Sprint - Test Flights 6 and 7 

What’s that rocket that’s next to echo?

Flight #6: _"Looking good...not looking good."_ I think _every rocketeer_ has uttered _those words_ at least once....

Flipped Z-axis XD As a hobby game dev I can relate

Your determination all these years in creating and developing rocket models is truly extraordinary. Just wanna let you know you are a big inspiration for me. Keep it up!!

I had to check the merch section after you mentioned the shots with the attractive model, and damn, that's hot!

Can't wait to see future landing attempts.

BPS.Space: “AVA does exactly what it sounds like” Me: um... what does it sound like 🧐

Went to check the attractive model. Wasn't dissapointed

The AVA Logo is freaking amazing

Can't wait for that 1km flight !

Now that your Kalman filter works, I'd love to see a 3d visualization of the path the rocket takes!

I was just about to write: "Time for a kalman filter!" then you said you were writing a kalman filter XD I love this channel!

Hello! I’m Russian teenager, now I’m building my own rockets. My rocket flew on 2 km! I like yours videos. If you can make Russian subtitles, then I will be very grateful to you! Thank you, Ernest.

YESSSSSS JOEY IS BACK!!!! I’ve missed you!!

It’s really nice to hear that this problem was solved, wish you no more in-flight failures of sprint

from an engineering thought and education perspective, this is one of your best videos by far! suuuuuper interesting dude! :D

I really wanna see sprint flight 8 right now. Great project. Keep it up man. I love your stuff.

Good to see you again! man, it's been a while since you uploaded, I hope you are doing great.

Great job Joe! Definitely can appreciate the failure analysis and the "Forensic Engineering" you've done as I've done it to try and figure out why machines fail during manufacturing.... Also, every time I see "AVA" I always have to say it like "EVE" in WALL-E. Keep up the great work!

You are very close of the 200k goal, just realized. Congrats!!

Can't wait for the day you manage to launch and land one successfully, no matter how long it takes. Keep up the good work!! I'll be watching all of it!

I'm new here on the channel, I'm already enjoying it a lot, very good content!

I love it that the videos are getting nerdier by the hot minute!

I see a new model down there

Found this channel after your fail compilation vid. Very impressive stuff. Hurry up with your international shipping as I want a t-shirt.

A lot of that went straight over my head, but I got the gist of things and watched it through to the end. I did think that for a moment you were going to say that you needed to design and build AVA II, but pleased to hear that it was resolvable through software and look forward to the flight 8 video

you know your into space and rockets when you dive into someone's model rocket Failure Analysis and cant wait to read it....

Absolutely love these videos was looking into doing this for a while but I am thinking otherwise now

love your work, it inspires a lot!

I really like your content, in line with my hobbies in electronics and chemistry and technology, thank you, keep on creating content, I've subscribed to your channel since 2018 until now😁 Sorry typo 🙏

Frick yes your failure analysis documentation is lit brah

You're right man i love slow rocket launches .They look epic and make us feel that the rocket's powerful

I can totally understand the annoyance of the wrong negatives. I have been working on some code for a competition called student robotics and I spent the best part of 4 hours or so getting my tracking to try and work, nope that code saying when to add and when to subtract doesn't work, back to the drawing board. And that is where I can simulate it and see if it works the moment I have done it!

An absolutely awesome content, keep up the good work 🤙👌

I hope people have a grasp on just how crazy it is that Joe wrote a Kalman filter. I had to "teach" Kalman filters as part of the curriculum on Inertial Navigation Systems in the Navigation course I provided for the Navy. Being a newly trained individual in the Navy certainly has downsides...like you end up teaching things that you've never understood yourself. To describe it in layman terms, a Kalman filter is a way to combine many inputs with varying levels of confidence in such a way that the output is not just the highest confidence input, but a mathematically derived, weighted average total answer that considers all inputs. The "confidence" for each input is also mathematically generated based on equipment accuracy, noise, update rates, and drift. It's not exclusive to navigation theory, but is certainly standard in high-accuracy navigation equipment. INS in ships is actually very analogous to yours, they're full-up mechanical IMUs updated by GPS which uses Kalman filtering to give the most accurate position. Despite having "taught" it, there's no way I fully understand what is going on. The fact that you wrote one yourself is positively insane. I'm a little curious why you couldn't dead reckon between GPS updates using your INS data? It's definitely not as accurate, but would have saved you an awful lot of headache!

