Fail your way to success

First flight is in the books, folks! Here’s how it went:

Or, to summarize:

Not too well! Which is kind of what I expected, to be honest. I figured something would fail. In software engineering, we have a sense that, if you run your code after making a complicated change, and it “just works”, then something must be wrong. Easy success is too good to be true. You learn to expect failure along the way.

So what happened here?

Well, it seems like one of the rotors isn’t turning as fast as the others. I’m not certain where the issue is—it could be in the ESC, but it’s more likely that it’s a problem with the motor. I may have to replace it, though if I do that, I’ll just replace all of them. (The kit I bought was sketch.)

Nevertheless: this is a big deal! The drone is complete. Have a gander.

Dunnn duh-duh-daaaaah!

There it is! My first drone, complete (sans propellers) and I made it from scratch a crappy kit.

Does it look good? Not really! And that’s OK. I crashed it immediately, and that’s OK! Definite progress.

My next step is to try to tweak the PID controller—that is, the self-correcting feedback loop which helps smooth out the changes in a moving system (PID stands for proportional–integral–derivative). In this case, it makes sure that, as the motors spin faster and faster, the vehicle doesn’t suddenly shoot off into the sky as soon as its thrust beats gravity. I think. I’d explain more, buuuut I haven’t worked with this concept since robotics in college! I hope to have more knowledge on the subject for next time.

Toodles!

Glossary

ESC

• Electronic speed control

All about drone guts

Note: This post was largely written a few weeks ago. It has good stuff in it, but it’s not current anymore. So don’t be surprised when(/if) I’ve made an incredible amount of progress between now and my next post!

I’ve continued to work on soldering. I now have the power distribution board (PDB) connected to the battery and one of the electronic speed controls (ESC). An ESC powers, and regulates the signals to, one of the motors that spin the rotors. So this seems like a good time to talk about the main components in the drone design.

Power distribution board (PDB)

The power distribution board accepts power from the battery, and provides it to all the other components in the drone. Each of the motors will be hooked up to this board through its ESC.

Electronic speed control (ESC)

The electronic speed control gets power from the PDB and instructions from the flight controller, and in turn powers and controls the motor it’s connected to.

Motor

Even if you don’t know how electric motors work (which, by the way, totally reasonable. I only have some idea. I think it’s magnets?) you probably know what they are. This motor receives power and instructions from its ESC and spins a propeller.

I’ll get into this more another time, but we want two of the motors to spin clockwise, and two to spin counterclockwise. Surprisingly, it doesn’t matter which motor wire you connect to which ESC wire. It’ll work however you wire it up. If it spins the wrong way, you just swap two of the wires, and it reverses direction. Magic!

Flight controller (FC)

The flight controller is the (pretty dumb) brains of the drone. It gets control input from the receiver, and figures out what the inputs mean in terms of how the motors should spin. It translates those inputs into instructions which it sends to the ESCs.

The FC also contains a couple of important sensors. It has a gyroscope so the drone can tell when it turns, and an accelerometer so it can tell when it moves. It also supports a number of other components, like GPS, and a camera, but I’m not using those right now.

Interestingly, the flight controller can be powered straight off one of the ESCs, instead of being hooked up directly to the PDB. This is one of many, many surprises I’ve encountered.

Instructions not included

You know how, when you got a Lego set as a kid, if you misplaced the instructions, you were basically lost? Or if you lose the instructions to your new IKEA cabinet, you will have no idea how to put it together? I have had a fun* time attempting to answer the question “so how do I actually connect this stuff?”. The FC, as you might imagine, is a very important component. (I mean, they’re all important; there really aren’t any optional components in this thing. I’m not even bothering with a camera yet. But it is the coordinator, so you need to know how to connect component to this flight controller.)

So I was miffed that there were absolutely no instructions in the kit on how to hook up the flight controller. Or, well, any of the electronics.

It took me days (of working on other things) to realize that the baggie the FC came in had ‘CC3D’ printed on it. Once I searched that online, I did find a reasonably good manual. The internet is a wondrous place, eh? Even with the manual, it took me a while to figure out how to plug in the next piece, the radio receiver (Rx).

