Ok, I'm just going to cut to the chase here...  I bought into this gear because of what I thought it would/could do and as everyone should know by now I lost my quad on day-2.

My intentions for building this thing in the first place was that I wanted to do FPV flying while simultaneously recording HD video.  Clearly I needed to build my gear, learn to fly, and get comfortable with FPV first, but the desired end-state is the same.

The gear I ended up with was:

  • 3DR Quadcopter kit from DIY w/APM 1  (APM2 was still buggy)
  • 2x Turnigy Nanotech 2650mAh 35-70C batteries
  • Turnigy TGY9x tx/rx (borrowed from the late Kevin Finisterre)
  • GoPro HeroHD Camera

My plan was to later include hardware to extend my range (amps, tracking, ground-station, etc), telemetry, and probably a laptop/ground-station setup running Mission Planner.

After losing my gear to what we thought was a problem with my shoddy solder job I posted here, and after a bit of noob bashing it was my cheap radio equipment that was to blame, or so we thought.  Based on the comments we later determined that it was a "bug by design" in the APM that more than likely led to my quad pulling a Houdini.

Since then I have purchased a small cheap laptop to run the MP, another 3DR kit, and more batteries.  However, there has been a LOT more discussion, bashing, blaming, banning, and general covering up of the fact that the APM may actually be responsible for serious safety issues, fly-aways, etc.

The bitch of it all is that I'm sitting here wondering WHAT if ANY hardware could be used to make this thing (relatively) safe to fly, let alone accomplish my initial desired end-state.  Some suggested spending some loot on a nice Spektrum rx/tx, but we've clearly determined that even it's failsafes do not work in certain scenarios, which was the whole point of bringing this APM bug to light.  Having demonstrated a properly working failsafe on an expensive radio apparently being blocked by the APM PPM encoder behavior, could someone please explain this to me because I'm confused - this demo is in direct contrast to all the bashing I have received lately.


So despite all of this, low and behold patches are being developed and released to address this issue, but where does that leave me?

Based on my observations on this board for the last few weeks I don't see any way around this issue. It would appear that no matter what gear is used a fly-away could still be experienced.  The patches mentioned above wont do anyone any good if they don't have the knowledge or the hardware to flash the PPM encoder.  I've already lost 600$ in gear and I have purchased another 600$ in gear to help mitigate a problem that I apparently cannot solve.  If a fly-away is still possible I don't even want to attempt flying this gear.

So another observation that I have made is that a lot of people seem to be doing exactly what I would like to do.  How is it that their gear isn't flying away?  Are they just taking the risk or are they truly safe?  Do they even know about the issue?  Are they even using diydrones gear?

So time for the BIG question for all the "experts" in this community...  What gear would you recommend to ensure that this type of thing doesn't happen again?  

I would like to see more than just a preferred list of gear; I want to know WHY you think it's better than any other.  Price/budget is a factor so don't just toss out all the expensive shit because it's "better."

My plan of action included the FrSky modules for the Turnigy9x that I already have, a 2.4ghz amplifier, a directional/tracking antenna and ground station, a laptop with a ground station module for the use of MP if all else fails, long distance telemetry gear, et al.

I would also like to make the APM loiter (gps/altitude hold) if radio signal is lost with the option to auto-land exactly where it's at if control cannot be regained.  Hopefully this wont be an issue with MP active on-site.

So post em up, lemme know your thoughts on enabling me to fly without the fear of hardware failure, and the possibility of losing more money and gear. (related safety issues aside)

Views: 3116

Reply to This

Replies to This Discussion

I think the new PPM encoder firmware should fix the problem.  It loads with a USB boot loader, so I think it's probably not that difficult to flash.  I'd also assume with all the heated discussion that it will be included on new APMs very shortly.

But nothing is going to have the safety you're looking for.  Everything is still very experimental and there are lots of bugs that occasionally pop up even after dozens of perfect flights.  Just read some of the threads here!

