compareeHub

YOU CAN BUILD THE PROBE A DRONE DROPS INTO AN ARCTIC FJORD - AND UNLIKE THE COMMERCIAL ONE, IT COMES BACK

You can build the probe a drone drops into an Arctic fjord—and unlike the commercial one, it comes back.

by Ebbe Poulsen, Mathias Eggertsen, Erik H. Jepsen, Claus Melvad, Soren Rysgaard

FULL CAD BOM FIRMWARE DOCS

ScienceOpen-hardware

Built withTeensy3D printing

difficulty
●●●●●
time
a month
license
CC-BY-4.0
repo
repo FINISHED0 stars

WHAT YOU’LL NEED

1
Jump to section

COMPAREE VERDICT

The ARC-TOP is a recoverable oceanographic profiler designed for high-risk deployments where you cannot put a ship—glacial calving zones or anywhere the bottom is worth more than the tool. It weighs 2.6 kg, costs €3,153 in published component prices, is rated to 250 metres, and the authors tested it in August 2021 at two glacier fronts in NE Greenland. The core idea is simple: drop it from a drone, let it sink while logging conductivity, temperature and pressure, then release the ballast by heating a polycaprolactone rod with a small heating element. The probe surfaces, transmits a LoRa position beacon, and gets picked up by the drone. The thing most likely to go wrong is the sealed hull: the aluminium pressure hull and potted cable penetrators must hold at depth, and the design includes leak sensors for exactly that reason. The second trap is calibration. The sensors are low-cost off-the-shelf parts (Atlas Scientific conductivity probe, Blue Robotics temperature and pressure sensors), and the authors ran their own calibration in a tank with known temperature and salinity to get research-grade numbers. If you have access to a test tank, a reference instrument, and a reason to profile water columns in places a ship cannot go, this is a legitimate tool and the documentation will get you there. If you wanted a weekend sensor project, this is not it.

GOOD TO KNOW

  • —Full assembly instructions, circuit diagrams, 3D print files, Arduino code and a validated bill of materials are in the HardwareX paper and Mendeley Data repository.
  • —The paper is unusually complete: it includes the operating concept, assembly steps, the calibration procedure, a side-by-side test against commercial CTDs (no significant difference after calibration) and a list of suggested improvements.
  • —Licence is CC BY 4.0, commercial use permitted.
  • —This is a research instrument build, not a weekend kit. You will machine aluminium parts, calibrate conductivity sensors, tune ballast to local salinity, and test the sealed hull for its 250 m rating.
  • —The conductivity sensor is an off-the-shelf Atlas Scientific probe and EZO circuit. Inside the pressure hull, two custom PCBs handle control, connectors and electrical isolation; the data repository has four KiCad PCB projects in total. Calibration is up to you.
  • —No specific drone is named. The mounting and release hardware is 3D printed and can be adapted to other UAVs; you need lift for 2.6 kg and three free servo outputs.

Parts to buy

11 items

From our check of the build. Exact quantities and part numbers are in the creator’s BOM.

  • Teensy 4.1Find
  • Atlas Scientific Mini Conductivity Probe K 1.0 with EZO circuitFind
  • Blue Robotics Celsius temperature and Bar30 pressure sensorsFind
  • Quectel L86 GPSFind
  • RFM98W 433 MHz LoRa modulesFind
  • SD cardFind
  • Heating elementFind
  • PCL rodFind
  • BallastFind
  • Cable penetratorsFind
  • Leak sensors and a 4-cell 18650 battery packFind

BUILDS OF THE WEEK

Five open-source builds worth your weekend, every week.

Checked like this one: what’s really in the repo, what it costs, how hard it is. One email, unsubscribe anytime.

Can I build this?

