This week wrapped up sprint 4. Some things didn\u2019t go entirely as planned, but we still made progress across both hardware and software. Here\u2019s a quick look at what everyone\u2019s been working on\ud83d\udc47
<\/p>\n\n\n\n
![\"\"](\"https:\/\/dronesonen.usn.no\/wp-content\/uploads\/2025\/10\/Hydrawlics-Logo-5.png\"){kind=logo}
<\/figure><\/p>\n<\/div><\/div>\n<\/div>\n\n\n\n
\n\n\n\n
Fredrik:<\/h2>\n\n\n\n
Since this week is the end of sprint 4, we were supposed to have a full hydraulic test. Unfortunately, this is not the case. I have been fully preoccupied with another course this week, which has led to very little focus on the hydrawlics project from my end.<\/p>\n\n\n\n
However, most of my work this week was things that could run in the background, that for the most part being 3d-printing components. I have been helping out the other members here and there, but my task was mainly to make the four cylinders, and the remaining cylinder attachment rods that I designed last week. This meant that I could do some of the busy work for this project, while also working on the other course. The printed parts are mainly the ones that I designed in week 6, that being the hydraulic cylinder components, with the design changes I made last week.<\/p>\n\n\n\n
![\"\"](\"https:\/\/dronesonen.usn.no\/wp-content\/uploads\/2025\/10\/Printing-1-768x1024.jpg\")
<\/figure>\n<\/div>\n\n\n\n![\"\"](\"https:\/\/dronesonen.usn.no\/wp-content\/uploads\/2025\/10\/Printed-components-1-768x1024.jpg\")
<\/figure>\n<\/div>\n<\/div>\n\n\n\nA while back I bought a long tube of diameter 20mm pvc pipe at Clas Ohlson, and used this as the base of the hydraulic cylinder. Now that I know the exact dimensions and features of the cylinder, I can finally put it to good use. Me and Erling used the prototype lab at school to cut the pipe into 80mm segments with the bandsaw, and drilled 4.5mm holes on the tube surface with the column drill. Erling also designed his manifolds to use the same tube, so we made his segments as well while we were at the lab.<\/p>\n\n\n\n
![\"\"](\"https:\/\/dronesonen.usn.no\/wp-content\/uploads\/2025\/10\/Drilling-1-edited-scaled.jpg\")
<\/figure>\n<\/div>\n\n\n\n![\"\"](\"https:\/\/dronesonen.usn.no\/wp-content\/uploads\/2025\/10\/Pvc-pipes-1-768x1024.jpg\")
<\/figure>\n<\/div>\n<\/div>\n\n\n\nAs of the writing of this blogpost, I have not put the cylinders together. But the components are printed, the tube is made ready, and the epoxy glue is bought. The only part missing is buying some more o-rings. This means that early next week, the cylinders should be in place. So keep an eye out for the next weeks blogpost, as we might have been able to do a full hydraulic test of our system<\/p>\n\n\n\n
\n\n\n\n
Lisa:<\/h2>\n\n\n\n
After last week\u2019s test, where the valves and pump finally worked together as intended, I had one of those moments where everything looked great, but something still didn\u2019t fully click in my head. The system behaved exactly like it should, but I realized that I didn\u2019t quite understand<\/em> why.<\/p>\n\n\n\n So, this week was less about running new tests and more about connecting the dots. I wanted to understand better what actually happens inside the system, from the moment a new target angle is received, to the point where the arm moves, the valves switch, and the pumps’ state changes.<\/p>\n\n\n\n Initially, it was a bit confusing. I kept mixing up which relay did what, and the logic didn\u2019t seem to add up the way I thought it did. But after some helpful whiteboard sessions, and with some help from the team (and a few \u201cohhh, that\u2019s how it works!\u201d moments), things finally started to make sense again.<\/p>\n\n\n\n Here\u2019s the big realization:<\/strong><\/p>\n\n\n\n Each valve isn\u2019t just one channel, it actually uses two! One for Extend (E)<\/strong>, and one for Retract (R)<\/strong>. And the pump? It\u2019s its own little controller that only turns on when the system is in motion. No movement, no pump. Then there\u2019s the potentiometer<\/strong>, which gives us the live position of the joint, in our case, that\u2019s what we call the currentAngle()<\/strong>. The code compares that to the targetAngle()<\/strong>, which is where we want<\/em> the arm to be.<\/p>\n<\/div>\n\n\n\n If the two angles don\u2019t match, the system automatically decides what to do:<\/p>\n\n\n\n The coolest part? This all happens automatically. If you type in a new targetAngle()<\/strong> for example, in the serial monitor, the system instantly figures out which way to move and does it \u2014 no manual toggling needed. Hello hello! This week was a bit all over the place, both in terms of tasks and available time for work. Continuing from last week’s efforts, I finished the inverse kinematic logic. However, implementing the last bit of code for the end effector proved more complex than I expected.<\/p>\n\n\n\n What remained was to fix the end effector segment of the arm, and make sure that its target angle always kept the segment horizontal. This was straightforward to implement, as we had some of the angles available already as bi-products of the inverse kinematics formulas. Specifically, the expression turned out to be;<\/p>\n\n\n\n Furthermore, I needed to implement the math for the coordinate system offset. Last week\u2019s blog post showed how the maths works, but the actual implementation required careful attention to detail (tunga rett i munnen). There isn\u2019t really much to say about this process, other than it being tedious, iterative and time-consuming. After finally completing the logic, the arm moved how I wanted, with the tip of the end effector placed at the given coordinates:<\/p>\n\n\n\n![\"\"](\"https:\/\/dronesonen.usn.no\/wp-content\/uploads\/2025\/10\/ChatGPT-Image-27.-okt.-2025-21_53_47-1-1024x683.png\")
<\/figure>\n\n\n\n\n
<\/li>\n\n\n\n
Once the two values are close enough, everything stops. That means that the valves go quiet, and the pump powers down.
<\/li>\n<\/ul>\n\n\n\n
And honestly, I think that\u2019s what made this week so valuable. I already had a general idea of how everything was connected between hardware and software, through the potentiometer, valved and pump, but this week it finally made sense.
Next week, my main focus will be on refining the setup code and making sure everything runs as expected.I know the wiring and connections are correct, so for now it\u2019s just about making the software align perfectly with the hardware. Once that\u2019s in place, we\u2019ll be ready for a new round of testing to see the full system in action again!<\/p>\n\n\n\n
\n\n\n\nSyver:<\/h2>\n\n\n\n
joints[2].targetAngle = 180 - phi_1 - phi_2 - phi_3; \/\/ satisfyingly enough<\/code><\/pre>\n\n\n\n