• Replace the Propeller Adapters
• Replace the Motor Mounts
• Enclose the Electronics

Replace the Propeller Adapters

The jDrones kit includes propeller adapters that are pretty difficult for me to install properly.  The adapters are the collet type and they kept falling off sending the propeller into the air and the copter to the ground.  Here is a video of the propeller falling off.  The camera was mounted on the quad and was pointing directly at the propeller that fell off.

I eventually replaced these adapters with the E-flite EFLM1930 adapter with set Screw (1/8″ rotor).  The set screw holds the bottom portion of the adapter firmly to the rotor.  The top portion of the adapter can be tightened independent of the bottom portion.  I use the blue thread-locking fluid on the top half and have never had another propeller fall off.

Replace the Motor Mounts

Many of the crashes resulted in the plexiglass motor mounts being cracked or completely broken.  I took the step of mounting the motors directly to the arm by adding another mounting hole like this:

I used the motor mount template from the kit to locate the additional mounting screw

Two 20mm M3 metal machine screws, lock washers and spacer washers. The space washers provide clearance at the bottom of the motor for the rotor.

The mounted motor

Enclose the Electronics

I plan to replace the landing gear, but the kit landing gear protects the electronics complex.  Here is my solution:

Use a DVD/CD spindle to enclose your electronics

The DVD/CD has been machined to replace one of the fiberglass boards in the frame.  Here are the steps I used:

1. Get yourself a 30 disc (very tight fit) or 50 disc DVD/CD spindle
2. Download and install Google Sketchup
3. Download the model for the main plate from the 3D warehouse
4. Add a circle to the center of your model that is the same size as the inside diameter of the bottom portion of your spindle case.
   

   This image might work for you with some scaling

5. Print the model so that you have a 1:1 exact pattern of the main plate (compare it with the actual main plate to make sure it is right).
6. Use scissors to cut around the circle on the printed model so that it can fit inside the bottom portion of the spindle case.
7. Use a Dremel or similar tool to carve out anything on the bottom spindle that is in the way of having the printed main plate circle from laying flat.  You might need a few spindles for practice while you perfect your technique.
8. Use your favorite brand of white glue (e.g. Elmers) to completely cover the blank side of the printed main plate circle and secure it to the bottom portion of the spindle.  Let it dry.  It is important to make sure you get the surface completely covered so that your pattern stays in place during the next step.
9. Use a Dremel or similar tool to drill and carve the pattern of the main board into the spindle bottom.
10. Use the same tool to allow room for the arms to lay flush with the new main board.
    

    Clear away material to allow the arms to lay flat on the surface of you new main board.

11. Assemble your quad using the bottom portion of the spindle as the top main board.
    

    The machine spindle base as a replacement for a main board

12. Screw the DVD/CD case top onto your quad.  You may need to rearrange your receiver and GPS to make it fit.  I also added a hole at the top to make enough room.

The quadcopter fully assembled and tested - ready for the first flight.

You can see the credit card sized camcorder that will soon take a ride.  You cannot see just how jumbled the wires are so here is a different shot.

Yes, it is supposed to look like this ... it's a prototype. I have ideas to make this look slicker later.

Today we took it out for the first flight.  All the technology built into this thing was no match for the novice pilot.

There were a few hard landings causing some minor damage – a few of the polycarbonite screws on the landing gear sheared off.

This is what the front looked like.  The drifts are about three feet.

After about an hour, I was over heating so I switched to lighter jacket.  It took about 1.5 hours to get this far…

Give Sue’s recent remarks about my lamenting about the blizzard, here are some manly shots…

After two hours, the driveway and mailbox is cleared.  The sidewalk will have to wait for another day.

Prior to my arrival, I booked the Total London Experience with Golden Tours. The tour was fast and furious with the bus barely slowing down at some sites.

7:30 AM – Golden Tours is already late and finally picks us up at 8. We spend the next 45 minutes picking up other people from other hotels in the London area. The streets are quiet on a Saturday morning.

A typical London street early on a Saturday morning

9:00 AM – After being dropped off at a central bus station, we are now on our tour bus (not a double decker). Our guide highlights narrow streets originating in the middle ages, Oxford Street and Hyde Park.

Some streets are narrow because they originated in the middle ages

Earlier in the week, I had taken the time to walk to Oxford Street to buy a sweat shirt at Primark. That was a unique shopping experience that I can best describe as organized looting.

We drive through Piccadilly Circus with barely enough time to snap this picture:

Piccadilly Circus in two minutes – the duration of one traffic signal

We literally drive past Trafalgar Square and I got these pictures while the bus was moving!

9:45 AM We arrive at Big Ben and the Houses of Parliament. Since this is such an iconic landmark, I took extra pictures of it.

We coasted past Downing Street long enough to glimpse the residence of the Prime Minister.

The black house on the right is where the PM lives

10:00 AM Our first actual stop is at the Household Cavalry on Whitehall. There we were able to take pictures, interview a guard and tour the museum.

11:00 AM We arrive at Buckingham Palace for the changing of the guard. The crowd is already swelling and we are only able to see part of the procession.

Noon We arrive at Covent Garden for lunch at a local pub named Marquess of Anglesey for a chicken dish and pint of Young ale. We have about an hour and a half for lunch and exploring.

1:45 PM We arrive at St. Paul’s Cathedral. This is the most stunning stop of the tour, but no pictures are allowed inside. A wedding is underway with a double decker special awaiting the wedding party.

