HT-FPV Frame – Second Flight!

Got to take the HT-FPV out to the park today, finally there wasn’t any rain to put a stop to proceedings.  This is the first full-range test I got to do.

First, range checks.  The TM1000 seems to be very short range (20 meters or so!), which may be on account of where I’ve put the antenna.  I’ll have to work on that.  On the other hand, range testing with the AR8000 showed control still working at 300 meters (30 meters on reduced power).  I’ll do more testing on that later, but it’s within spec for that I need at the moment.  I’ve currently got the failsafe configured to just cut power – I’ll have to mess with trying autoland at some stage.

Second, hover checks.  Hover was fine, altitude hold held altitude great, even with quite large attitude shifts when maneuvering around.  There seems to be some yaw drift, which may be either magnetic interference or insufficient P (the drift isn’t always the same direction).

Lastly, loiter testing.  To start with, I left the quad in altitude hold at about 1 meter, then kicked in loiter.  The quad immediately started heading off to the west.  Since I have limited room, I didn’t let it get too far before kicking back out of loiter and bringing it back.  Tried it again, and this time it went great – it loitered around for a good minute or two within about a 5 meter radius area.

My theory here is that I probably didn’t have a good GPS fix (I waited for the 3dfix light).  If I had more room and I just let it go, it probably would have loitered at some other point nearby, but with my limited space I didn’t want to let it go.

Still some vibrations which indicate that tuning is still required, but other than that, a good result.

HT-FPV Frame – First flight successful!

Had the first flight of my HT-FPV based quad on Saturday.  No tuning, just went with factory default PIDs.  I installed the sonar under the frame – potentially too close to the power distribution, but it appears to be OK.

First flight was fine!  Takes off evenly, stabilization is good, altitude hold works great.  It wobbles a bit, so obviously the PIDs need adjusting somewhat.  Voltage telemetry on the Dx8 works well, and the LiPo alarm goes off at 10v.

The current monitor on the Attopilot doesn’t seem to register in the Planner, but that appears to be a noted bug in ArduCopter 2.6 .  The hardware definitely works since I note a voltage on the current pin that varies as the current draw varies.  Hope it gets fixed soon.

The battery didn’t seem to last very long, although I’m wondering how much of that is me losing track of time.

Next up is to get it all tuned up and to test loiter mode and auto-land for failsafe.  That’ll probably have to wait for the weekend, it’s dark by the time I get home from work.

HT-FPV Frame Assembly – Part 2

Got the last (yeah, right) order of parts in for the new quadcopter!  As such I’ve finished assembly, up to the point where I’m ready to attach props and start the testing process (in these photos the props are just sitting on the adapters and not actually bolted on).

Assembled HT-FPV Frame (3/4 view)

The frame is designed to hold a GoPro camera up the front, hence why the battery is sitting so far off-center.  I have quite a large battery on it (5000mAh), and since I don’t have a GoPro, I instead attached some ballast.  That ballast comes in the form of a roll of old solder that happens to be the right weight (200 grams).  Ideally, when I get a GoPro I shouldn’t need to adjust the center of gravity at all.

The TM1000 telemetry module is visible just behind the solder roll, and its antenna runs alongside the landing gear on the left side of the quad.  The AR8000 receiver is attached on top of the HT-FPV top-plate.  I currently only have six of the channels connected though.  You can see the power select jumpers on the APM2 Adapter board in there too.  The satellite receiver is attached to the front right landing gear leg with some double-sided tape and a zip-tie.

Assembled Frame – Top View

The telemetry radio is visible in this shot on back left leg.  It’s position is such that it’s far away from the receiver and not in a position where it might obstruct a propeller.  I know that having landing gear on a HT-FPV offends people’s sensibilities, but I tend to be a rough lander and I fly on often wet grass.  So I really want landing gear.  Due to the balancing of the center of gravity, I should probably push the skids back a bit further.

There’s a bit of free space under the battery, that is likely where I’ll install any FPV transmitter equipment I get.

Next up, pre-flight checks and test flight.  Hopefully the weather will hold out for the weekend.

APM2 Board Adapter

I received my APM2 Board Adapter boards last week, and have now attached them to my HT-FPV frame.

APM2 Board Adapter
Attached to HT-FPV middle plate

The board came back from Osh Park exactly how I wanted it.  Fits perfectly, although I noticed that ESC inputs 1-2 obstruct access to the USB port on the APM2.  That’s no problem, because for a quad you can just use inputs 5-8 (as I’ve done in the shots.  Attachment was done with various nylon washers and nylon M3 screws.