Receiver (Rx)

The receiver is a chip which listens for instructions radioed to it, which it picks up with its antenna. (Aside: electronics are incredible. Remote communication? What? Yes, I know we’ve had radio communication for a century. That doesn’t make it mundane. We did not evolve to find this stuff intuitive. There’s so much for us to wonder at in the universe 😄.) The receiver passes these instructions on to the FC.

Transmitter (Tx)

The Rx communicates with the transmitter (Tx), which is usually, and in this case, a handhold remote control. It’s similar to a remote control car controller… but way more so.

The first prototype of the rocketdrone will definitely be using a handheld controller. Having the vehicle land itself is one of those problems I am deferring into the far future. But it occurs to me that if and when I do implement the autonomous landing functionality, I don’t have to calculate it onboard the vehicle. I could instead run that code on a separate computer which transmits instructions to the drone. This means the vehicle won’t need the extra computer onboard.

H(oo/ec)king up the receiver

(hooking/hecking) Yes, I’m rather pleased with myself 😁

The receiver is supposed to connect to the flight controller, passing along the instructions from the transmitter in the remote control, telling the vehicle how it’s supposed to be moving. This particular receiver has eight places to connect wires. So, which ones am I supposed to use?

Power and negative are easy, they’re 5V IN and GND (5 volts in, ground). I’m actually not sure which of the other terminals I’m supposed to use 😟. I’m not too worried about messing those up, though. Crossing those wires will only make things not work, it shouldn’t break anything.

Uh, whoops

I did manage to break the receiver chip, though 🙃. I’m not totally sure how? My best guesses are that I somehow mixed up the wires at some point, or that one of the shorts I definitely caused (though they were outside the chip itself, which I’ve been told should be safe? ¯\_(ツ)_/¯ ) was unsafe, and broke the chip. Or, most likely, when I was trying to power up the receiver chip to pair it with the transmitter, I connected the chip directly to the battery. Which proooobably overloaded it 🤦‍♀️.

At any rate, now when I plug the receiver to the flight controller in a way that I’m pretty sure is correct, part of the chip gets warm, which is promising, but neither of the LEDs light up, which is unpromising. I’m pretty sure the LEDs are supposed to light up when connected.

So I have more receiver chips on the way. Yay for breaking things!

My next post will probably be about my experience connecting and testing the components. Later!

Glossary

• PDB

  • Power distribution board

• ESC

  • Electronic speed control

• FC

  • Flight controller

• Rx

  • Radio receiver

• Tx

  • Radio transmitter

Soldering on

Electronics work often involves connecting different components together to form an electrical connection, and there are a number of ways to do that. We’re all familiar with a very simple connection: wall plugs. A device that needs electricity in order to run needs an active electrical current, and if it doesn’t have its own power supply (like a battery or a car engine), it needs to connect to an external source. When you plug your vacuum cleaner into the wall, you temporarily connect the electrical innards of your vacuum with an electrical circuit that runs through the walls of your house.

The vacuum is a pretty simple case, because the only thing it needs from the electrical connection is power. Data can also be transferred through these connections. The cable that runs from the back of a desktop PC to the monitor uses the electricity to send data that the monitor can understand as instructions to display a certain pattern on the screen. Some cables can only send power, some can only send data, and some can do both. USB is an example of this last category.

The connections I’ve mentioned so far are all designed to be easily reversible. You push the plug into the socket, and it stays there until you pull it out. Maybe you use built-in screws to make it more secure. But the idea is that you want to be able to move a plug from one socket to another without too much bother. There are other connections that are more permanent, and that’s what I’m learning about right now. A central component of these connections is called solder.

Solder (pronounced “sodder” in the US) is a metal alloy that you can use to form secure electrical connections between pieces that start out separate. Solder conducts electricity, so if you have two wires and solder connects them, current can flow from one wire, through the solder, and into the second wire. Solder also has a relatively low melting point, only a few hundred degrees Celcius. It can be melted with hand tools, and applied to your components. This forms a solid connection, essentially making your wires into one connected piece.

Behold my very first soldering connection!

