Public Lab Research note

Turbidity Sensor Assembly

by MadTinker | March 26, 2018 17:44 26 Mar 17:44 | #16019 | #16019


This research note documents the build and deployment of a turbidity sensor with a data logger. The data logger is a 5V Arduino pro-mini. The turbidity sensor is the DFRobot SEN0189. The unit was assembled and deployed in conjunction with the Electrical Conductivity (Total Dissolved Solids) sensor and data logger described here. The goal is to get both of these in the stream before construction starts up-stream on a new flume and also before spring run-off along Willow Creek in Creede, CO.

Data Logger

Construction of the data logger is pretty straightforward. The only difference from the the other data loggers common at Public Lab (e.g. Cave Pearl or Mini-Pearl) is that this one is required to be 5V due to the power requirements for the turbidity sensor. This means a change in the Arduino processor from 3.3V to 5V and an upgraded microUSB card that handles both 3.3V and 5V. You can follow instructions provided for the Arduino pro-mini 3.3V provided by either Ed Mallon (Cave Pearl) or Chris Fastie @cfastie (mini-pearl).

[Will add Fritzing picture of new connections when I learn Fritzing]

Parts List

This is just a representative sample of sources. A little bit of searching on the net will likely result in a variation of sources and prices.

Micro SD Storage Board SPI for Arduino ($0.77)

DS3231 RTC/AT24C32 eeprom ($1.18) Note power saving modifications suggested here

Arduino pro-mini 5V ($3.00) This package contains the FTDI Adapter

FT232RL FTDI to TTL Serial Adapter ($1.78)

DFRobot SEN0189 Turbidity sensor ($9.90)

Battery pack ($1.99)

4x AA ($2.50)

CR2032 coin battery for RTC ($0.29)

Misc pipe, coupling, etc for packaging ($12.50)

image description image description

Turbidity Sensor

Inside of Turbidity Sensor: mostly a couple of LEDs. Definitely not waterproof - per the warning by DFRobot.

image description image description

image description

The sensor was removed from its shell and the outer shell was epoxied into a 3" pipe plug. A small dab of epoxy was used to hold the sensor to the inside of the outer shell. (The black outer cover would not fit into the recess.)

Software coding was easy, but hardware is currently being contentious ...

Status 4/14: Still waiting on response from DFRobots on lack of analog voltage response from sensor.

Status 4/15: Problem solved and ready for initial deployment

image description image description

Some might recognize the small circuit board on the bottom right as containing the start of Coqui-type electrical conductivity circuit for a STEM class on water quality being planned. Waiting on a speaker...

Initial Testing




Preliminary Data Analysis



Hmm! Maybe @cfastie could help with this issue? Very cool to see this in-progress!

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

@warren Thanks. Perhaps, but if the info in the link is true, the solution is just finding a new RTC module that runs at 5V. I misread the power requirements for the sensor and thought it was 3.3v and built a standard 3.3v mini-pearl. So I've got to start over. The good karma is that an eBay supplier mistakenly sent me 10 5V pro-mini boards that must be 100 years old. The negative karma is that, while the mini-pearl RTC modules are rated at 5V, ugly things appear to happen when they are overpowered. I haven't even started to look at alternative RTC modules; maybe tomorrow.

Reply to this comment...

Ah, Arduino. Those DS3231 RTCs are designed to be used with a rechargeable coin cell (LIR2032) and they trickle current into the battery when current is supplied. Ed Mallon's standard procedure is to remove the 201 resistor (flip it off with a soldering iron) and use a non-rechargeable coin cell (CR2032).

Remove the resistor on the right to disable trickle charging of the coin cell. Removing the one on the left disables the LED.

Ed Mallon typically uses these with 3.3v Pro Minis, but he also uses them with 5v Unos. He has not reported any problems running them at 5v. I have used one of these RTCs with a 5v Nano without issue, and I was powering the Nano with a 9v battery. But that is not to say that the coin cell will not explode the next time I power it on.

I don't remember if the RTCs I sent to Dave had the resistors removed or not. Now I always remove them before including an RTC in a kit.

Dave is correct that a 5v sensor cannot be run from a 3v Pro Mini (without fancy mods), so the Pro Mini might be the wrong platform. However, your 5v Pro Minis could work if the RTC will tolerate 5v. Although the Arduino Forum posts suggest that those RTCs will not tolerate 5v, Ed Mallon suggests they work at either 3.3v or 5v.

You might be able to use a 5v Nano which conveniently has a 3.3v output pin (VCC is 5v), so you can power modules which need either voltage. But Nanos use a lot of battery power and will not last long in the field.

I have used those RTCs without removing the resistor for deployments of many weeks without issue. I was powering them at 3.3v, not 5v, but the trickle current into the CR2032 coin cell did not explode it. So there seems to be a lot of variability in how those RTC behave.

So there are multiple solutions depending on your requirements for battery life and whether you want to test your faith in the 5v tolerance of those RTCs or wait to get another type of RTC.


Reply to this comment...

@cfastie Thanks Chris, I'd had forgotten this particular mod by Ed Mallon, and didn't notice the mods on your modules. I need to go back and review those discussions more closely. The RTC modules I am using are not the ones I rec'd from you. When I get back to this, it won't hurt for me to highlight the mods again on the research note. Tentative data logger package: 5V pro-mini, mod existing RTC, (maybe) drive with 9V battery.

Reply to this comment...

That will be an interesting project to make a 5v Pro Mini data logger. I guess it will also include a 5v microSD module (the ones in the Mini Pearl Logger Kits are 3v).

I was using a 9v battery for that Nano-based logger because it included a Modern Device wind sensor which requires 8v. There is generally no good reason to use 9v batteries for this type of logger unless you need more voltage than the main board needs. Powering a 5v Pro Mini will probably work well with four alkaline batteries (4 x 1.5 = 6v) or six NiMH (6 x 1.2 = 7.2v).

Reply to this comment...

Wow, this is a cool project. By the way, have you solved the problem that the RTC module can not be compatible with 5V? I also used this RTC module with 5V voltage. It works fine. According to the datasheet, the DS3231 chip can work fine with 2.3~5.5V. And in order to charge the battery onboard, 5V voltage is needed.

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

@youshen I think with the modifications that have been suggested, the particular RTC module that I've been using will work fine and support running at 5V. I'm going to disable the battery charging portion of the circuit.

Reply to this comment...

@MadTinker Great project and approach here. I just finished my first prototype of the same thing, only the sensor is tethered to a weather proof box via 10ft cable. Did you ever deploy this in CO? How did the data look? Wondering if this approach can lead to reliable measurements long term. My preliminary data suggests the data isn't all that reliable for long periods, mostly due to light leakage, scum buildup, and debris interference. Let me know how your deployment went and any data you care to share!

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Where for buy container to turbidity?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Login to comment.