Development Blog
After our last workshop we decided it was time to tackle the mechanical design while the electronics were getting reworked. So we went through the design and made a list of everything we want to change. We also reached out to our users and tried to find pain points and points of failure. After talking to lots of people we boiled it down to this list.
- Protect the MS5803 and EC sensor from impacts
- Remove strain from the cables and rethink face plate/ structure
- Make sensor alignment repeatable during deployments
- Refine turbidity cover
- Prevent leaks
- Increase battery life
So overhauling the Smart Rock internals we did. We started with a new internal design. Using acrylic cement we created a structure made of 6 pieces that are fused together replacing the original 3 piece sled and separate face plate.
With the new structure we are able to double the battery capacity of the Smart rock along with improved positioning of the turbidity sensor. The electronics now have a dedicated strap to prevent the boards from separating and the cables sit nicely in the channel running through the center of the core.
We separated the EC probe and the MS5803 to their own pass-through because the face plate on version 3 had the potential to leak if impacted or assembled incorrectly. We are still testing different meshes/ covers for the sensors to prevent direct hits but not interfere with the EC sensor. We also added the small tab sticking out of the side of the core to allow it to index with a small grove on the enclosure. The index allows the core to be removed from the enclosure, maintained, and then returned to the exact same orientation/ position. Overall this change gives the smart rock increased capabilities while shrinking the volume it takes up inside the enclosure making more room for descant.
The enclosure we kept the same with the idea that we want to make upgrading old sensors as easy as possible. However we wanted to add an index to the outside of the enclosure so that when deploying the sensor you are able to maintain the sensor and have a consistent deployment position.
Then using a 2.5" U-bolt to mount the smart rock to the mounting plate. The mounting plate uses small U bolts to mount to rebar that can be hammered into the steam bed. The mounting plates we have tested so far have been made from 1/4" acrylic however we hope to test metal versions for a stronger mount and more impact resistance.
With these changes made, our last point of leakage seems to be the interface between the core and the O-ring of the union fitting on the enclosure. We have found that increasing the friction in the threads with wide teflon tape worked it had to be reapplied and was annoying to get undone at times. We have since moved on to try and better understand how tight the union fitting needs to be. It seems that with the proper torque the fitting may be able to be tightened where it doesn't leak. How tight someone is able to get the union fitting should be the same for everyone so we are working to find or manufacture a tool that will assist with providing proper grip. Over all we are super excited to have all of these changes working together. If you have a version 3 or 3.5 of the smart rock upgrading to version 4 will just involve migrating your feather and hypnos to the new core and modifying your enclosure to have the indexes. More information on procedures and testing results to come.
- Colin Hale-Brown (halebroc@oregonstate.edu)
After getting our new 6.35mm and 3.00mm cast acrylic in, we realized that some slots and tabs need to be adjusted accordingly. After finding our laser cutter kerf, we adjusted the dimensions of our acrylic so that the tabs fit snugly in the slots.
We finished a ~month long deployment of a fully assembled SmartRock version 4.0 to test how long we could theoretically deploy a device for. Usually, the devices are programmed to take data every 15 minutes, but for this test, we programed it to take data every 30 seconds. Using a multimeter, we measured a fully charged Lithium Ion Battery Adafruit - 3.7V 10050mAh to be 4.01 Volts, and then to verify, the first line of our .csv file's Analog Vbat reading is 4.004150391 Volts. The SmartRock stopped recording at 3.572314501 volts on 4/6/2024 so for simplicity's sake, we are going to assume we were able to get about a month's worth of lifetime. Extrapolating it out to