It is not great. It is exposed, and therefore not really safe. But it will conduct electricity from one wire into the other. Probably.

Solder itself is not entirely safe. It often contains lead, which is poisonous to humans when ingested. I’ve been using a Tin/Lead solder, so I’ve had to be careful to wash my hands before touching food or my face holes. But it’s hard! I’m touching my phone as I work on electronics, so I’m probably getting lead on it, and then I use my phone throughout the day. Is this safe? I’m not sure! 😬

Soldering: Day 2: Hup to, Solder!

On this day, I learned that soldering is hard, sometimes painful, and potentially expensive. I also learned that making mistakes can be fun!

I’m building a quadcopter drone first, before I even get started on the rocketdrone. There’s a lot that I don’t know about what goes into a drone, so it’s really hard to make informed design decisions about the rocket drone at this point. Like I discussed last time, you can split a problem up into separate components, and figure out how to solve problems more locally. I don’t want to be figuring out how drones work for the first time when I’m building the rocketdrone; that way leads to overwhelmedness and frustration.

The first not-just-practice things I tried to solder were two wires and an XT60 connector. This will help me connect the battery to the rest of the drone, which will let me power on components as I add them. This means that I can test the components right away, so I can see if I’m connecting things correctly.

The XT60 is a standard electrical connector often used in drones and other remote control products. I’m using it to plug the battery into the rest of the device.

And soldering it was hard. I was trying to get the wire to fit into a copper receptacle that wasn’t much bigger than the wire itself.

In order to make a connection, you usually want to apply solder to the two surfaces you’re trying to join (this is called “tinning” them), heat them both up so the solder melts, and then join them. The solder cools very quickly, so in seconds, you have a solid connection. To melt the solder I’d applied to the plug, I was sticking the soldering iron—a heated, pointed metal stick—into the wire receptacle. The video I watched made it look pretty easy. Unfortunately, it seems that copper conducts heat very well, and once of the things that was heating up was the plastic of the plug.

So the plastic kinda melted a bit, and the metal pin kinda moved a bit.

And this little Yossi went wee wee wee all the way to the hobby store to get new connectors. Which really wasn’t that bad!

Taking chances, making mistakes, getting messy

I had a conversation with my therapist the morning of the day I ended up trying to solder for the first time. He said that, if you’re trying to invent something, which I guess I am, there are going to be failures. You are going to break things. You’re often going to learn something when you break something, but it will be broken and you will need another one.

I’ve realized recently that, for all that I identify as a scientist (someone who doesn’t have any absolute, unshakable beliefs, someone who is willing to change their mind when they’re wrong, someone who treats mistakes as opportunities to learn something new), I hate making mistakes. I feel like I’m going to be found out to be a fraud, or it feels like I’ve proven that I’m not good at things and I’ll always fail. Growing up, I think I often felt like I was disappointing someone. And that reinforced to the point that I’m reluctant to try new things, because I might not do them well. I might mess up. The perfect becomes the enemy of the good, or the good enough, or even the better than before, and I just don’t try.

And yeah, I totally did make mistakes soldering. I ruined some of my hardware, and had to buy more. I burned myself a little, on the heated pins, on the soldering iron, and on tiny balls of melted solder. And you know what? It felt amazing. Trying something new, screwing up, knowing that I’ve just hit a little bump but damn it I’m going to overcome it, and then actually overcoming it, felt so good. I have been missing so much by playing it safe. From now on, I’m going to try things, and mess up, and keep trying. I’m going to learn, and improve, and it’s gonna be great.

So the next thing I did was give up and go home 🙃.

Getting by with a little help from my friends

I’m not always going to know how to make progress. In this case, I didn’t seem to be getting anywhere with soldering the wires to the XT60 connector. So I arranged to meet up with a friend who’d offered to help me learn various shop tools.

Having that help was a big asset. Sure, I was stalled for a couple days until we would meet up, but I wasn’t stalled due to uncertainty. I knew what the next step was, and waiting and being patient was actually part of the process.

It turns out that I hadn’t been tinning (pre-applying solder to) the soldering iron. When the heated, liquid solder on the soldering iron contacts the heated, but still solid, solder on the wire, the solid almost magically becomes liquid. This makes things so much easier.