Wow can not wait for more videos yku are ever so inspiring! Excellent👍

Im having an online test for physics right now and what are the odds!😂...im skipping it for rocket science. Edit: I took the test and got more than 95%...and now im finally going to watch the video completely.

Very Nice update. Looking forward to flight #8. :)

Really excited that you're deep into control systems engineering now. Ye old PID controller just isn't always up to snuff for more complex stuff.

Its impressive to see how close it actually is to landing autonomously, respect :)!

Whoop whoop, svery day is a learning experience!!

Amazing work! I would like to suggest a very useful control method called "Sliding Mode Control" or SMC. A controller using this method can make your rockets follow a certain path or trajectory quickly. It basically minimizes the error immediately and converges it to zero in finite time. There are many research papers that use SMC on TVC which proved that it is much better than other controllers. I hope this will benefit your work :)

I've had my own fair share of reversed/faulty control inputs on my RC projects, so I know how that goes. But I never spent weeks writing/debugging the code that keeps them aloft. Keep fighting the good fight. This is fascinating to watch. Cheers :)

*Joe : looking good* *Also Joe : Not looking good* *Behind Great failures ,always have a Great Success*

Keep up the good work Joey B!!🌟

People : watching this to know why it failed. Me : enjoying how it fails.

I love your videos! You are awesome! Rocket man!

"Hey I'm Joe Barnard and I built those rockets" is the best.

Been waiting for your next video

seems like the rocket is kinda light as you can see the wobble and the gimbals over compensate. I'm wondering if you have a weight in it that can shift center weight up or down to keep it from wobbling. Love it B!

I also love the slow lift off

Man you jst inspires me so mch♥️. I myself is trying to built some space state model of the rocket.

Hay man, I can imagine a massive hard work to calculate the position in real time. But hay, you did it. Congrats man. Keep up! By the way, there is anything that you could kind of program to simulate most of the variables you have to deal with? Just curious if you ever heard of such thing.

"We are looking good. We are not looking good..."

I don't know much about it, but Fuzzy Logic can prevent aggressive maneuvers i belive. It was a kind of control technique I guess can be used in situations where inputs may be uncertain. And love your work man the content you produce is really inspiring ..

Congratulation in advance for 200 k.

You missed a sentence at the very end.... Something about masks, I think ;)

I would really appreciate a good video on Kalaman filters, if its half as good as your explanation of PID controllers it'll be better than anything I've come across! Great video!

I don't understand anything of his explanation.. But it's really enjoyable to watch your videos.. Keep the good work up man!!

if you don't wanna use a kalman filter, Ivensense has a great digital motion processor (DMP) library for the famous mpu6050 and mpu9250 chips which is really easy to use.

I check the Merch and the model is damn good looking

keep up the good work mr. Goddard would be proud!

It's really enjoyable to hear about the control system, lag and feedback and so forth. I don't mind that :) A question for you: Do you use a frequency domain model of your system? And if so, does the flight behaviour match theory?

Beepis Space

From 5:45 to 6:48 I was shouting 'Kalman filter!' at the screen 😂

Hey Joe, you should develope yourself a optical rocket tracking system. using the altitude via the telemetrie downlink, and a tilt servo, you should be able to track the rocket automaticly.

I’ve used kOS in ksp for 10 minutes so I know aaaalll about z axis flips I’m basically a brofessional on them. Ps it mean to type broffessional but I’m gonna roll with it.

Very big inspiration bro

I think grid fins can improve the stability during lift off and later you can improve it more. Just waiting for perfect landing on landing legs

I really like the music you're using for your videos, especially the outro. Is there a way to listen to it in full length?

Interesting video! I think you might be able to leap forward in terms of preventing these kinds of failures by working towards hardware-in-the-loop testing, or running (parts of) your control software in a software simulation. I'm not experienced with Simulink, but Python might be easy enough to run some basic simulations in. This kind of failures can then be caught in a simulation instead of real-life. I've also wondered if GPS-RTK might be benificial for you, since it can increase GPS accuracy a lot (at least relative to a known place, like the launchpad).

nice work

Finally your own ego-motion estimation! How does it compare against the VN300? A video about the comparison would be really interesting!