I’m quite pleased with how it’s turned out, looks good and it does the job nicely.  I’ve got two spares, don’t know what I’ll do with them though…

HT-FPV Frame Assembly Part 1 – Photos!

I forgot to attach photos to my last blog about the HT-FPV frame!

Assembled frame, topplates removed

Above is the assembled frame, with the topplates removed.  In this shot, the power distribution board hasn’t been added, but the assembled arms are on the frame.  The blue heatshrink-wrapped things are the Hobbyking F-30 ESCs.  The motors are the Turnigy 2217-16’s.  For connectors I’ve used Hobbyking’s shrouded 3.5mm banana plugs.  They make a good tight fit.  Notice how the servo lines from the ESCs are supplied with a ferrite choke.

Motors with cut-off shafts

Above is the motors, with the ends of the shafts cut off.  I simply used a pair of bolt cutters – don’t hold onto the motor when you cut otherwise you might bend the shaft.  After cutting, I put them on the laser balancer, and they all appear to be fine with minimal adjustment.

Power distribution & Attopilot

Regretfully I forgot to take a photo of the Attopilot before heatshrinking it.  In the green heatshrink is a 90A Attopilot, used for current and voltage sensing.  That then attaches to the power distribution board.  Due to the length of the ESC cables, I had to attach small 3.5mm banana plug leads to the distribution board, unfortunately.  I was hoping I’d be able to solder the ESCs directly to it.

Power distribution installed onto bottom plate

The power distribution board is attached in this view.  Coming up through the middle is the sensor cable for my TM1000 telemetry module, and the current/voltage sensor cable for the Attopilot (current is white, voltage red, ground black).  Everything fits, although it was a bit dodgy with the large banana plug connectors.

Plates with APM Adapter mockup in view

The printout is a mockup of the APM Adapter board I’m having manufactured.  As you can see, it fits.  I had to shift some of the spacers that connect the very top plate in order to free up the 60x60mm holes.

At the moment the HT-FPV is sitting aside waiting for the APM Adapter to arrive.  Once that turns up, assembly can start in earnest!

APM2 Failsafe Configuration

As discussed in my last post, I installed a new APM2 PPM encoder firmware and intended to set up the failsafes on my APM2 and my Spektrum Dx8 transmitter.

Following the excellent guide you can find at DIY Drones (link here), I configured my Dx8 / AR8000 in exactly the same way.

Anyway – everything works, exactly as desired!  When I turn off the radio (ie, the AR8000 failsafe triggers) the PWM signal to the APM2 drops to ~ 911us, which is short enough to trip the failsafe.  When I pull the throttle line out of the APM2 (simulating a cable failure, which is what the new PPM encoder firmware protects against), the PWM throttle signal drops to ~900us, which is also low enough to trip the failsafe.

Thinking about it, there’s other failure modes possible – namely that you might lose other lines besides the throttle line.  I guess the solution there would be to either flip your mode switch to RTL, or simply turn off your radio thus triggering the normal failsafe and an appropriate response.

APM2 – PPM Encoder Firmware Update

The APM2 appears to have a serious (well, I consider it serious, the designers don’t appear to) issue.  There’s the normal failsafe, where if the throttle channel goes to a certain value (below 960us or so) then the failsafe code can kick in.  That’s great.

But if something happens in-flight and there’s absolutely no throttle signal at all, like say a wire breaks or a connection pulls out, then the PPM encoder on the APM2 by default will just keep sending the same throttle signal it got last through to the APM2 flight processor.

The end result?  Your quad stands a decent chance of just flying away into the wild blue yonder if you have this kind of failure.  Note I say IF.  It’s pretty unlikely you will have that happen, and the designers are right that it’s up to the user to make sure their hardware is reliable.

That all said, there’s a fix!  The latest version of the PPM encoder firmware for the APM2 fixes this issue.  Installation instructions are at this link;

http://code.google.com/p/arducopter/wiki/APM2Encoder

In addition to the resources stated there, you will also need the Arduino Mega 2560 DFU driver.  It’s a bloody nuisance extracting that from the Git repository file by file, so you can find a zipped up copy at my GoogleCode repository;

http://zencoding-blog.googlecode.com/svn/trunk/ardupilot/dfu_driver.zip

Enjoy.  Tonight I’ll be setting up and testing the failsafes to make sure everything’s going to work right, and will edit this blog when I have some more info.

Edit:  The PPM encoder update worked fine!  Details in the next post.

HT-FPV Frame Assembly & APM2 Adapter

I got myself a HoverThings HT-FPV frame for the new quadcopter.  The frame is made entirely from G10 fiberglass, aluminium spaces, and #4-40 bolts.