After my friend’s help, I made much better progress. Take a look at this chip!

This is the power distribution board (PDB) for the drone. It takes power from the battery and supplies it to all of the components in the drone: the flight controller, the motors, and the cameras if present. You can see that I’ve successfully attached the wires to the yellow XT60 connector by now. Further, I’ve soldered the connector wires to the PDB. Note how I’ve tinned the pads of the board. See how the solder is only on those copper pads, and not on the black material surrounding the pads? I don’t know what this black material is, but it seems to repel the solder, which means that you don’t have to be as careful when applying the solder. What you want to avoid is the solder from one pad and wire coming into contact with another wire, or “bridging”. This can cause an electrical short. The solder-repelling material is a big help.

My next step is to connect one of the motors to the PDB and figure out how to activate it. I’ll have more on that soon!

Composing a suite vehicle

Let’s get to it! What are we actually working on here?

The project goal is to launch a vehicle vertically using a model rocket engine, then descend and land under rotor power. The vehicle itself will be one, combined piece, but we can think of the different components separately, as things that can we can figure out independently, and then put together to form the working whole.

So what are the pieces? The project goal tells us what each component is supposed to do:

1. Ascend using a model rocket engine

2. Descend and land under rotor power

This is a bit of a cheat; we’ve already constrained what we can use to solve the problem. If we hadn’t already decided to merge, specifically, a model rocket and a drone, we might propose various solutions, compare them, and then pick the ones that seem to work best. As it is, there’s mostly one kind of thing that uses a model rocket engine—a model rocket—and one kind of thing that uses rotors to move vertically—a multicopter (or less generically, a quadcopter, aka drone). In fact, let’s start by removing the rotor requirement:

2. Descend and land under power

Hmm, these requirements are pretty vague. How will we know if we’ve succeeded? Sure, if our vehicle pops up two feet on a tiny explosion, then turns on rotors and falls without hurting itself, that might technically meet the requirements. But it’s not what we really want. Let’s set some more specific goals:

1. Ascend to 1000 feet in altitude in under 10 seconds

2. Land with a final vertical velocity of under 2 feet per second, within 5 feet of a target point

3. Don’t separate into pieces

This is much better. Now we’ll be able to check whether the choices we make fit well with our goals. Note that the numbers we’ve selected are pretty arbitrary; after all, this isn’t a government plane or rocket whose contract is written with the vehicle’s use in mind. So these numbers might change, and that’s ok! If they do, we can reevaluate our choices in light of the new requirements. It’s pretty easy to make mistakes when that happens, though; some earlier decisions might be carried forward into the updated design without checking that they’re still appropriate, and that can cause problems. So we’ll try not to change things too much.

Admittedly, we’ve chosen these numbers so we pretty much have to use a rocket engine to ascend, and rotors to descend. This is another cheat, so to make amends, let’s at least propose alternative solutions and see how they do.

Ascent

The first requirement is to reach 1000 feet in under 10 seconds.

Solutions for ascent:

• rotors

• rocket engine

• one of those stompy rocket things for five-year-olds

Based on what I can find online, most drones are going to max out at something below 70 mph, which would just get us to 1000 feet in 10 seconds, if we didn’t have to first accelerate. I’m also guessing that this max speed is for moving horizontally; going upward, against gravity, might be slower.

Can model rockets reach this height? Some definitely can, but many of them are very light. We’re going to have the added weight of the quadcopter pieces, which model rockets typically (obviously) don’t have. I think it should be possible, but I haven’t done the math yet! It could be that this simply isn’t possible without making a really big rocket. This is a part of the problem I’m going to set aside for now. If we find out that 1000 feet in 10 seconds isn’t doable, we’ll change the requirements, and we’ll be sad about it 😞 but we will persevere 💪.

Can we use a stomp rocket? No. No we can’t. Duh. But now I know what I’m getting my goddaughter for Chanukah!

Descent

The second requirement is to land softly, and near to a target.

