My journey into the world of DIY drones

3689634510?profile=original

Hi! I'm new to DIYDrones, but I want to tell you about my journey from total drone-n00b to... well, still a drone n00b, but at least a little bit more educated, with a proper respect for drones. It involves a tricopter you might have heard of: The Pocket Drone.


Back in the beginning of january of 2014, Airdroids launched a kickstarter compaign to fund "The Pocket Drone". On the kickstarter page you can still read their original promises.
For people like me, people who have never flown a drone before, this seemed like a great opportunity to get a compact, easy to fly drone, for a reasonable price. It was still a lot of money for a luxury product, so I actually hestitated a while... and the funding period ended. It was a great succes.
After seeing this, I found that they also took pre-orders for pretty much the same price, so I pre-ordered right after the Kickstarter project closed and the wait began.
My wife and I had planned a year-long trip to Australia starting that summer and this would be a great drone to take with us: compact, yet able to take great photos and videos through the use of a GoPro. The promised delivery date was in june and we were leaving in early september, so that all worked out great.
I was really excited when summer was nearing! I bought a GoPro and was ready to learn how to fly and all that. But then the delays came.
I'll save you the details, but in the end I had to leave for Australia without the drone. When it finally came, it was january 2015, and I was actually one of the very first to get it (some people still haven't received theirs). Why? Because I couldn't wait and I bought someone's "early bird" kickstarter spot and was in contact with Timothy Reuter, one of the three founders of Airdroids, to get both my pre-order and the early bird delivered at the same time.
So, halfway through january 2015, both my drones arrived. In Australia. We had already done most of our traveling, so we couldn't use the Pocket Drone to film it all anymore, but I was excited nonetheless. My first drone!
Boy, was I in for a let down...

The included manual was very meager. But, I didn't know any better. As I was one of the first, there were no reviews or videos of the Pocket Drone yet, so I set out to be the very first to film it flying!
By now, someone had posted a link to the RCGroups forum so I got a little bit of support there, and with that information, I went out to the park to get her up in the air.
This is the report that I wrote later that day:

******************
My Pocket Drone experience so far.
Yesterday I received my Pocket Drones. I originally ordered my Pocket Drone after the Kickstarter had already finished, but later on I bought Me's drone. As he was an early backer, it shipped last weekend.
Edit: Clarification, as there are some questions about it: I bought the early backer spot of "Me" on Kickstarter. You might have seen his comments about him being willing to sell his drone: I bought it from him. He changed his address to mine. This is why I got my drone so early without being a backer.

