into the Conduit and issue the command ps -auxww | grep node

View the Node-Red log by going to var/log/app (then) cat node-red.log

Regarding range, Which advertised distances are you referring to?
I am assuming the Gateway and the endpoints have been configured for optimal
range. Maximum Transmit Power configured on Gateway and endpoints.
Maximum Spreading Factor (Data Rate) configured on endpoint for optimal
range.

Distance depends on conditions, configuration, antennas, desired throughput,
and usage frequency. In dense urban environments, a typical range is 1-2
miles. It is capable of 2-way single-duplex communication across distances
of up to 8 km line of sight, deep into buildings, or within noisy
environments.

With the Conduit indoors, centrally located on 2nd floor within a brick building
with the hinged, right angle antenna shipped from the factory, I achieve
connectivity anywhere inside our 175,000 sq ft facility, and about 300 meters
outside in our urban location.

In the same scenario with an antenna mounted on the roof of the 2-story
building we achieve 3km to 5km range depending on obstructions (leaves,trees
buildings, hills).

I have customer with an antenna on top a 6-story building successfully
covering an entire college campus.

I have another customer that achieves 13km line of sight, with the antenna
elevated at about 300 ft.

Multitech manufactures a MTDOT-BOX and MTDOT-EVB for the purpose range
testing. These devices have a program that can perform a 16 point site survey
cycling through a range of transmit power and spreading factors (Data rate).
These useful tools will store survey data in the device as well as send it
to the Conduit Gateway.
http://www.multitech.com/brands/multiconnect-mdot-evb

-Hope this is Helpful
-Best Regards

[Szymon Zieba]

Hi Steve,

The problem with Node-RED is that I need to have assurance that it turns on consistently as it severly hampers testing if it doesn’t automatically activate the script when turned on. Is there an easy way to turn the Node-RED script into a shell script, or a way to generate a shell script that can listen to the LoRa ports for incoming information and log it that way?

With range, the advertised ranges I am talking about are the ones listed in the mDot data sheet, where it states a 1-3mile (or 2km) in Building Penetration and 16km maximum range. We weren’t expecting to get the perfect results due to signal attenuation, however we were attempting to get at least 5km (ideally 10km) in our rural test from one of the highest points easily available to us, which is about 85m above sea level going off Google Maps. As previously stated, the best result that was obtained was ~3.56km during our testing. This was on a signal that was alternating through DR1 to DR4 and sending GPS Data to the Conduit. Thank you for the information of previous tests, however, it at least gives us an idea of the limits.

As for the mDot Box, I have found that it appears to have a shorter range than the standard mDots and I am not convinced that the survey mode works correctly on the Box that we currently own. As such we created a simple mDot GPS device to send the distance data to the Conduit as this appears to be more effective than the Box.

Regards,

[Szymon Zieba]

Following on from above, I have implemented a Node-RED script which gives a heartbeat to whether any script is running. Basically I have the program turning on/off one of the signal lights on the front of the Conduit due to the lack of usage of these lights at this time. I then measured how long it took to have the Node-RED script begin running, finding that it took approximately 4 minutes to start running Node-RED script. Is there a way to improve this time, or should I just start writing shell script in order to reduce load times?

[Steve Kovarik]

Hi Szymon

Yes, that is normal. On power-up it can take up to 4-5 minutes before Node-Red is loaded and running. When used with LoRa, Node Red uses moderate processor resources. The latest Conduit firmware version 1.3.3 supports a new
feature that allows loading and running a custom application. This would
allow you to disable and not use Node Red, and create, load and run a custom
application that may accomplish the same. With Node Red disabled, you would
notice increased performance.
http://www.multitech.net/developer/software/aep/creating-a-custom-application/

-Best Regards

[Author]

Posts