This makes it very hard to call anything "safe".  I'd suggest that you better take lots of safety precautions AND be willing to accept the fact that you might loose your gear.  That's why I fly with dirt cheap gear as much as possible.  If you build your quad from some simple wood or aluminum from the hardware store you don't have to worry as much because the cost is low and you develop the skills to fix and rebuild anything that breaks.

As far as trying to make things as safe as possible I suggest that manual control is a must.  If the APM fails you have to have a backup.  I'd suggest starting with a quad that can actually be flown manually, then adding an APM.  That would mean having a RX->stabilizer system->APM setup.  The APM is a complex setup, so there's tons of things that can go wrong.  When people fly with a quad that can only fly if the APM is working perfectly they are asking for trouble IMHO.

I love watching youtube and reading guides on building stuff for cheap.  Look at PBFs (pizza box flyers) and oldskool quads.  These guys are literally building stuff out of scrap and risking very little.  Almost everyone I know crashed their first handful of flights.  That's totally to be expected.  You see people flying like pros and it looks easy, but they got there by crashing MANY times.  I just crashed myself a couple weeks ago from a stupid mistake.

Since you haven't run crying to mommy I think you've got what it takes to fly.  My best suggestion is to learn on a simulator and forget about performance/range, or having nice gear until you're good at flying.  Build a quad out of two sticks and the cheapest stabilizer board you can get.  Order spare parts ahead of time and keep at it until you're good enough to need, and willing to risk, a larger investment. 

I re-read your post... You need to do one thing at a time.  Start flying low and slow, maybe even tethered to the ground.  Don't even think of adding video, antenna tracking, TX boosters, etc. until you get everything working perfectly.

Slapping such a complicated system together and expecting everything to work is almost guaranteed to end in disaster.  Worse yet, you won't learn from your mistakes because you'll have no idea what went wrong.

Even when you do start adding stuff, keep going back to square one and testing everything like it was a new rig.  Lots of people here put on a helmet and fly their gear indoors for testing.  Chances are you can find a gym or auditorium to test in also if you spend a little time looking.  If you have a local college there's probably places to test fly in, especially during the summer.  Just go into an empty lecture hall and fly until you get kicked out.  Chances are that nobody will care anyways.  Some places also have enclosed tennis courts or other areas that would be great to fly in.   

As I said in the other thread:

- FrSky Tx Module

- FrSky Rx, either DR8SP or D4FR

- Use CPPM from either of these into the APM


- FrSky system has excellent noise rejection.  It's probably more robust than the Spektrum system.

- Telemetry to warn you of impending signal loss, battery status, etc. Before it becomes a problem.

- CPPM only requires a single wire between the APM and Rx.  It's cleaner, and less failure points.  And if the single wire fails, you will go into failsafe. Period.  No single-wire-missing problems.

- It's inexpensive.


You were extremely unlucky. I wouldn't count on being so unlucky again, but I would consider the possibility that *any* autonomous system can do a fly away.

As everyone else has said, keep it cheap, keep it simple. Start low, slow, if possible indoors. 

Test *everything* before takeoff, without propellers, including failsafe responses (turn off TX after arming)

Make sure you LOCK GPS before taking off. Otherwise RTL is RTSomewhereWhoKnowsWhere.

If you can wait a couple of weeks, the new geo-fencing software will be out. It won't be a magic bullet and it won't be foolproof, but it adds a layer of protection by initiating RTL for you even if you are not in radio contact. 

I would expect that the first five or six flights and first five or six hours of flight will give you a 1:1 ratio of flying to crashing. That is my experience. If those crashes are from 2-5 feet and low speeds, you can fix a part or two and fly again. So, assume you will crash and make sure you are at a speed and height that will allow you to recover a crash. 

Good luck and all the best. Despite the heated conversation, I think many people have a lot of empathy for your unfortunate beginning to the hobby and genuinely want to help. I certainly do. 

