Month: March 2015

The Clocks Go Forward. It’s Summer Time.

You wouldn’t know it was the start of British summer time looking out of the window. There’s a 30MPH wind and it’s raining again.

No flying today then, so I’m going to do some more work on the flight simulator and build a bit more of the autogyro. I’m now on the tail section now, so it’s starting to come together. At least I’ve got one working quadcopter that I can fly indoors now that I’ve fixed the motor on the Hubsan X4. If I get time I’ll have a look at the broken motor on the Q4 and see if I can fix that one as well.

How to Replace a Hubsan X4 Quadcopter Motor

This article shows how to replace a motor on a HubSan X4 quadcopter. The items required for this job are:

  • A replacement Hubsan X4 motor, H107-AO3 (mine were purchased as a pack of 4 for £14.95 from RobotBirds).
  • Small Philips Screwdrivers
  • Tweezers or small pliers
  • Soldering Iron and solder
The tools required

The tools required.

These are the tools required for the repair. I’m actually using a 60W soldering iron, which I use for most jobs, but a smaller one will do.

My Hubsan is over two years old now, so it’s the original version without the LEDs on the motors. Also, as it’s getting rather old, the black plastic case is on its last legs and is missing a few bits of plastic here and there.

The top right motor is the faulty one which is going to be replaced.

The Hubsan X4 being operated on - the top right motor is faulty

The Hubsan X4 being operated on – the top right motor is faulty.

The first job is to turn it over and remove the outer case. This is done by removing the three Philips screws and popping the arms out of the motor pods. I find that it’s easier to remove the single screw at the top, pop out the top two arms and then remove the final two screws at the sides.

Bottom casing removed, note the damage to the plastic case.

Bottom casing removed, note the damage to the plastic case.

Next, carefully remove the two screws holding the PCB onto the top casing. While it is possible to make the repair with the PCB in place, I found it easier to remove it in order to access the wiring.Notice that I’ve put the screws back into their threads in the top casing in order to avoid losing them. The bottom case screws have also been inserted into the case, which I’ve put to one side.

The PCB is removed.

The PCB is removed.

The motor is removed by pulling it through the top, i.e. the propeller “upwards” direction. You can see from the picture that mine suffered damage by striking a wall from the top, driving the armature through the bottom of the weak plastic motor back plate.

Make sure to free the wiring from the motor to the PCB first so that the motor is free to move. In my case, the back plate and motor can were in two separate parts, and, with the propeller still firmly attached to the motor, it was an easy job to pull the motor out of its mounting. If this is not the case, use duct tape (or sellotape), stuck around the exposed top part of the motor can to pull it free. You could use pliers to do this if you don’t mind damaging the motor though. In my case, I had forgotten to take one of the earlier pictures, so I had to remove the good motor and put it back again later. The motor is just a friction fit, so the tape method works very well.

The defective motor is pulled out through the top

The defective motor is pulled out through the top.

Now comes the soldering part. Make sure the iron is fully up to temperature before attempting the soldering. On a PCB this small, temperature is critical. DO NOT allow the joint to heat up too much, or the PCB will be damaged. I’m using a 60W soldering iron, but I do have a lot of soldering experience. The hot iron allows me to get the heat into the joint very quickly to melt the solder, but I’m using less than a second of contact with the iron on the solder pad. Use a smaller iron if that’s what works best for you.

Anyway, introduce a LITTLE heat to the plus and minus solder pads where the black and red motor wires connect and allow the wires to come away from the PCB. The wiring can now be pulled through the motor mount hole and the defective motor removed completely.

Use as little heat as possible to de-solder the motor wires

Use as little heat as possible to de-solder the motor wires.

Motor wires removed.

Motor wires removed.

And the defective motor is removed.

And the defective motor is removed.

