How do I collect data from a 3x7 grid of (21) FS3000-100 air velocity sensors?

I want to get individual readouts, which probably can be done using individual arduino boards with readouts. But how do I get those numbers into a single board and get the timed average over x minutes?

In outline, send to the “single board” a measurement along with its sensor identifier, collect those in an array indexed using the identifier, and simply calculate the average.

The details depend on how you design and configure the system, so tell us more about that.

jremington:
In outline, send to the “single board” a measurement along with its sensor identifier, collect those in an array indexed using the identifier, and simply calculate the average.

The details depend on how you design and configure the system, so tell us more about that.

The design is dependent on the quiic system? Since this is the only air velocity sensor I could find. I've only worked with arduino and raspberry pi boards before. I would have connected all of the 21 sensors to RPi if that was a possibility. Apparently I2c addresses will be an issue too, right?

If you’re asking about the physical grid, the sensors will be tied to a metal mesh. I am trying to get the timed average air velocity across a fume hood face. On a side note, there’s also a requirement for easy calibration of these sensors. Hope I could explain it.

I assume you are considering using the [Sparkfun Air Velocity Sensor Breakout - FS3000-1005 (Qwiic). That has a fixed I2C address. So if you want to have 21 of them you’ll need 21 separate I2C buses! That is a lot of buses. Multiple software buses probably won’t work too well (you’ll need 21*2 free digital pins), so you’ll be needing a I2C multiplexer of some sort. I don’t have any experience with it, but perhaps something like 3 of [Adafruit TCA9548A 1-to-8 I2C Multiplexer Breakout will work (each one gives you 8 I2C buses, so with 3 multiplexers, you will have 24 I2C buses: a few more then you need).](Overview | Adafruit TCA9548A 1-to-8 I2C Multiplexer Breakout | Adafruit Learning System)](https://www.sparkfun.com/products/18377)

What does “requirement for easy calibration of these sensors” mean?

The only way to accurately calibrate those sensors requires placing them in a calibrated, laminar flow, constant temperature wind tunnel.

jfbauer:
I assume you are considering using the [Sparkfun Air Velocity Sensor Breakout - FS3000-1005 (Qwiic). That has a fixed I2C address. So if you want to have 21 of them you’ll need 21 separate I2C buses! That is a lot of buses. Multiple software buses probably won’t work too well (you’ll need 21*2 free digital pins), so you’ll be needing a I2C multiplexer of some sort. I don’t have any experience with it, but perhaps something like 3 of [Adafruit TCA9548A 1-to-8 I2C Multiplexer Breakout will work (each one gives you 8 I2C buses, so with 3 multiplexers, you will have 24 I2C buses: a few more then you need).
[/quote]
That’ll be a good start :slightly_smiling_face:. I was also looking at anther forum and they basically suggested powering on/off each sensor somehow and read from only one sensor at a time.

jremington:
What does “requirement for easy calibration of these sensors” mean?

The only way to accurately calibrate those sensors requires placing them in a calibrated, laminar flow, constant temperature wind tunnel.

I have a TSI hot wire anemometer that goes through yearly calibration. My solution is to see how close each of the 21 sensors read to that independent TSI instrument and input the deviation in the code? What do you think? Guess that's a far cry from my 'calibration requirement' from earlier.](https://learn.adafruit.com/adafruit-tca9548a-1-to-8-i2c-multiplexer-breakout)](https://www.sparkfun.com/products/18377)

My solution is to see how close each of the 21 sensors read to that independent TSI instrument

I don’t see how that can be a “solution”.

What is your plan to distinguish genuine differences in air velocity across the opening, from differences due to errors in individual sensor calibration?

Not sure if I can ask this in the same post…

is there a way to connect the FS3000 sensor to a raspberry pi?