AERO 307 Blog

Monday was the big day. Over the weekend my other group members picked up the motors and escs and soldered on the necessary connectors while I went home to grab some batteries and a transmitter so we can actually test it. This was because the batteries we ordered came in monday night.

We then met up at Derrick’s house since everything was there. We attached the motors and mounted the escs and struggled a little bit with wire management. After that we attached the top plate and soldered up a battery harness to allow us to connect 2 lipo batteries in parallel. 

What happened after this was probably one of the most frustrating experiences in recent memory. So, to finish setting up the hexacopter you have to program the escs to have the escs know what the max and min points of the transmitter are. This lets the motors spin up at the same time and roughly match the RPM. So we followed the process to what we thought was correct in which a series of rapidly decending beeps would make it so we had to pull the stick down.

Wrong. We were so,  so wrong.

Derrick and I spent the next 2 ish hours troubleshooting wondering where we went wrong. After awhile we decided to just reread the instructions to the programming and realized that series of beeps we heard early actually made it so the escs recognized the connected batteries as NiCad and not LiPos. So then Derrick and I finished recalibrating all 6 escs individually and then the system armed and ran for the first time.

Once Justin returned, we walked over to the local park and proceeded to have our first successful test flight seen below.

I then picked up the larger LiPos from Dr. Doig that night and charged them in prep for more testing tuesday. We metup again at Derrick’s, soldered the correct connectors on the batteries and got the Telemetry system working. The Telemetry system allows us to remotely connect to the hexacopter and eventually input autonomous commands. After that we headed over to the soccer fields near Eng IV where Tynan met us as we began some more “rigourous” testing. We attached a 120g weight on a long shoelace to the bottom of the hexacopter to simulate the magnetometer and to test the stability. As you can see it actually flew pretty damn stable and the repeated crashing by the original owners was most likely due to pilot error. 

After that we calculated that with that much wind, you can get 8 to 10 minutes of flight time before replacing the batteries would be necessary. Then we went home and prepped the presentation that we absolutely killed the next day.

Now for some reflection!

a) If you were to pick one thing that you feel like you understand pretty well about aerodynamics now that you didn't at the start of 307, what is it and why?

I felt like I understand stall conditions way better than I did before. This was mainly due to some confusion on my end and the first 2 labs we did quickly cleared that up for me.

b) If you were to pick one thing about aerodynamics that still confuses you, what is it and why?

Nothing really confuses me really when it comes to aerodynamics itself its more of the testing methods itself and why you use such method to get a specific result.

c) What was your 307 highlight?

100% gettting the hexacopter flying and flying it myself. It was a genuine blast working with Derrick and Justin.

d) What was your 307 lowlight?

The entire flow viz football project. That was just torture and week after week of stuff going wrong. We pushed through though!

I wish I could say a lot of things happened with the hexacopter. But to be honest not a lot happened. We are pretty much held back by the fact that Cal Poly’s internal package processing center is run by some very special people and it took them the entire day to get us the Motors and ESCs needed to ultimately try and get this hexacopter flying.

All that really occured was on Wednesday was wait around for the motors and escs while starting to build the hexacopter platform. Derrick printed some custom motor mounts that allows up to more easily mount the motors to platform. I then spent an hour with some acetone and olive oil removing some old double sided tape from the smaller platform that the receiver and other small electronics go on. That was pretty interesting. At the same time we called the shipping department and tried to track down where the package was and was basically told they will deliver it later today.

The day ended with being pissy about the shipping situation and choosing the batteries necessary for the hexacopter. We settled for 2 3 cell 5200 mAH LiPo betteries that will be connected in parallel for a total of 10400 mAH of power to supply the power system for around 10 minutes.

onday was mainly spent in 322 going over possible replacements for the motors and ESCs. This was because I took them home over the weekend and tested every single one. Each motor had some sort of issue with it. One motor had loose magnets causing terrible performance around certain RPMs while another just flat out would not turn. 

The ESCs were a different story. They flat out would not arm. In case you didn’t know about a brushless ESC’s arming process, generally they have a long initial beep then an amount of short beep equal to the amount of cells plugged into the ESC and a final beep to signal that arming is complete. These ESCs would flat out not arm at all. I would only get a single beep and then nothing. After troubleshooting for awhile, I gave up on them. 

So then the group and I tried finding inexpensive replacements. Hobbyking is generally the place to go to for quality (ishhhh) rc parts for projects like these. Replacements were quickly found that had the same rating as the original parts and we decided to present that idea to Dr. Doig later and then finished for the day.  Later that day I decided to update Dr. Doig on the hexacopter situation and ask about ordering replacements. Out of nowhere he asks the original owner of the hexacopter to pay for the replacements and the man did it! It was a pretty cool surprise that the dude dropped 140 so quickly. So that was a good way to end the day.

We then met up again on Wednesday, since the project is basically now just waiting on the new parts, we decided to just go through the parts and find out what random hardware needs replacing/remade. Derek took home some parts to create a 3d printed motor mount that goes over the end of each arm (seen below).