Yesterday I unpacked both of them (as you can see here: http://imgur.com/a/fzmfb), one was an RTF and the other was BNF, and started toying around with the ready to fly one.
I unfolded everything, put in the battery, and tried to get it to work.
The transmitter was properly bound but the drone seemed to respond very sluggishly to the transmitter. Half of the time arming and disarming didn't work. There was no manual for the transmitter (Flysky-T6) included, but I found it online: http://site.hobbypartz.com/manual/FlySkyT6Manual.pdf
As this is my first drone and I'm a total noob, I checked with the people on the RCgroups forum (http://www.rcgroups.com/forums/showthread.php?t=2072873 page 15 and on, I'm the user "DeltaBlast"). It turns out that this sluggishness is normal because you start in the mode "altitude hold" and it tries to work on compass and gps and you don't have free control over the drone.
On the APM site I read that the first mode should always be "Stabilize" (which means you have full control over the drone), but apparently the PD people thought that was unnecessary.
To use stabilize or any other mode, you first need to connect the Pocket Drone to mission planner. For this, a USB telemetry stick is included. It didn't work for me at all, even after installing the proper drivers, but in the end I got it to work by reinstalling the drivers through device manager. So if you can't get mission planner to connect, check device manager for errors.
So now that I was connected to Mission Planner, I could see the two flight modes by going to Initial Set up>mandatory hardware > flight modes.
Now if you want to use more modes, you need to bind them to a switch or knob. The problem is: all the flight modes are handled on one channel and you can only assign one channel to one switch or knob. But a switch only has two positions. You could cycle through all flight modes on a knob, but it would be very difficult to see on which flight mode you are, so that's not an option either.
Luckily another forum told me how to use another channel to modify the first. This is called Mix on your transmitter.
I wrote this little tutorial to tell you how to set this:
***
By default, Switch A (SWA) is assigned to altitude-hold (0) and loiter (1). You can see that this works in Mission Planner via Initial Set up>mandatory hardware > flight modes: if you're connected and you flick the switch, you see the mode changing.
You can set another switch to modify the behaviour of Switch A, by going to Functions setup (hold the scrollwheel on the transmitter), then to Aux channels. Set Channel 6 to the switch you want to use to modify (I used SWB). Then go back to functions and scroll to the second page to Mix.
Turn on Mix 1, set master to Ch6 and Slave to Ch5. This means that 6 now modifies the output of 5.
Set positive mix to 50% and negative to 0%
Now turn on Mix 2, same master/slave and set pos to 0% and negative to 75%.
I got these values from another forum, so dunno if these are the best, but they work for now :P
***
Right, that was yesterday (and part of this morning), today it was time to try to fly for real!
First, I stuck on the GoPro mount. Note: there is no GoPro mount included! While there was a GoPro mount visible on all promotional material (it was part of the 3Dprinted frame), apparently they decided to go with a mounting plate instead, so you have to stick on your own GoPro mount. Also, once this is placed, it doesn't fit neatly into it's protective box anymore :'(
Anyway, I went on my bike to the park and got started.
First I taped the struts to the Pocket Drone with ducttape. Later that turned out to be a very wise decision!
I did a lift-off on altitude-hold, and immediately the drone drifted away so I switched to loiter (for no effect). The drone still drifted off so I switched to Return To Launch (which is not bound by default, I did that manually). RTL mode lowered my drone near me and then smashed it into the ground twice before crashing. Nice.
The landing gear cannot be called landing gear. It can barely be called "launching gear" as it does remind one of the supports of the spaceshuttle, immediately falling off after launch :)
After take-off, the second part of the landing gear was just dangling immediately, even though I secured it properly. This means that landing with this gear will *always* have the tail rotor hitting the grass. On landing all struts shot loose and one even escaped the ducttape.
So: DO NOT TRY TO FLY THIS DRONE WITH THE CURRENT LANDING GEAR.
Anyway, I captured all of this on video, unfortunately I forgot to turn on the GoPro.
This first video can be viewed here:
Anyway, I'm persistent, so I tried again.
This time I tried to takeoff using droid planner from my phone, but that didn't work at all. No idea why.
So I tried another lift-off, this time with the GoPro turned on. The PD went up, drifted, then went down (I didn't even touch the throttle) and started chopping grass.
This is all on video too, it can be viewed here:

Third time, I tried to go with stabilize mode I think, but it just crashed sideways immediately.
Short video here:

I tried some more but it was pretty much all crap. I captured one try on the GoPro:

Then I summarized this in a video here:
As a last try I wanted to put in waypoints and fly from those. I put in the waypoints on my phone through droidplanner and pressed the button, but nothing happened. Whatever I did, nothing happened. It just stood there, idling (armed).
Turns out, the transmitter needs to be on as well (I guess that makes sense :P).
So, I gave the command and the Pocket Drone launched, flew up 10 meters, flipped out and just spun to the ground, killing it completely. Unfortunately that wasn't on camera :'(
Result: 4 clipped props, a severed antenna (or whatever that red thing was, hanging from the frame) and the frame has a big tear in it so one of the front arms can now bend further than it should. I think it still works, but I've had enough for now. The damage can be seen in this video:
and also on these images: http://imgur.com/a/oIvVo
Just to make it clear: at no point during my flights did I even touch the right stick, only left to launch..
I hope someone with more experience can get this to work properly and tell us what to do. Also, if someone can look over my videos and tell me what I did wrong, it would be much appreciated. I'll be over at the RCgroups forum :)