3:00 PM We arrive at the Tower of London for the longest queues in the tour. Here we meet a Beef Eater, see the Crown Jewels (no pictures allowed) and spend time exploring the area.

The Tower Bridge as seen from one of the sentry walls at the Tower of London

5:30 PM We are in a queue for the tour of the Thames on the Thames Clipper.

6:15 PM We are in a queue for the London Eye. There is a short 3D movie that was pretty cool, followed by another queue. Lots of pictures on the Shutter Fly site link above.

7:35 PM I walk across the bridge to the Embankment Tube Station for my first taste of the underground. It costs 4 GBP ($6) for a 15 minute ride back to Paddington and the train is not air conditioned.

8:15 PM I have dinner at the Monkey Puzzle pub. Dinner includes Biggles Sausages (aka bangers) and mashed potatoes (and a pint of another English ale).

9:30 PM Back at the hotel to unwind. Little do I know that the next day’s travel will be 19 hours.

Event-driven programming enables you to focus on robotic behavior while the event infrastructure notifies you when something interesting happens. The event infrastructure is typically part of the operating system or application framework. In this case, HB-ICE provides the infrastructure. HB-ICE is just the minimum to monitor the Handy Board buttons and digital inputs. Other extensions will be added later.  I intend to use this to implement reactive behavior, hopefully with subsumption.

Button Events
Let’s start with an example. Here is the code that is notified when a button is released:

void handleButtonReleasedEvent(struct Event* event){
    int eventId = event->id;
    switch(eventId) {
        case EVENT_ID_START_BUTTON:
          start();
        break;
        case EVENT_ID_STOP_BUTTON:
          stop();
        break;
    }
}

When a button is released, a pointer to an event is passed to the function. The id of the event is used to determine what action to take.   Events are structures like this:

struct Event {
    int type;
    int id;
    int value;
    long timestamp;
};

The type field helps distinguish button events from digital or analog events. The id further distinguishes the event within the type. For example, a button released event needs to indicate if the start or stop button was released. The value field is a copy of the sensor, input or button state. The time stamp field is assigned to the event when it is signaled. I plan to use that for meta-sensing in the future.

In handleButtonReleasedEvent, the program either calls start() or stop(). This is used to start and stop the robot behaviors. So how does this all get started? Let’s look at the example program main() function and supporting functions:

void initialize(){
    initializeEventQueue();
    intializeDigitalEvents();
    initializeMotors();
    startMonitoringButtons();
}
void handleEvent(struct Event* event){
    int eventType = event->type;
    switch(eventType){
        case EVENT_TYPE_BUTTON_RELEASED:
          handleButtonReleasedEvent(event);
        break;
        case EVENT_TYPE_DIGITAL_CHANGED:
          handleDigitalChangedEvent(event);
        break;
    }
}
void main(){
    struct Event* event = NULL;
    initialize();
    printf("press start\n");
    while(looping){
        event = dequeue(eventQueue);
        if(event != NULL){
            handleEvent(event);
        }
    }
    printf("done\n");
}

The main function calls intialize() and then enters an infinite event loop.  The initialize function initializes the event queue, digital events, motors and then starts monitoring the buttons.  The buttons are monitored in a separate process.  When a state change in a button is detected and the button is released, an event is added to the event queue. The event queue is a classic first in, first out (FIFO) queue.  The event loop in the main function takes events off the queue and calls the handleEvent function.  This function routes the event to the appropriate handler based on the type of event.

The HB-ICE distribution also supports the digital inputs. I have been using this infrastructure for the Handy Bug exercises in Fred Martin’s book entitled Robot Explorations and it appears to work.

Source
The source is available as a zip file via MediaFire hosting or from the Source Code page.

My RCX died from battery acid that had leaked from batteries that I left in years ago. Since I have the Handy Board, I have no use for the RCX. The RCX has six very specialized electrical bricks. The RCX has two rows of three connectors of these electrical bricks for Lego wires. After taking apart the RCX, I found that all six are easily popped out of the RCX shell. I have hacked these bricks to provide an easy way to preserve my current Lego wires and add additional length for interfacing to the Handy Board.

The RCX electrical brick in various stage of the hack

The brick on the far left is the unchanged brick out of the RCX. The brick is held in place in the RCX by two tabs. I scored the tabs with a knife and easily bent and broke them off. This is shown in the middle brick. I used an old 28 gauge ribbon cable to create a custom cable for the adapter. I soldered the two wires to the metal tabs of the electrical brick. Care should be taken not to apply too much heat or the brick could melt. The results of the soldering are on the right side. Next I bent the metal tabs down and applied electrical tape as shown below.

After bending the tabs down

After applying the electrical tape

Now this adapter can be connected to an existing Lego wire without cutting it. I can also extend the length by using a different length of ribbon cable.

This is how the robot from chapter one looks with the sensor adapters installed.

The sensor adapters are installed under the robot and wired to the Handy Board ports

The IC environment is primitive in comparison to the integrated development environments I deal with in my day job.  I have not experimented with it much but it appears good enough.  I can’t complain about the price either (free).

My next step is get the parts I need to connect my old Technics motors and sensors to the Handy Board so I can start to follow the book in earnest.  Those parts were not at my local Radio Shack, so I had to order them online.  The Handy Board uses a standard 0.1″ circuit board header as the template for connectors.  You have to cut four pins off of the header and solder a few wires but that should be easy.

Here is a brief review of the product I ordered.

I plan to use the Handy Board and the book entitled Robotic Explorations by Fred G. Martin to get started.

Today, I ordered the Handy Board from Gleason Research.  It should be here in about a week.