Now, locate the correct replacement motor. I found that, in my pack of four, there were two with red and blue leads and two with black and white leads. This matches the diagonals on the quadcopter (top left and bottom right, top right and bottom left looking from the top). My old motor was one with red and blue leads, so I selected an identical replacement from the pack of four. I can only assume that this is something to do with the direction of rotation, but I don’t know what the differences are. It is possible that the timing on the different motor directions is different, but I don’t actually know the reason for the difference, or whether a white/black motor can be used as a replacement for a red/blue one. They only cost a couple of pounds each, so its sensible to replace like with like.

The new motor is pushed through the mount into position, after which the wiring can be put into position.

Four motors.

Four motors.

Wiring goes in first.

Wiring goes in first.

Then push the motor through. Make sure of the orientation so that the wiring sits flat.

Then push the motor through. Make sure of the orientation so that the wiring sits flat.

When the motor is pushed into position, make a note of the rotation of the motor in relation to the wires. The blue and red wires need to be in line so that the wiring sits neat and flat in the motor pod. Look at how the wiring for the other motors has been accomplished and copy that. Here, it’s worth pointing out that the wires on my new motor are slightly thicker than the ones on the motor that it is replacing. While this is good in terms of quality, it also makes it a bit harder to push the wires into the slots that need to hold them securely in place.

In order to solder the wires in place, I fixed them in place at the motor end before attempting to solder. I also freshened the solder pads on the PCB and ends of the motor wires very carefully with some solder just to wet them and increase my chances of a good joint.

Don't use too much heat!

Don’t use too much heat!

Solder complete.

Solder complete.

Hold the wires in place while soldering any way that you’re comfortable with. Use fingers if you can do it without burning yourself (also prevents you using too much heat), or use tweezers of small pliers. My advice is to stick one wire to the pad any way you can, then do the other one. Now both have a mechanical joint, it’s easier to hold both in place together and add a little more heat if you want to remake a joint that you’re not happy about. Solder joints should be shiny and rounded, not dull and pitted, which is an indicator of a dry joint. You can give the wires a little pull if you want to check.

My soldering here isn’t actually all that good, but the key is good enough. Heat is the enemy, don’t melt the PCB and make the tracks come off.

The next thing to do is check whether this actually works. Put the propeller back on and connect up the LiPo.

It blinks, it beeps, it binds!

It blinks, it beeps, it binds!

Put the propeller on and see if it spins.

Put the propeller on and see if it spins.

OK, it seems to work, so put it back together and see if it flies.

Place the wiring back neatly in the guides.

Place the wiring back neatly in the guides.

Loop the motor wires underneath the PCB.

Loop the motor wires underneath the PCB.

Place the wiring back neatly into the guide slots in the motor pod, arm and joint into the main body. I used a small flat blade screwdriver for this, but something non-metallic would be better. Make sure to loop the motor wires underneath where the PCB sits, ensuring that they do not obstruct the PCB’s screw hole. When all the wiring is neat and tidy, screw the PCB back on.

Screw the PCB back on.

Screw the PCB back on.

And finally, put the bottom half of the case back, making sure to be careful when popping the arms back into their correct places. My case is so badly wrecked, it probably doesn’t matter any more.

The little Hubsan is as good as new.

The little Hubsan is as good as new.

That’s it, now go and fly the X4 to make sure it’s all good. You might have to do the “Expert mode, full right rudder, zero throttle, waggle ailerons left and right until the eyes blink” trick to reset the trim on a horizontal surface. Mine flew fine as it was though.

I would say it flies like a bird, but actually, think, “angry hornet with attitude…”.

A Busy Morning’s Flying

We all thought it was a nice day this morning, but again, the forecast of mainly cloudy prevailed as the Sun rapidly disappeared and the wind picked up throughout the morning. By lunch time it was trying to tear the model out of my hand, so it was really time to go by then. With the direction of the wind as well, it was blowing through the trees and very turbulent low down, flying anything was a bit of a challenge.