To summarize:
* Do not fly with the current landing gear. Your props *will* hit the ground on landing.
* Do not fly in any sort of wind, ever. There wasn't much wind, but wind was probably the reason for everything.
* A GoPro mount is not included.
* If you don't know what you're doing, like me, wait till others try this first and see what they come up with.
* The transmitter seems unresponsive in arming and disarming, and everything else really. But I have nothing to compare it to, so maybe it's just me.
* The manual for the transmitter can be found here: http://site.hobbypartz.com/manual/FlySkyT6Manual.pdf
* Look for more info and updates in this thread: http://www.rcgroups.com/forums/showthread.php?t=2072873 page 15 and on.
* My combined videos are found in this playlist: https://www.youtube.com/playlist?list=PLirj0w75mWNw74yneTgAHd_REvbd4qQ2F
*****************************

As you can see, it went badly. Very badly.
After posting the above, I was contacted by someone that lived close to where I stayed in Australia: Quadzimodo

This guy turned out to be a godsend.
He asked me to come over so we could try and get the Pocket Drone to work. I went there and I learned a *lot* about drones. It turns out that everything wasn't as easy as I (and other people with me) had imagined, and that drones could be dangerous in the way I was handling things.
We tried to get the Pocket Drone flying, but in the end we didn't really succeed. We vowed to try again though!
In the mean time, from the RCGroups forum, Bo Lorentzen launched a Facebook group for easier communication after receiving his own Pocket Drone. That Facebook group quickly filled with people who all asked the same questions and the answers got buried under new questions. Everybody tried to reinvent the wheel, and because of that, no real progress was being made. To remedy this (and also for my own overview), I decided to combine all questions and answers into the Pocket Drone Community User Guide. While this allowed for some progress, we still didn't see much actual proper flights. People were having problems with the yaw and blaming it on the APM, so most flights were more like short hops.


Two weeks later I went to quadzimodo's place again. This time, while facing the same yaw-problems again, he realized that the problem was in the tail clutch: Airdroids had poorly designed the entire tail section and for pretty much everyone, the clutch was slipping in mid-air. He devised a fix and afterwards, flying was suddenly possible, with relative ease!
Another writeup was placed, with new videos:

*********
Great news everyone!
We seem to have fixed all stabilisation issues!
Today, Quadzimodo and I had another go at trying to make the PD fly properly. After having constant trouble with the yaw, Quadzimodo recognized that the issue was that the midpoint of the yaw seems to move around. Why?
The problem seems to be the rubberized slipper-clutch which forms the mechanical link between the yaw-mechanism and the servo.
The clutch is designed to limit the torque that the tail rotor can impart on the yaw servo so to protect the yaw-servo from damage during a crash or hard landing. Unfortunately, this clutch is too loose and allows the tail-rotor-assembly to slip under normal use. This manisfests itself generally by poor yaw-control which degrades over time whilst in the air, generally leading only to very short flights (or worse) before having to reset the tail yaw position.
So, to remedy this, we opened the tail servo-housing and removed the plastic part of the clutch (actually, we first tried tightening it, but it was no good: the screw thread would just weaken and overscrew), cut off the edge with the screws so it can actually close around the rubber properly (shown in the photos below), put it back around the rubber and put zip-ties over it and tightened them up real good.
After this, no more spinning during flight! On stock firmware, no changes, we had perfect stable flights! This looks like the drone we we're promised! Personally, I had my first *real* drone flight today, and it was awesome. The guys that made the flight controller really did a great job.

Here you can see the clutch of the original tail servo (the top cover is removed from the tail):
vOMI7I1.jpgThese are the parts we clipped off:
aXnLkyo.jpghttp:

And this is the result after putting on tie-wraps:
57kcKjx.png
2C1CrUd.png

And here are the videos:






************

Sadly, right after this, I had to return home to the Netherlands. So this is where we are now. Quite the experience!
I've left out a lot of details, but in the course of this all I learned a lot about the world of DIY Drones: I now know how the software works, how all parts work together, and I could build myself a new drone now if I wanted.
I wrote all of the information related to the Pocket Drone down into the Pocket Drone Community User Guide and I continue to update it (I welcome feedback!).
So while nothing went as planned and the Pocket Drone project was (is) a disaster, I learned a lot more than I would have when I would just have bought a standard DJI drone. I respect drones now and I hope to learn even more in the future.