Justin and I messed with the ardupilot board mainly to update it and see if it still works. It still worked completely fine, the only issue is figuring out how the GPS works but thats a future us problem. 

The week began with gathering data on two NACA 4412 airfoils using a force balance. The only difference between the 2 is that there one has a larger aspect ratio than the other. We tested the 2 airfoils at 2 independent speeds to match Reynolds numbers (roughly 200,000) which allowed us to more directly compare the airfoils.

High Aspect Ratio (Left) vs. Low Aspect Ratio (Right)

We then adjusted the angle of attack in 5 degree increments from 0 to 10 and then 10 to 18 in 2 degree increments. Overall testing went very well with Kyle. Data collection was pretty quick despite the slightly annoying process in how the excel files are written and saved which everyone in the group got to experience. 

Seen above is the test setup used. One issue we experinced was that the mounting system was kind of stripped out. As in once you got to a certain point in tightening the bolt, it would loosen up pretty much immediately. This made it pretty hard to level and sometimes caused the airfoil to rotate left or right and vibrate a little bit at high angeles of attack. The wobble and vibration can easily effect our results and will be addressed in the future report. Overall it was still a pretty interesting experiment to run purely due to getting more experience in the tunnel and I look forward to more work in it.

The week ended with Mini-Project group announcements and meetup. I was assigned to the hexacopter group. The hexacopter project is basically bringing an old hexacopter back to flight readiness. It was used and abused by a phd student that was hanging a magnetometer off the bottom which caused a pendulum effect when it moved and ultimately a bunch of crashes. A couple years ago it was revived by a couple of Cal Poly students and flown but then taken apart and put into a box. So thats how we got here. 

Nothing really exciting happened this week. Monday was presentation day. Will and I just went over our project and presented our findings. They weren’t completely conclusive and definitely was not as impressive as other groups that were in our lab period. To summarize, the boundary layer was smaller when the football spun and the laces created a sort of corkscrew aerodynamic effect. I am overall very pleased with how the project went and spinning that football in the windtunnel was very very cool despite the complete rollercoaster it was. I guess it was more of a constant car crash in motion. 

Thursday was a bit more tame than presenting, since my group wasn’t testing today we just met up and did a bunch of prep work. Damien created a pretty sweet matlab script that quickly got what calculations we wanted while Kendrick and I helped with the theory. Then Bob and I began messing with Xflr which is basically a more advanced xfoil. It can simulate flow, lift, and drag. It can even produce force and flow vectors across the airfoil. 

Monday was pretty uneventful. After going through our mounting plan and seeing what happened when we actually spun the football, it was decided that we needed something way more solid to mount the football. This was to try to reduce the amount of oscillation created by the rotated football. 

We did this by attaching the motor, esc, battery, and receiver to this large wood block that had mounting holes that matched up with the super solid linear drive. There was a minor-ish detriment with doing this though. This would mean that the block would be a blunt face that would affect the flow visualization after the flow went over the football. Despite this, it was decided that this was the best approach going forward due to time constraints.

Moving on to Wednesday and actual testing day. It was basically the last day we could get some actual data to present on the next Monday. Will and I quickly mounted the block onto the linear drive and tested the spinning football again. The amount of oscillation was very reduced, it was still there but it wasn’t so bad that it was shaking our entire mounting system. Due to the line laser being burnt out earlier in the day, we had to rely on making the test section super bright with 2 floodlights to adequately illuminate the test. 

Will then setup the go pro and we finally got ready to get some actual visualization. I turned on the windtunnel and then communicated with will to adjust the smoke to where it contacts (roughly)  the center of the football. It actually looked good! The flow was attaching to the football and you could clearly see what is going on. We then did the scary thing and began spinning the football. At a quick glance, the boundary layer got smaller and there was definitely some different behaviors going on.

We then moved the gopro to a different view and repeated the test. Overall I would say it went much better! Hopefully he presentation goes well next Monday.

We began the week pretty excited. Will and I had a pretty good idea on how to mount the football to the motor by using some pvc pipe and a LOT of epoxy to make sure the football cant fly off.  On Monday we started with going over the plan and assembling the setup shown below. 

We fould out that we needed 1 more nut to attach the football successfully. So Will and I took a field trip to Home Depot to try and get what we needed. That took over an hour due to the threaded rod threading perfectly fine on a 6mm nut in the testing block in Home Depot but the moment you test an actual 6mm nut it would not thread. So that began the saga of Will and I running in and out until Will by chance found that the exact size that escapes my mind currently. Basically it was not metric but standard and was very stupid that it took us an hour to figure that out. After that we worked on the report some for group 1 and then decided to group back up wednesday to finish the final prep before testing on thursday night.

Next was checking if our glue was hardened and ready to go (we used an entire bottle of gorilla glue epoxy attaching the football to pvc) and having a friend of ours create a motor mount so we can use their aerodynamic post that would be in the floor on testing day. It was 3d printed and seemed to do the job ok but we wouldnt know until thursday night.