PrintThe top and bottom hull sections, internal frames and the UAV mounting hardware, printed in PETG on a consumer FDM printer. The printed hull is a flooded design: only the machined aluminium pressure hull is watertight, and 3 mm carbon fibre rods tie the printed sections together.
BuyTeensy 4.1, Atlas Scientific Mini Conductivity Probe K 1.0 with EZO circuit, Blue Robotics Celsius temperature and Bar30 pressure sensors, Quectel L86 GPS, RFM98W 433 MHz LoRa modules, SD card, heating element, PCL rod, ballast, cable penetrators, leak sensors and a 4-cell 18650 battery pack. The full BOM in the data repository lists suppliers.
Tools3D printer, soldering station, multimeter, access to a machine shop for the aluminium pressure hull (hard anodised), potting supplies and a vacuum pump for the penetrators, a pressure test option, a calibration reference, and a drone with three free servo outputs.
SkillsThis is a research instrument build. You need circuit assembly and debugging experience, familiarity with I2C sensor integration, confidence machining and threading metal parts, and the patience to iterate waterproofing until the seals hold at depth. If you have never built a waterproof enclosure before, this is the wrong first project.
TimeThe paper does not give an assembly time. Realistically: a weekend for initial assembly, a week for firmware integration and bench testing, then an unknown number of iterations on the waterproof seals and ballast tuning. Budget a month if this is your first pressure-rated build.
CostThe paper gives 3,153 euros ex. VAT, based on producing two sensor units and one set of UAV mounting hardware, without development and assembly time. The custom-machined aluminium pressure hull accounts for 45 percent of that; the Atlas conductivity probe and EZO circuit together are about 155 euros. The drone, pressure testing and calibration equipment are not included.
SafetyLithium battery in a sealed enclosure underwater—if the seals fail and water reaches the pack, you have created a device that can vent or ignite at depth. Test the housing dry, then with ballast, then submerged in a controlled environment before any field deployment. The paper recommends a leak detection circuit for exactly this reason.

Build at your own risk. Projects involve tools, electronics and sometimes mains voltage — follow the creator’s safety notes.

More builds like this

All projects
The Enactive Torch: an open-source device that turns distance into touchScienceYou can build an open-source algae photobioreactor with a built-in fluorometerScienceYou can 3D print a microscope that sees individual atomsScience

Gallery

sciencedirect.com
sciencedirect.com
sciencedirect.com
sciencedirect.com
sciencedirect.com

Start here

Navigation into the creator’s own docs — we don’t rewrite the guide, we route you to the source.

  1. 1.Read the full HardwareX paper (The paper is the build guide. The Hardware description section covers the hull, buoyancy engine, pick-up system, electronics and software; it is followed by the design files, the bill of materials, step-by-step build instructions, operation instructions and the validation tests.)
  2. 2.Download the design files from Mendeley Data(STEP models, KiCad projects for the four PCBs, Arduino code and the complete BOM (Naming_and_BOM.xlsx) are in the Mendeley Data repository referenced in the paper (DOI 10.17632/zdvb5hzv2x.1). You will need those before ordering parts.)
  3. 3.Source a pressure test chamber or find a facility that will let you use one(The authors tested the hull and penetrators to 250 m. If you cannot verify your seals hold at the depth you plan to deploy, do not deploy.)
  4. 4.Build and bench-test the electronics before any waterproofing(Get the Teensy logging clean sensor data to the SD card on your desk before you seal anything. The firmware is provided but assumes familiarity with Arduino and I2C debugging.)

KNOWN ISSUES

  • The paper does not name a specific drone. You need a UAV that can carry 2.6 kg, drive the three servos the release and pick-up hardware use, and hold station in the wind you will actually face at a glacier front. The mount is 3D printed and meant to be adapted.
  • The conductivity sensor is a low-cost Atlas Scientific probe, and out of the box its accuracy is not research grade. The authors ran a two-point calibration in a tank with known temperature and salinity. If you cannot do the same against a reference, treat the salinity data as indicative only.
  • The sealed aluminium hull and potted cable penetrators are rated to 250 m, and the design carries four leak probes because a sealing failure ends the instrument. Budget time for potting and test cycles, and test your own hull at depth before a field deployment.
  • The ballast release melts a 2.8 mm polycaprolactone rod at about 60 °C with a PID-controlled heating element, and the firmware retries if the depth does not start decreasing. It is the one mechanism that brings the probe back, so test it repeatedly on the bench and in shallow water before a real dive.
  • The 3,153 euro figure is from the 2022 paper, excludes VAT, development and assembly time, and covers the instrument and the UAV mounting hardware, not the drone itself, a pressure test or calibration equipment. Budget for those separately.
  • This is not a beginner project. If you have never built a waterproof sensor package, debugged I2C sensors, or worked with lithium packs in sealed enclosures, start with something simpler. The paper assumes you know how to do all of those things already.

