RoadTest: Raspberry Pi 2/3 Weather Sensor Shield
Evaluation Type: Development Boards & Tools
Did you receive all parts the manufacturer stated would be included in the package?: True
What other parts do you consider comparable to this product?: Weather Station
What were the biggest problems encountered?: Influence of the Raspberry Pi board on temperature and humidity sensors. use/ blocking the whole RPi GPIO connector.
The MEAS Pi weather shield from the point of view of electronic design is relative simple device.
We have the set of four sensors:
temperature (thermopile) - contactless.
all connected to supply line as well as I2C bus.
It has standard 40 pin header compatible with Raspberry Pi GPIO connector.
The form factor responds the RPi "second layer" board. It can be even closed in most RPi cases but from the point of view of measurement accuracy it doesn't make any sense.
Moreover, the MEAS Pi fits to any system that utilizes RPi compatible expansion ports configurable for I2C communication.
Unfortunately, after inserting the board, all the GPIO connector is occupied - blocked for additional wiring etc.
I wish the weather shield is of "through hole" type, that would make remaining pins available
The MEAS Pi weather shield provides the necessary hardware to interface:
the HTU21D digital relative humidity sensor;
the MS5637 digital barometric pressure sensor;
the (TSYS)01 digital temperature sensor;
the TSD305-1C55 digital thermopile sensor.
HTU21D, MS5637, TSD305 have also additional temperature sensor fully accessible using I2C.
Receiving the package I realized, that the MEAS Pi weather shield is shipped as simple electronic component. It is packed in "classic" silver/ anti ESD bag. No box, no extra stuff typical for electronic devices.
I didn't find any instruction nor manual in the package. To be honest I was surprised by the lack of setup instruction etc. On the other hand, I must admit, that the shape of the board, the characteristic interface should make the montage easy and intuitive. Nevertheless If you need the support in the matter of installation and setup, go to the TE site. You can realize, at last, that Weather Shield came with quite comprehensive documentation available on TE page.
Details here: http://www.te.com/usa-en/product-CAT-DCS0036.html
Surfing on that page, under the "Related Materials" I found, that there are official links to both GITHUB and IBM Bluemix projects based on the board.
These links are available also on the datasheet:
Python framework and example source code available on Github.
IBM Bluemix recipe available on developerWorks.
At last I decided to launch the github project and put the life to my weather shield
4 temperatures + contactless temp. measurement + humidity and air - barometric pressure.
Not much (well, well temp makes the big number!) but it's good enough for several experiments!
Here are my results from the first approach:
I simply installed libraries from GitHub and run the python code from there, easy and simple at all! - nothing more, if you remember to set i2C in your RPI of course.
Following you can find several screens from my RPI weather station setup:
The results of basic python code for reading the sensor measurements, below:
Just adding the code for timestamp, the basic target is achieved now!
Temperature TSY: 29.0
Temperature HTU: 30.2°C
Temperature TSD: 30.2°C
Temperature 5637: 30.5°C
Object temperature: 19.7°C
And here I have the first surprise - why TSY01 differ almost 10%? well, OK it's separated a little from the rest of the board. But the influence of the Raspberry heating center is visible. The ambient temperature in the room mustn't exceed 21°C according to independent, accurate measurements!
The "object" is the ceiling - I supposed its temperature should be below 19°C! but in this case I can accept the results.
I have suspected that there is the influence of the board (in the meaning of the heat) on the readouts from temperature sensors.
TSYS0101 temperature sensor seems to be slightly separated from the board, thanks to "cutting-ready" solution
I decided to spoil the board a little and break-out this part, separate it from the board and generally minimize the influence of RPI thermodynamics...
Breaking out the TSY part from the board and extending it using the wires, improved the reliability/ credibility of the readouts!
Now I have one remark as the soldering is not easy here, the paths are narrow, the best solution is to press tightly 4pin connector.
Having the unit improved I started to Think about its readouts collecting method and... ThinkSpeak solution came to my mind.
Thanks to that idea, It was extremely easy to log data from my humble Raspberry Weather Station.
Above results came from simple solution based on "one line" python code and use of cron in RPi
Now - let's compare barometric pressure chart from my Pi to neighbourhood weather station
above my RPi+Weather Shield station readouts (pressure)
below, the reference pressure chart
comparison of the above charts
Temperature sensors accompanying the HTU21D; MS5637; TSD305-1C55 give almost 30°C - I treat it as the relative board temperature. And frankly, these three, quite similar measures - it look quite reliable... (<3%error)
I tried to analyse a little my temperature data.
