We made some quelab dalek buttons tonight and then were talking about justifying them by actually building a dalek, which seems like a potentially fun project.

There are some old BBC instructions floating around:

http://www.aceldama.com/~tomr/media/dalek/blueprints/

(Yup, it really is a plunger...)

And I hear Adric has a wheelchair base we might be able to build on, or I was quite amused by the idea of making a dalek roomba...

Thoughts? Links?

Is that two reversible DC motors? So we need two large H-bridges (24v, 15A) I propose an Arduino controller, with Bluetooth serial (like I used on my smoker). Then an Android app to drive it. Any Android developers listening?

Materials needed, from the BBC instructions:

28 lbs modelling clay
paper/polystyrene
1 roll 500 mm bandage
28 lbs fast-setting potter's plaster
4 sq yds hessian scrim
.5 pint shellac
1 tin car wax
.5 pint PBA release agent
2 oz accelerator
2 oz catalyst
strips of glass matt
2lbs gelcoat resin
6lbs layup resin
acetone
soap and water
1 tub barrier cream
sink plunger
2 car parking lights (for flashing lights on head)
2 6V .3amp bulbs and holders
6 volt battery

1.5mm ply 1 sheet 5'x5'
6mm ply 4 sheets 5'x5'
9mm ply 1 sheet 8'x4'
15mm ply 1 sheet 10'x4'
12mm wooden dowel
27mm wooden dowel
2 wooden balls 95mm diameter

24 polystyrene balls 100mm diameter
6 ball bearings 6mm
30 gauge fine aluminum mesh 275mm x 1470mm
24 gauge large aluminum mesh (2 strips) aluminmum 655mm x 180mm
1425mm x 150mm
and two strips sof aluminum 651mm x 10mm
40mm x 170mm

36mm aluminum or plastic tube 455mm
40mm aluminum or plastic tube 615mm
3mm steel rod 315mm
15mm rod 270mm
brass shim strip
aluminum angle
ribbed rubber flooring foam strip
3 plastic rotating castors 1.5in to 2in in diameter
perspex 2mm 80mm x 80mm
2 brass rings (internal diameter 28mm, cut to 10mm long)
screws, nuts, bolts, snap rivets, fast-drying enamel paint

Expensive? But it says the approximate cost is £15!

What's a vaccumformer?

(photo found online, both cute, and may have useful construction/clues.

Quote from Walter on August 9, 2012, 14:14
Expensive? But it says the approximate cost is £15!

What's a vaccumformer?

Expensive! that parts list looks closer to $200+

As for a vacuformer, they use them on mythbusters all the time, (also just about any blister packages containing any gadget you have ever had was vacuformed.) its a cheap way to make a mold or pull a shape into hot plastic.

you put an object down on something like a small airhockey table, (but instead of blowing out air, you suck it) then you place a sheet of heated plastic over the top and it gets sucked down tight. then it cools.

Ordered power MOSFET's. When do we start?
Would it be a mess to bring it to Maker Faire, and play with it there?

Plan: Arduino + Serial-over-Bluetooth module. Dual H-bridges. ASCII serial command (at least for testing) so we can drive it from a bluetooth terminal emulator.

Should I use 4 DIOs for each H-bridge and do all the direction switching in software, or should we add some logic to do a "direction" pin and a speed pin (2 DIO's per H-Bridge).?

Here's the schematic for 1/2 of the isolated H-Bridge design. DIR is a straight DIO, while EN is PWM. Software will need to note that when going forward, EN is active-high, when in reverse, EN is active low. So when going backwards, command analogWrite(255-speed).

This only needs 2 wires coming from the uController board.
Suggest putting 120 ohm resistors in series with DIR/EN lines both for safety, leaving the low-power board, and as current limiters to optocoupler LED's.

For other side of H-bridge, just flip the DIR and EN leads.

They recommend 470ohm for TTL, but since each of our diodes goes through 2 resistors (one on each side), and each line drives 2 parallel diodes, divide by 4, which is roughly 120 ohm.

Here is the test circuit for Arduino. Still waiting for opto-coupler chip to arrive.