When I showed up there were already two French guys there who I’ve never seen before. They had a small helicopter and a blue UMX Piper Cub, but it was already too windy for either of these to fly, so they just watched. I was closely followed by someone I haven’t seen for about six months and his young son, carrying a yellow and black Weston Magnum and a Twin Star MK2 in Slovakian air force markings. The Magnum is very fast in the air as you would expect and it had a number of flights in the windy conditions. Next to arrive was the guy with the EFlite Advance and his new(ish) Multiplex Xeno Flying Wing that folds in half and comes in its own bag. I really like this and it was amazingly only the second time that he’s flown it. In the terrible conditions it was going so well that it didn’t want to land and he was able to squeeze a little power on and go around again with ease. The fact that it folds in half means that you could actually carry it on a bike without any problems and the wing area and chord are fairly decent.

Next up was the only pure glider flyer of the bunch, carrying a Gentle Lady, another glider that I can’t name and a bungee. He had some really good flights, but we still can’t persuade him to try an electric plane. He did also bring a rocket which we watched him launch though, which was quite fun. Last up were an Italian family with lots of young kids bringing another blue UMX Cub with ailerons and a Discovery ST. Along with these guys was the owner of the Blade 350 quadcopter who does the aerial filming. I noticed that he now has a 12v lead acid battery like mine to charge his LiPos from at the field. He said it was his first time since September, so he just got in the air, had some practice and packed up around 1pm when the rest of us gave up due to the wind.

I managed to get in 3 flights with the RS352 and some really decent landings. That is apart from the last one where I was feeding in small amounts of elevator to correct, but nothing was happening. To avoid hitting the ground early I had to over control to gain about 6 feet in height, then continue the landing run. I was quite close to hitting the ground very early and fast, but I think it was just the windy conditions and maybe the flex in the airframe that causes this problem. It does the same thing in the air sometimes, so I’m quite used to it by now. In terms of flying, I managed a knife edge loop, which is something I’ve never done before. I was practising knife edge left and right, then thought, “OK, why not try the 360?”. The RS352 has much too much rudder movement, so it’s fairly easy to do as long as you have the height and can keep the aircraft on the knife edge. My problem is always with it moving away from me on the elevator as I find that very hard to judge.

I would have had a go with my HubSan X4 as well, but both my Q4 and my X4 are broken at the moment. The Q4’s front left motor just stopped working, so I suspect a motor failure or wiring problem as the LED on that corner still lights. With the X4 I managed to slam it into a wall top first, so the motor armature has gone through the bottom of the plastic motor back plate. When you look at the motors, the back plate is only cheap plastic, with two brushes that are little more than wires. There is no damage to the motor can or armature, only the back plate and brush wires, so it’s really annoying that it looks to be unfixable. I’ll have to get a replacement motor. I’ll just have to settle for the Quadcopter simulator until the replacement motors arrive. This weekend I went out and got a Mad CATZ CTRLR for my Nexus 4 phone so that I could practice with some controls which are closer to the real thing. The reason for all this quadcopter practice has to remain a secret for a little longer, but I’m hoping everything will have worked out by next weekend.

Anyway, the ATOM build is still progressing slowly, with the boom now being assembled and glued.

The ATOM with the boom fitted

The ATOM with the boom fitted

Top view of the boom

Top view of the boom

You might just about be able to see all the masking around the spruce longerons to stop them sticking to the main pod structure. I just hope that the glue hasn’t seeped into the structure and glued it permanently, as it now needs to be removed for the next step. It’s only just dried so I haven’t checked yet.

How Apt. It’s Raining.

It’s raining and I’m sitting at the computer writing a web-based rainfall radar application that uses the Met Office’s DataPoint open data. Anyway, enough about the PhD I’m currently writing in geospatial data visualisation, this is supposed to be a blog about model flying.

The ATOM with the soft balsa base added and sanded

The ATOM with the soft balsa base added and sanded.