Can I skip the pressure test and just seal it well?

No. The pressure hull is designed for 400 m with a safety factor of 1.3, but the authors only tested hull and sensors to 250 m, and they still fitted leak probes that trigger a ballast release. The only way to know your own housing is sound is to test it at pressure before a field deployment.

What happens if the LoRa signal does not reach me when it surfaces?

The paper does not give a backup recovery plan beyond visual search. LoRa range depends on line of sight, antenna quality and local interference. If you lose signal, you lose the probe. This is why they designed it to be affordable enough to risk.

Can I use a different microcontroller instead of the Teensy 4.1?

You can, but the firmware is written for Teensy and uses its specific I2C and SD libraries. Porting to another platform means rewriting the sensor integration and data logging, which is not trivial. If you are asking this question, use the Teensy.

How do I calibrate the conductivity sensor without a lab salinometer?

You need some reference. The authors did a two-point calibration in a 150-litre tank with known temperature and salinity, using a Guildline Portasal salinometer as the reference, and got an expanded uncertainty of 0.07 mS/cm. Without a reference instrument or known solutions, treat your salinity readings as uncalibrated.

Community builds

No community builds yet — be the first, we feature the best ones.

Discussion1

FROM THE COMPAREE TEAM

The paper lists a build cost of 3,153 euros, a 2.6 kg instrument rated to 250 metres, and field tests at glacier fronts in NE Greenland in August 2021, but it leaves the choice of drone to you. What would you fly it on?

CompareeTEAM15d agoedited

Practical notes from our verification: this is one of the most complete oceanographic hardware papers we have catalogued. The data repository linked from the paper contains the full BOM (Naming_and_BOM.xlsx) with suppliers and prices, STEP models of the assemblies, four KiCad PCB projects (two of the boards sit inside the pressure hull) and the Arduino firmware. The paper does not name the drone model: the profiler is designed to hang under any suitably sized UAV with connections for three external servos, and the mount is 3D printed. Sealing is taken seriously: the pressure hull uses a double O-ring seal between endcaps and hull, mechanical testing of the hull and sensors was limited to 250 m, and Blue Robotics SOS leak probes at four points on the bottom endcap trigger a ballast release if water gets in. The low-cost conductivity sensor was calibrated after the cruise in a 150 L tank against a Guildline Portasal salinometer. If you have the skills and equipment, the documentation will get you there. Correction (4 October 2026): we re-checked this page line by line against the project's own repository, documentation and videos, and fixed errors in earlier versions.

Ebbe Poulsen, Mathias Eggertsen, Erik H. Jepsen, Claus Melvad, Soren Rysgaard

Poulsen and team are at the Department of Mechanical and Production Engineering and the Arctic Research Centre, Aarhus University. Very few measurements exist close to calving glacier fronts because of the risk. Earlier work used single-use XCTD probes dropped from helicopters, limited by the cost of replacing equipment and hiring the helicopter. ARC-TOP was designed to be dropped by drone, profile autonomously and be picked up again.

Star the project on GitHub

DISCLAIMER

  • Comparee is not the author of the projects featured here. All rights to each project belong to its creator — every page links to the original source, and we never host creators’ files.
  • Information is provided without warranty and may become outdated as projects evolve. Prices are indicative bands only — always check the creator’s parts list for current costs.
  • Building and operating any project is at your own responsibility. Protective equipment, safe workshop practice and compliance with local regulations are the builder’s responsibility.
GitHub repo