If you're worried about radio failure, try this for a dual transmitter-dual receiver buddy system:

I just got one for myself.  It allows you to switch between two receivers.  One is the master, which means one of its channels determines which set of inputs is used.  Basically a servo mux.  I plan to use mine to make a dual controller system with DJI Naza and APM1.  I'll use the mux to switch between the primary flight controller, and hot standby.  Haven't yet decided on which one I trust more.

Here's the link to the manual.

An example of how to use it.

Gravity always wins. ;)

I would advise against using a dual controller. 

As stated above, you must KEEP IT SIMPLE. Two autopilots make for twice the problems, not half, in my opinion. 

It just creates complexity, without reducing the failure issues. 

Agree.  This is just spiralling out of control.

Advice taken.  But, I'm still going to do it. ;-)

It's just so cool!

I admit, this probably is not something that Zen would do, given his lack of experience.  Also, I was recommending a dual radio in his case, not dual controller.  I'm the one using it for a dual controller.

I think you're going down the right track.  The first two steps I would implement:

  • FrSky Rx/Tx.  I got mine when Andreas had a flyaway with his Zephyr and landed in the San Francisco Bay.  It connects quickly as well.  Flipping the Tx power off and on gets a reconnection in a second or so, and invokes the failsafe very nicely.  BE SURE AND SET THE RX FAILSAFE according to the instructions.  We've got the 7dB antenna as well.
  • 3DRobotics radio.  Unlike the XBee (which I never did get working well) it has been rock solid.  It's technically interesting to be able to monitor your unit while it's in the air, and it's a great backup in that you can send commands (such as return to home) via this channel if your main radio fails.
  • bonus step which I have not tried: add the FrSky telemetry with battery monitoring.  This can also be done through the APM, but with some more work.

My logic for these two being the first steps:

  • simple -- not a lot of complicated interactions to understand/troubleshoot.
  • adds value in addition to the safety aspects.
  • cost effective.  The FrSky upgrade is about $40 and the 3DR radio is about $75 for the PAIR of ground/air units.

Step 2 assumes you've got a laptop available of course.

Good Luck!

Having demonstrated a properly working failsafe on an expensive radio apparently being blocked by the APM PPM encoder behavior, could someone please explain this to me because I'm confused - this demo is in direct contrast to all the bashing I have received lately.

Not entirely true, the demonstration shows the PPM encoder having a design weakness by keep the last known value when there is a physical signal wire loss.

The price of you radio has nothing to do with it simply because the throttle channel of the receiver is disconnected, preventing any failsafe value from reaching the APM board. When the throttle channel is reconnected the Spektrum failsafe clearly works as intended again.

The only radio receiver I know of that supposedly stops sending a signal on the throttle channel (same effect as physical wire loss) is the 9X. And even here there are conflicting reports depending on firmware and receiver revisions. All other radio systems that I know of, will either transmit last known or failsafe throttle value when the radio is turned off.

So to make it clear. The design flaw (I will admit that, but it was by design for technical reasons mentioned in the failsafe thread) is only in effect when there is a broken throttle wire, or with certain 9X radios/receiver combinations having uncommon failsafe throttle handling.

Anyways, with the new firmware the wire loss/9X throttle problem should be solved.

I have experienced this as well, had a loss of control that saw my quad fly away. I am on here almost every day and knew of the issue and have been following it closely, yet i still continued to fly with my stock 9x.

the wakeup was when i lost my quad, i was very lucky that when i walked (kinda running and panicing) to the way the quad dissapeared i got the signal back to the xbee's then after working out where it was from the GPS i drove till i got the video signal back too. found the quad undammaged on someones front lawn, luckily it didnt hit their BMW in the drive.

Who knows how far it would have gone if the battery had more charge left i guess i was lucky on more levels than one!

Since then i have modded my 9x and set the failsafes. and i guess another thing is having the xbee's and making sure the battery is secure so it doesnt get ripped off in a crash.

I have been considering getting a GSM GPS tracker, they are small and light, independant battery etc. 1 quick sms and it will send you a sms back with its gps position speed etc. they have about 48hr battery life (rechargable) and work almost everywhere.


Reply to Discussion


© 2018   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service