HoverThings HT-FPV Frame (black)

As you can see, it looks excellent.  The frame is intended for use with FPV.  I don’t have any FPV equipment, but that will be changing.  Anyway, during the assembly, I noticed a few issues;

  • Battery holder isn’t high enough to hold a 5000mAh Zippy Flightmax battery.  This should be resolvable with either a smaller battery or higher spacers on the top plate.
  • No landing gear.  However, I’m planning on putting on some landing gear (this and this) which I already use on my current quad.  I really don’t like the idea of having no landing gear at all.
  • There’s an aluminium spacer directly under the motor mounts.  Unfortunately the Turnigy 2217-16 motor has a shaft sticking out at the bottom of the motor that interferes with that spacer.  I don’t want to remove the spacer because it will affect the rigidity of the frame.  The solution here is to cut the shafts off with bolt cutters – they aren’t used for anything useful.
  • The middle plate has hole spacing for many popular board sizes.  Unfortunately it has no hole spacing that’s suitable for an ArduPilotMega2!  The solution appears below…

So, to resolve the last issue, I decided to go and make an adapter board to allow the APM2 to bolt onto the HT-FPV frame with either a 45mm x 45mm or 60mm x 60mm spacing.  Then I realized that I need to make some adapters so that the BEC power from my four ESCs only gets to the APM2 from one of the ESCs.

Enter the APM2 Board Adapter v1.0!

APM2 Board Adapter Top Render

APM2 Board Adapter Bottom Render

I’ve ordered three of the boards from OshPark this time around so that I’ve tried them out.  The adapter board itself has all passive components (I was debating adding some capacitive filter caps to the power section, but I decided against it), and has the following features;

  • Direction guide so you get your APM2 installed the right way around (on both side of the board)
  • Hole spacing for 45x45mm mounts and 60x60mm mounts
  • Up to eight ESC inputs and a separate BEC
  • Configurable power selection so that you can move a jumper to select which ESC supplies power to your APM2 (or none!)
  • Splits out the ESC signal line to a separate pin header
  • Large ground plane on the underside of the board to reduce interference caused by power systems under the board
  • No ground plane on the top of the board, and no traces inside the APM2 mounting area, so that the APM2 can be mounted hard to the board and scratches won’t reveal traces and cause shorts

Assuming I actually got all my dimensions right (I’m pretty sure I did), I should be able to attach this to the HT-FPV frame using the 60×60 holes, then attach the APM2 directly to it.  From there, I can connect my four ESCs to the ESC inputs, jumper one of them for power output, then I can take the PWM inputs off the board and the VCC out and I’m done.

We’ll see what it’s like when I get the boards.

AeroQuad Frame Redesign

As discussed earlier, I rebuilt my quadcopter frame.  The original design worked fine, but I had a few new design goals with this frame;

  • Integrated motor mounts with the arms for simplicity
  • Separate arm holding plates and electronics module for the ESCs
  • Bigger central FCU container (for an AQ 2.0.7 shield)
  • FCU board mounting direct to a levelled metal plate
  • Vibration decoupling between the arms and the electronics module

Anyway, here’s the finished result;

Finished frame with AQ shield installed

The finished frame appears above.  Design is very similar to the old design, except now there’s four plates.  Plates 1 & 2 bolt together to hold the arms in place.  Plates 3 and 4 bolt together with spacers to provide an enclosure for the ESCs and other electronics.  The ATmega attaches to the top of plate 4 with nylon standoffs, and the FCU container bolts onto Plate 4.

Finished frame, side view

The landing gear is a set of TRex 450 (I think) landing gear that I got from HK for like $3.  The battery straps on with velcro under the landing gear.  The end result is quite a lot taller than the original design, but it has a lot more space inside and can be disassembled in a much more straightforward manner.

AeroQuad 2.0.7a Custom Board

Got my customized AQ 2.0.7 board, and I spent the weekend transferring all my existing sensors and such across to it.

AQ 2.0.7a Custom Board

So, it turned out that the board actually has an error on it!  Eagle went and isolated a piece of ground plane, so you have to hook a small air wire from GND on the barometer to GND on the LLC.  Oh well.  Other than that, everything went pretty swimmingly moving my sensors across to it.  I had to completely ruin my old AQ v1.9 board to get everything out intact though.

It also turns out that the LEDs that are supplied with the AQ kit are not normal LEDs, they are actually LEDs with built-in resistors.  So my current board has no LEDs on it while I wait for my order of special LEDs to arrive.

The board worked pretty well straight off.  I discovered that the magnetometer is basically useless in its current position, too much interference from the power systems under it.  I’m thinking of desoldering the magnetometer board and relocating it onto a mast above the quad so it can get clear of the interference.