As you can see, the ATOM build is progressing, if rather slowly at the moment. After what seemed like an awful lot of sanding, the soft balsa base is almost finished and the fuselage lines now resemble something like the photographs in the magazine. I’m almost at the point where I can attach the boom and tail, so it’s starting to look like an aircraft.

I’ve also been doing rather a lot of quadcopter practice with the HubSan Q4, but I can’t say why just yet. Let’s just say that I’ve watched Star Wars The Phantom Menace and Return of the Jedi again and if everything comes off, racing though the forest moon of Endor will be absolutely manic. A friend recommended the “QuadcopterFX” app for Android to practise on for FPV experience, which I now seem to be addicted to. I’ve also been looking at the journal articles written by the researchers at the GRASP lab at the University of Pennsylvania with a view to building my own quadcopter model for my own flight simulator. These were the guys who did all the crazy YouTube movies of quadcopters doing extreme manoeuvres controlled by a computer and vision system. It’s actually much easier to model a multirotor than an aircraft, so I’ll give it a go when I get some spare time.

Flight Simulator Update

It’s been a while since I posted any updates on the flight simulator, but I’ve been too busy with other things the last few months. Despite this, its current state is that my Great Planes transmitter joystick can now control the plane which flies, but using a simplified wing model which I’m not keen on releasing until I’m happy with how it flies. I’ll publish some more details on the flight controller later, as it’s basically an old analogue PC joystick that was designed to work though a computer’s gameport, but is now plugged into a USB converter.

I wanted to get my RS352 model into the simulator before going any further as the Gloster Gladiator is just an untextured grey model. I had already done most of the work in Sketchup, drawing around the photos to create an accurate model, so the next step was to figure out how to get the model into three.js. Sketchup has a Collada export function (.dae), but the model needs some additional work, so I had to import the model into Blender first. The results are shown below:

The RS352 model imported into Blender from a Sketchup Collada export

The RS352 model imported into Blender from a Sketchup Collada export


The textures are still on the model, but I’m having some trouble getting the export from Blender to Collada to load the textures into the flight simulator:

Blender's texture render of the RS352

Blender’s texture render of the RS352

I don’t know why Blender creates such a grainy render, but that’s all I can get out of it at the moment. It’s not like 3DS Max, which is much more intuitive to use by comparison.

Here are the flying shots of the untextured RS352 flying around the chessboard cube World:

The RS352 on a flypast through a 1980s computer game

The RS352 on a flypast through a 1980s computer game

Another shot of the RS352 in an audition for Tron 2015

Another shot of the RS352 in an audition for TRON 2015

 

It’s actually incredibly difficult to use the screenshot software on the Gnome desktop in Fedora 19 as I have to fly with the controller on my lap and try to click the mouse button to take the screenshot with my other hand. I really must find an application that works from a key combination.

So that’s it for now. I need to get the textures on the aircraft (and other objects), then get the flight model back up to scratch with something that reads the digitised wing polars from a real aerofoil test. I’m using the coefficient of lift Cl = 3.45*(Alpha-0.25) at the moment which isn’t good.

And then I’m thinking of fractal trees to make it look really weird…

Skylarks

It’s the first flying in March and the Skylarks are back again. There were quite a few of them around tweeting, darting about and then doing their trademark dive into the ground. There’s no grass yet, so they can’t be nesting, but it looks like a good few of them are back this year.

Anyway, back to the flying and it was quite a busy morning bearing in mind that the weather was horrible. Everybody had looked outside before coming out and thought, “it’s sunny and there’s no wind”, but then arrived to find the clouds had covered the sky and the wind had picked up. Despite this, I turned up to find the helicopter guy had beaten me to it, then was closely followed by the two more people who had a Hobbyzone Advance 25E and a pair of Art Tech Cessnas (one was a different type, which I didn’t see flying). We had some fun trying to reprogram the ESC on his Cessna to change the battery cut-off point, but ended up putting it back to its original setting. Having played around with the throttle quite a bit, I’m left with the impression that this particular ESC has a really good linear throttle response. I don’t know what the type is, but you could see that it was producing usable power over the whole range of the stick in a very progressive way and responsive way. The aircraft also flies extremely well.