The next step is to transplant the usable Pocket Drone parts into a trifecta frame.

My journey into this new hobby is only just beginning!

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones

Comments

  • RCTimer 2208-17 was the motor used in the 3d-printed prototype I had a chance to mess with before they switched to the injection molded stuff; a short-shaft version of that motor is likely the one used in final production. 

  • Chance, I know there were some design decisions made based on a price point, or tech available at the time of design.  I would like to know what compromises in design were made, and what other choices you would have made if price wasn't a big limitation.

  • Heya Chance, very brave of you to show your face here indeed, but it's much appreciated :)We would welcome you in the Pocket Drone User Group.

    There was a working prototype before you left, right? Can you tell us what was in it? It seemed to fly much better, so I assume that the motors and battery and such where probably a bit more expensive or at least a better combination of parts. I don't know if you are allowed to divulge this information, but it could be great for people looking to improve their PD without doing a complete transplant.

    The promises that were made on flight time and weight and such... even though being very optimistic, I assume they were made based off the prototype, so it seems to me it would've flown pretty well (also seeing the FPV footage made with it a long time ago).

  • Quadzimodo - it is all good.

    Chance - you are fearless! Figuring out how to get this thing working would be a good service. Is this just a better servo clutch and a landing gear away from being usable? Do those folding props add a lot of vibration? As you may have gathered from above I was a tad cynical when I saw it touted but I definitely would have purchased one when it had been released and was pronounced capable of doing what was promised. Why do you think the flight times do not measure up? Were cheaper motors used in the production batch? I have a small fleet of  custom-made AP machines that get used for semi-pro AP. I am still looking for that elusive ultra portable/foldable APM machine that could do decent aerials and be in my car at all times.

  • Marc - You are absolutely right about what you can and can't comment on.  You have also perhaps been too generous to me with your concessions.

    I really enjoy participating in discussions on this site due to the fact that mugs are really quite thin on the ground (when compared to general alternatives). Your comments here just reinforce this in my view!

    Chance - I must admit that I am a bit blown away by your post.

    While your comments and invitation for discussion on the delivered product is most welcome (as it would also be elsewhere), I fear that most people (backers/customers like DeltaBlast and spectators like me) will be far more interested in mining you for information on just about anything and everything else.  Perhaps the best way to head off these potential questions might be for you to construct a blog post (which you could post here on in thePocket Drone User Group) covering your experience and understanding of the situation to the extent that is possible so that you can get in early and clearly define the boundaries set by your confidential separation agreement.  That way you can provide something for people to read and head off any potentially contentious questions that might be put to you in the future as a result of (admirably) popping your head up.

    A replacement set of landing gear would be extremely helpful to the many hundreds of backers out there dealing with this incredibly inconvenient (and somewhat laughable) problem.  Sticking them up on thingiverse is good, but handing the design to the remaining team members and compelling them to forward a set to each backers would be far more helpful.

    What do you wanna know about the clutch modification that isn't covered in the post above?

  • FWIW, I have not been a part of the AirDroids team since April of 2014 (less than a month after receiving the KS funds and before spending a dime of it). I cannot discuss much as per the confidential separation agreement I signed, but I have been completely out of the loop since then. I have no access to any customer service channels or emails. Needless to say, I would have tried/liked to do things differently.

    I AM a recipient of one of the units and it was already bound to the included transmitter (As RTF as I could have expected). I did change the default mode to stabilize (I don't think alt hold should be the default setting) and it left the ground a few feet without any problems (I haven't had a chance to take it outside yet). The redesigned landing gear leave something to be desired for sure (I am considering designing some 3D printed replacements to post on Thingiverse). I'm interested in the tail boom modification that was discussed here to see if it's necessary.

    Provided I don't receive an onslaught of negativity here from backers or DIYdroners, I will continue to have an open discussion and review the product as an informed consumer.