In the period of 7 days I get following results
I checked the range of results:
maximum TSD temperature value is 28.83°C
minimum HTU temperature value is 26.08 °C
but maximum difference between "hottest" TSD and "coldest" HTU sensor is 1.424°C with the minimum difference of 0.72°C
Introducing the mean value, I calculated the maximum deviation equal to 2,57%
It's rather risky, to state that above calculation shows the measurement error level. But even if , the calculated error is relatively small and fully acceptable. Especially taking into account that I have analysed ancillary temperature transducers.
Well, after two months of adventure with Raspberry Pi 2/3 Weather Sensor Shield I should be ready to give the resume and some conclusions
I need to share my doubts regarding manual and documentation at all. Honestly, the first impression and the lack of instruction in the package caused some hesitancy in evaluation of that matter.
Maybe I'm too harsh maybe my bad habits with extended expectations regarding setup instructions/ manuals as well as shipment and packing standards lowered the assessments.
On the other hand shipment was safely delivered, then I could start my successful experiments soon. I found all necessary instructions and even examples on the net.
Another subject is my disappointment with temperature/ humidity measurements disturbances related to the influence of the RPi mainboard. The construction of the board, its shape and basic montage idea is wrong in my opinion. I decided on brutal move with breaking-off the temperature sensor. After all I wish I use 40 wire cable (instead of 4) and run the weather-shield in "adiabatic" separation from RPi. But the idea of compact set of RPi + weathershield would be lost. This case is still bothering me.
I would repeat that in my opinion the idea of use the whole 40-pin connector is wrong in the device which uses only power + I2C bus pins!
Unfortunately, after inserting the board, all the GPIO connector is occupied - blocked for additional wiring
Leaving aside the price issues, I timidly suggest to enrich the module in GPS module. That would be valuable signal for amateur "mobile" weather station which uploads weather data in the net.
Perhaps, an additional extension for rain/ snow level measurement (external sensor), would be an interesting option that would made our "shield" almost professional solution for a weather station.
Nevertheless these four transducer are good enough for simple, amateur applications, particularly, the quality of sensors exceeds amateur level. I'm glad to have the shield. I hope to continue the experiments exceeding the field of use beyond the realm of RPi. I will direct my next steps to the Arduino.
Thank you for the confidence and giving me the chance to conduct this test.
Thanks for the review, its nice to see how your testing progressed and the issues you found with the sensor being heated by the Pi. I guess that was to be expected, and it's a good job they designed…
Thank You for these words. At the beginning I should apologize for... let say "falstart" with publishing the report. Unfortunately this is not the first time, I hit wrong "button" - [publish] instead…
Nice review, you methodology follows exactly what I'm trying to do. The main issue I'm having is the address of the TSD305 is 0x00. The Raspberry Pi and all Linux Distros I've looked at for this…
Yeah I know what you mean, I've been caught out with the accidental publish before too! I wish they'd put a confirm dialog on the 'Publish' button when publishing an article for the first time!
I also don't think you are being harsh, you have to go with how you feel about the products performance etc and so long as you can back up the scores with reasons then there is nothing wrong with having them lower.
I wish you all the best for 2018 too!
Nice review, you methodology follows exactly what I'm trying to do. The main issue I'm having is the address of the TSD305 is 0x00. The Raspberry Pi and all Linux Distros I've looked at for this dont allow access to 0x00 as that is a reserved address. This leaves me wondering how did you manage to read this device. I'm still working to get my review out so any help with this would be very much appreciated.
Thank You for these words. At the beginning I should apologize for... let say "falstart" with publishing the report. Unfortunately this is not the first time, I hit wrong "button" - [publish] instead of [save draft]...
I tried to support my hasitations regarding "Product Expectations", "Demo" and "Support".
As written above, maybe I'm too harsh but generally speaking these are my first impressions. I decided to keep it, despite of finding supporting materials in the net.
I wish you all the best in 2018!
Thanks for the review, its nice to see how your testing progressed and the issues you found with the sensor being heated by the Pi. I guess that was to be expected, and it's a good job they designed it so the sensor can be broken out and wired further away. You got some good test results in the end!
I see you've been updating this a lot over the last couple of days and I can see how it's been progressing. I don't know how much more you are planning to do, I was wondering if you were going to write up a summary/conclusions at the end to bring everything together? It might be worth commenting on how you derived your scores a little, there's nothing wrong with them, it's just scoring can be a little subjective sometimes and when some sections are marked at 5 or 6 it's maybe useful to follow up with some reasoning for these and what could be done by the manufacturer in future to enable these to improve.