After picking up pizza from Fatte’s for Tynan since that was the only way we could convince him to come in to the windtunnel that late, we began attaching the setup to the tunnel. Then we plugged in the battery and receiver and decided to test the setup before we turn on the tunnel. This was just the beginning of a long series of just stuff going wrong, one after another. The moment you give the brushless motor throttle, the ball would behave like an unstable pendulum and wobble all over the place. It looked like it was gonna shake itself out of the mount and even break the post that was in the floor. We decided that it wont be possible to run the test as it is so I just attached a propeller to the motor for fun to get SOMETHING from our time there. 

After spending over an hour being just complete dumb and freaking out over the prop freely spinning in the tunnel, we finally got to getting some flow vizualization that was pretty much well, not cool?

We began the week with wake testing. Since we were trying to optimize angle of attack for max CL and max CL/CD, we chose to test at 7 and 14 degrees for the angle of the attack of the airfoil. Due to this, we had to adjust the starting position for the wake and how many increments we took. So instead of the starting position being 300 mm it was at 280 for the first test and 270 for the 2nd. The tunnel then ran at 15 m/s for both tests respectively.

Seen above is the wake test setup from the opening point of view. You could see how adjusting the starting posittion when changing the angle of attack becomes necessary.

This went pretty smoothly for the 7 degrees test. After some minor correction initally due to not hitting move (the function that moves the wake) in labview the overall data matched what we were expecting.  More interesting things happened when we did the 14 degree tests. The graphical results were not matching up to what we were expecting possibly due to the high angle creating more unpredictable results. You can see on the graph below that there is honestly pretty big error bars with a distinct lacking amount of smooth curvature past 300mm. Due to this we decided to recollect data and do it as fast as we can before we had to go for the next group to do their wake testing. Despite this, nothing really came of it, the resultant graph still looked as jacked up as the first.

Monday we finally got to acquire data to find the optimal angle of attacks for Max L/D and CL. I was tasked with adjusting the angle of attack for the first run and then controlling the RPM for the second run. We collected data for 0 to 15 degrees in 1 degree increments so we were able to collect data pretty quickly and finished with just enough time to spare. 

Seen here is hte infinite wing test setup and Will being super excited to begin testing.

There were a couple of issues we experienced that were quickly solved. One of them was when the infinite wing was reinstalled the pitot static tube was not setup correctly, producing an incorrect velocity. This was correctly fixed by Tynan but then there was Port 7 being well, completely wrong. It was stuck at 0 the entire testing day. This was fixed in the matlab code by taking Port 6 and Port 8 and averaging them.

During the first test run I was tasked with adjusting the infinite wing. It was pretty interesting since it got slightly harder to adjust the higher the AoA became. This also could be a source of error due to there possibly being an inaccuracy in the method used to adjust the infinite wing but I feel it is necessary to trust the accuracy of the method. During the second test run I was tasked with communicating to the guy adjusting the AoA to our desired angles and monitoring the motor controller. During this i got to see what was going on in the labview. It was interesting watching how the various ports behaved as the angle was changed. It could be observed that the data across the ports seemed pretty correct outside of a couple of ports. After testing concluded we began fixing what was wrong with the matlab code that was put together and attempting to get plots that make any remote sense. To be completely honest this last hour and a half was pretty unsuccessful in making any real progress besides setting up the report format and starting some general stuff.

The meetup on wednesday was met with much more progress. Luis and Bennett were successful in fixing the panel code to produce pretty correct plots that we can use for later analysis. Will began working on the report and created matlab code for the wake testing while I was working with xfoil to get data to compare to and approximate angles of attack we use to test on next monday. Xfoil is still a bit weird to me. I get how it works yet it feels like im just pushing in random commands to hopefully get a result I want. Maybe more guidance would be necessary in the future to fully utilize the program itself.

The last thing that happened wednesday was Will and I decided to group up for the next lab project. We found a nice model of a wingsuit and thought that it would be pretty cool to see how flow behaves around the wingsuit model. 

Nothing much really happened besides refining our test plan and diving into the matlab required to calculate CL and CD. We had to use our old 306 panel code we created as our final project in Dr. Marshall’s class. This required changing the geometry of the panel code to a NACA 4412 airfoil and to make it work for only 20 panels. 

For this first meetup, my lab group and I spent our time figuring out what we are going to test for and the deliverables necessary before we test next monday. The two things we are going to test for are:

1) Find max Cl and corresponding AoA

2) Find max Cl/Cd and corresponding AoA

We then decided that we are going to get data at 15 m/s and 30 m/s which gives us a Reynolds numbers of 257,432 and 514,864 for 15 m/s and 30 m/s. After this we began splitting up into different roles in preparation for the test on monday. William and I began working on learning how xfoil works and going through the tutorial. Although the ui seems primitive due to its command line format the program actually seems pretty useful for approximating data. Shown below is an example plot created using xfoil which also has the different coefficients needed for our approximations.

The other two group members were focused on working on the matlab code necessary to produce the plots and calculations needed for testing day. This seems a bit more complicated due to having to modify the panel code created in 306 for a different airfoil geometry but this would be tackled in more detail on wednesday due to running out of time. The goals for next meetup are to finish the deliverables, matlab code, and accumulate data in xfoil to approximate what angles we should be trying to get close to.