After that we were joined by somebody who I’ve seen flying for years. He had a Hobbyzone Striker F27 which I’ve seen before, but he’s added telemetry in the shape of a pitot which reports altitude and speed back to his transmitter. Here’s a flying shot I managed to capture as it looks much more impressive in the air:

The HobbyZone F27 Striker in its element

The HobbyZone F27 Striker in its element. It’s not easy to catch one though!

As for my flying, I had three flights with the RS352 before calling it a day. I had wanted to try out my HubSan X4 which I seem to have fixed somehow. It must be circuit board problem of surface mount component loose as all I did to fix it was open it up and put it back together. The wind put a stop to any flying though, as it wouldn’t move forward into the wind at all. I’m also still having problems with the new HiTec servos in the RS352. I’ve swapped to a second Futaba R617FS receiver with the same Futaba FF8+TM7 Module and still get the servo jittering. Then I switched to a FrSky Tx Module and Rx and it looked like it might have fixed the problem. I need to test this a bit more, but it’s only a 4 channel receiver so I can’t use it in the aircraft. It does seem to happen when you move the controls suddenly, but it is happening on all the surfaces. It might be that it doesn’t happen in the air because of the load on them, but I have no way to tell.

After practising prop hanging (badly), flick rolls (not good) and stall turns (pretty good actually), the final landing of the day was a bit eventful. You have to remember that it was seriously windy and that the previous two landings were really quite good for the conditions. During the third flight the wind had almost dropped completely, then came back with a vengeance as I was landing. The wheels touched the ground, then a big gust of wind launched the plane 10 feet into the air, at which point I hit power, ailerons, rudder, down, lots of different control inputs, then landed for a second time that flight. That’s the first time that’s happened to me, so it was rather interesting.

Near the end of the session, a new guy turned up with his wife and little boy, plus a Parrot AR Drone and a Horizon Hobby Delta Ray. These are fantastic aircraft for anyone wanting to get into fixed wing flying as they have the SAFE panic button and look absolutely amazing in the air. I was standing next to him as he flew it for the first time and put in a few click of down elevator for him while he was flying as it seemed to be flying with its nose in the air. He used the panic button a few times and, with the blustery conditions, you could see a few times when the aircraft was thrown around and he panicked, but he was doing really well with it. Unfortunately, something strange then happened and as I was watching the aircraft went 90 degrees nose down, he used full up elevator and the panic button, but nothing worked. It hit the ground at 90 degrees full on the nose, which inevitably broke off, along with damage to the two wing mounted motors. The really strange thing was that, as we approached the wrecked plane, the motors were still running despite the throttle being at zero. As he picked it up the motors were running quite fast and the only way to stop them was to disconnect the LiPo. Very strange and probably indicative of an electronic failure leading to the crash. It’s the first time I’ve seen a Delta Ray crash like that though.

As the forecast weather front started to approach we decided to call it a day. The ATOM Autogyro is continuing and I’m almost at the point where I can add the booms.

It’s Light. It’s Dark. It’s Light. It’s Dark Again.

The forecast was for 20 mph winds this morning. I wouldn’t disagree. The wind is blowing rain clouds across the Sun, so one minute it’s bright sunshine, then it goes very dark and threatening.

No flying this week then. I’ll have to get on with building the ATOM. I’ve got the curved rear fuselage pieces that go up to the mast on, so I’ve just got to add a soft balsa base and then I can start on the tail section. I soaked them in water last Sunday so they fitted nicely. Then I glued them in place and sanded the curve up to the mast largely by eye.

It seems as though most of this aircraft is soft balsa block sanded to shape, so it’s making a lot of dust. Over the years I’ve started to get really sensitive to balsa dust, so it’s time to make good use of that 20 mph wind outside. Snowstorm in the neighbour’s garden anybody?