Solutions for descent:

• rocket engine

• rotors

• parachute

Could we use another rocket engine, like SpaceX does? The short answer is no. The long answer is hell no. I think what SpaceX and Blue Origin are doing is really cool, and they’re part of the inspiration for this project. But model rocket engines can’t throttle or steer, and they don’t always ignite reliably or in a reliable amount of time. So it simply isn’t possible to use a model rocket engine. And I am not qualified to use something more complex. Let’s keep it safe!

Rotors are a good bet. Drones are capable of controlled vertical and horizontal movement, and of making minor changes in velocity. A good approach would be to make sure that the vehicle can hover, and ideally maintain altitude while moving horizontally.

How about a parachute? Well, that’s the solution we’re trying to improve on, so we already suspect that it can’t slow the rocket enough. Apogee Rockets suggests that if you want to descend on a parachute, you should shoot for falling at 3.5-4.5 meters per second, or about 11.5-14.5 feet per second. So the parachute won’t work. Additionally, a parachute wouldn’t allow the vehicle to aim for a particular landing site; instead, it would be at the mercy of the wind. Finally, a parachute would violate the third requirement, because in order to expose the parachute, a model rocket has to pop its nose off. This would mean separating into pieces. The pieces are connected by a cord, but they’re still distinct pieces in my book.

Conclusion

The two pieces, then, are

1. model rocket—our up-goer

2. quadcopter drone—our down-comer

Separating these components conceptually means that we don’t have to complete one before working on the other. There are many ways that decisions in one component will impact the design of the other, but we don’t have to know exactly what that decision will be in order to make good progress elsewhere.

I think this is a good place to leave things for now! I have more stuff written about aerodynamics, and identifying the front of the vehicle (not so simple!) but I’d like to keep these posts reasonably bite-sized, so they don’t overwhelm the reader. Not overwhelming people with detail upon detail of in-depth explanation is something I’m working on 🙂.

Until next time! Have a goodun!

Hi and welcome!

Welcome to Rocket Feels! I’ve never done a blog before, so this will be a fun learning experience for me.

Rocket

I’m starting this blog to tell you about my new personal project, which will merge two separate, independently cool devices: a model rocket, and a quadcopter. Why?

1. Cuz it’ll be cool

2. Cuz it’ll be fun

3. Parachutes don’t slow model rockets down enough

   1. Although honestly, I’m not certain this is a problem that actually needs solving. See 1 and 2 above for why that’s ok

When you launch a model rocket, it go up, and it come down. The going up is driven by a small solid rocket motor, which is ignited by an electric charge. It burns propellant for a few seconds of flight, then it coasts for a few seconds, then it ignites another piece of fuel that pops a parachute out the front of the rocket. The rocket then drifts down on the parachute, and lands at several feet per second.

A rocket which lands, in my opinion, too fast

The problem I’ll be solving is that the rocket is going too fast when it hits the ground. There’s a risk of something breaking. I would prefer that the rocket land more softly. Again, is this really a concern? ¯\_(ツ)_/¯ Doesn’t really matter! Cuz it’ll be cool, and it’ll be fun 😁.

A few months ago, I took the first step towards making this project real by returning to model rockets for the first time since summer camp as a kid.

As my 19-month old housemate would say, “kshhhhhh!”

It’s actually a good metaphor for what you can expect here: definite missteps, not always progressing, but we’ll get there in the end, and with good humor throughout. I hope you’ll join me!

Feels

The other thing you you can expect from this blog is my honesty, and an insight into who I am. I’m not very good at keeping myself to myself; when someone greets me with “How are you?”, I usually give them an actual answer. What I think saves me from the “Look, I’m just being honest” attitude, which can lead to real hurt, is that I always, always, try to be kind. To other people, at least. I’m working on being kinder to myself.

So I invite you along on my journey. It’ll be fun!

P.S.

The other other thing you can expect is puns and general silliness. I love puns. I rocket making puns, and I’m generally quite silly.

P.P.S.

I’ve gotten the subscription system set up, and it seems to work? I’d be honored and delighted if you would subscribe. If you check back once in a while, there might be new posts ready for you, but the bestest way for you to follow along is to subscribe 😁. But no presh! I get not wanting yet more emails in your life.