  • Well I stand corrected as you obviously know a lot more than I about this and the backgrounds of the people involved. Technically making prototype is easier but producing in on scale is often much harder. But yes, it is financially more feasible. The original specs still exceeds almost anything out there currently (size, weight, foldability) so this was never going to be easy.

    I do not think you are a sycophant - a poor choice of words. A pillar of the community more like it. I do however feel at liberty to comment of what anyone should or should not have said as long as I remain civil and I still think the original posting should have had a cautionary word precisely because of his position and influence in the consumer UAV community - it is not the only instance I have raised an eyebrow as I referenced some posts back. Maybe this was not quite so implausible as I once thought. Your point on that is well taken. Peace out.

  • Marc - Mate, I really hate to argue... but... The AirDroids team were not people with zero track records.  TJ Johnson is a qualified engineer (civil, minoring in mechanical), Timothy Reuter created Drone University (an online educational platform) and the Drone User Group Network (one of the world's largest networks of civilian drone users) and Chance Roth works for Brain Corporation (develops software, hardware, and cloud services for consumer robotics).  AirDroids actually posted a video showing a 20-minute flight with the prototype back in March last year.  Videos were also posted about the same time by AirDroids and a beta tester showing smooth aerials (as good as you would expect without a gimbal). When your clientele is prepared to front the cash now and receive their product later I can assure you that delivering 2000 units at $500 each with a reasonable margin is an order of magnitude easier than delivering just one.

    If I was a sycophant I would not have mentioned Chris' name in the first instance and would not now be engaged in this discussion with you. Perhaps you should read my response again so that you might realise that what I was commenting on wasn't so much what Chris might have foreseen but rather what he might (or might not) have said and why.

    If I were you I would not be claiming Chris should or shouldn't have done or said anything... and if you were the type of bloke that commanded a USD$100k+ speaking fee and was responsible for a company the size of 3DR then I suspect you would be a bit more considered about what you did and said than you (and I) are being here... hence why you are unlikely to be 'put in your place on this' by him here.

  • Quadzimodo - It definitely could have been closer to the mark than what launched. Just not very likely by people with zero track record. Which is my point. 20 minutes with gopro, foldable, weighing a pound minus camera, and takes smooth aerials. And has a controller included. Even now that is pushing it. I do not really agree with you at all. I believe YOU could make one that achieves this. 2000 of them? At the prices offered? If anyone would know how hard it really is to deliver on these specs and make a profit then it would be the person who heads the #1 US maker of such devices who, like all makers of multirotors, has had issues with hardware that does not quite hit the advertised specs. Maybe we are all wiser now but that seemed ambitious to me then and certainly worthy of a word of caution in the original post rather than unreserved, uncritical enthusiasm for something aimed at first-timers by people with no track record using marketing that was on first glance just 100% dishonest in the kind of aerial footage it could deliver.

    These suggestions that Chris Anderson could never have foreseen even the possibility than any of this could possibly happen seems mildly sycophantic. Maybe he can put me in my place on this. His enthusiasm got the better of him IMHO. But as I said, I am indebted to him for creating a community that sucks up all my spare time, so I say it with love. That criticism seem fair, even of nobody wants to air it here.

  • Marc - It was all perfectly achievable. AirDroids simply failed to deliver a well resolved product. Which is a real shame because if the Pocket Drone turned out to be even remotely reasonable they could have delivered 3000-4000 units by now, have developed a roaring trade moving spares and accessories, be working on their next product offering and would have positioned themselves well for further success in this incredibly exciting new market... In fact, their initial campaign plus direct orders taken within the weeks following completion numbered nearly 2000 units, providing them with upwards of USD$1.3M in working capital. By mid last year they were forecasting 2015 to be worth $8-10million (which, building on their initial success, probably would have been quite achievable). Instead... well... today AirDroids are no longer promoting their product, have shut their online store, pulled their personal profiles from Kickstarter and appear to have all but disappeared from the scene.

    Such a waste of a good opportunity that so many others here would not have squandered.

    Michal - No worries at all mate!

This reply was deleted.