Yessss new viddd

I would detect for a flight deviation, and deploy the landing gear for drag. it will cut down on distance and might give you an extra second to deploy a chute (better than a core sample). I would also add fins to the upper rocket to handle misalignment quicker to return the rocket closer to the vertical launch GPS coordinance. but where the fins really help is the freefall. you can guide the rocket much like SpaceX, and in theory help, you stick the landing with less horizontal drift due to local weather conditions (especially rotor from the wind hitting trees and hills). to stick the landing. there is an issue of Scale. if you were to scale up grass, you can see what it's hard to "stick" the landing. to help with the issue of scale for the landing, well unless your launching and landing from a Golf Course. You need to "cheat just a little" be adding thin wires to your landing legs to widen your landing leg stance for added stability. you can start with long, and shorten them to the minimal length as needed. now one area I would look into is radar. a decade ago they were talking about miniature radar (used for cars, for the purpose of parking and driving was said to drastically lower the costs. I have not followed up on them, to see if it's advanced to even lighter cheaper units since then. the tiny laser measuring units they have now (gutted) could also potentially work to detect the actual height above the ground to fire the landing motor. But, the delay to actually light and get the motor up to thrust is a variable and the Motor itself would also be a variable. they could be minimized, and you would always want to use the same Lot Numbers as much as possible. your more than 90% of the way there already (that last 10% is a bitch) I'm sure you have already considered my ideas above, but I had to put them out there for others to potentially consider.

Great model for those product pictures! ;-)

Would love to see a liquid engine rocket on this channel some day 😊

Yoooooo that motor just FLEW out of the rocket

Before I check the merch, lemme guess, it’s you as the model. Edit 4 min later: yup it was Joe. I’m not surprised, but I highly approve

Sir your work is great, have you ever thought of pursuing MS in Aeronautics?

Hey Joe, I've always wondered how you don't light the grass on fire when you attempt to land or launch your rockets?

Hey bps space, for your flight controller, if you don't do it yet, you can design and order pcbs cheap on some of pcb websites

Great store - and now I really want to buy your dog

8:23 man he was not exaggerating

Look the Virgin Orbit model rocket is next to Echo!

Love channel

Why use GPS on the rocket? You could create a few stationary ground radio transmitters with GPS in different locations so the rocket can triangulate its location, distance, and possibly altitude. I mentioned that the grand transmitters would have GPS, that would be so the setup would be more efficient. Also because the ground transmitters would be stationary, you don't have to worry about GPS lag. This would be a WAY better solution in terms of efficiency for positioning in less time.

Just curious: how does the legality of model rockets change when you make them guided rather than just stabilised? 🙃

i like that the legs stand on the microphone (if you look upside down at least :D)

Joe: as you may know, I own some sophisticated GNSS/INS algorithms - if interested, you know how to contact me.

So what is the ultimate goal of PBS? To reach space? Or just to get working rockets that can go hight and come back safey?

Just a random thought but would it be possible to put ir sensors on the rocket and say 4 ir banks on the ground to guide the rocket on x & z "assuming y up" then use the GPS as a conformation of location device? At least until the delay problem is solved.

He protecc He attac but most importantly he is now bacc

When joe makes a video Me :.....….......…......yaaaaaaassssss

I know it's completely off topic, but that watch you're wearing looks amazing. If somebody could tell me what that is, that would be great.

Been waiting for the kalman filter to show up.

Nice microphone placement !

Interesting comments about determining tvc misalignment on launch, what if however, there is no misalignment, but the rocket is not burning cleanly down the rockets axis? Would that not also cause a rocket to deviate, despite being no TVC misalignment? How would you know which case it was, rocket burn or tvc? If you compensated for misalignment, but then the burn cleaned up, you would be now out of alignment (but the controller should still try and correct during flight, but with the unnecessary compensation multiplier). Hope my question makes sense..

Possible dumb question as I'm watching about halfway through... Have you thought about something like a rifle laser boresighting system to get things dialed in to within a few minutes of arc?

I'm actually surprised that you got that far without a filter. Did you do all estimations manually so far or did you use some generic estimation library?