Heat sinks are connected to PN-pair drains, and will also serve as motor connection contacts.

Got an opto-isolated H-Bridge motor controller running. Here's a drawing of the physical circuit layout. P-MOSFET's are pinkish, N are blueish, ground bus is blueish, high-side bus is pinkish, heat-sinks are blue-gray, and optocoupler chip is green

with the values coming from BT-Serial, what happens if the dalek goes out of bluetooth range? does it continue or will it require constant updates from the controller/deadswitch?

Updated schematic, as implemented, for the full H-Bridge optoisolated controller, which can be driven by PWM signals:

for completeness, here is the source code for the motor driver.
There is one main method,

    MotorDrive::setSpeed(int x)

where x is a number from -255 to 255. Negative values are for reverse.

Code is typical minimalist uController stuff.
You need to directly set up exposed state variables in your initialization sequence, like...

#define MOTOR_DECEL_FACTOR 4
MotorDrive MotR;
...
MotR.Pin.REV = 4;  // DIO pin for REV signal
MotR.Pin.EN  = 5;  // PWM enabled DIO pin for EN signal
MotR.setSpeed(0);  // make sure we are stopped at init

Main class:

class MotorDrive 
{
public:
  struct {
    int REV;  // reverse direction indicator, 1==Reverse
    int EN;   // Enable/speed pin. 255 is fast forward, but stopped reverse
  } Pin;
  int speed;  // current speed
  // Set speed -255 for max reverse, 255 for max forward
  void setSpeed(const int spdReq)
  {
    int spd = spdReq;
    if (spd < -255) spd = -255;
    if (spd >  255) spd =  255;

    if ((long)spd * (long)speed > 0)
      { // no change in direction.  just change EN value
        if (spd > 0)
          {
            if (spd ==  255) digitalWrite(Pin.EN,HIGH);
            else              analogWrite(Pin.EN,spd);
          }
        else
          {
            if (spd == -255) digitalWrite(Pin.EN,LOW);
            else              analogWrite(Pin.EN,255+spd);
          }
        speed = spd;
       //Serial.print(F("\t\tsame direction\t"));Serial.println(speed);
        return;
      }

    // check current REV pin state
    int revState = digitalRead(Pin.REV);

    // Weather we are stopped, or have a change in direction,
    // shut motor down first.
    if (revState)  digitalWrite(Pin.EN,HIGH); // HIGH to stop, rev mode
    else           digitalWrite(Pin.EN, LOW); //  LOW to stop, forward mode

    int prevVel = (speed > 0) ? speed : -speed;
    if (prevVel>10)
      { // kind of fast.
        // wait a while to slow down before allowing more commands
        Serial.print(F("\t\tDirection change... wait a bit"));
        delay(prevVel*MOTOR_DECEL_FACTOR);
      }

    if (spd != 0)
      { // should do these simultaneous, but I think it
        // is safe to allow sub-microseconds of reverse direction drive
        if (spd < 0)
          {
            analogWrite(Pin.EN,spd+255);
            digitalWrite(Pin.REV,HIGH);
          }
        else
          {
            analogWrite(Pin.EN,spd);
            digitalWrite(Pin.REV,LOW);
          }
      }
    speed = spd;
    Serial.println(speed);
    return;
  }
};

I think that a loop antenna will be the easiest for me to use as a directional antenna. A full wave resonant loop will have a diameter of 22 cm. A simple single loop should do the trick, and have an impedance of about 100 ohms.

/wordpress/wp-content/uploads/2012/10/0071475745_chap05.pdf

http://quelab.net/wordpress/wp-content/uploads/2012/10/0071475745_chap05.pdf

I think we'll do a (mostly) bolted scaffolding to attach the shell, and electronics too. We talked about all bolted, but I think that the main T at the back, hanging over the drive wheels is a bit stressing for bolts. I'll just do one weld for that main, back T structure, with a 2" angle iron for the main post, attached with the U bolts as the seat was, and 1.5" angle iron for the main bar. The rest of the structure will be from hardened steel bedframe.

/wordpress/wp-content/uploads/2012/10/rcr433rp_Mouser433MHz_4800baudRX.pdf

Datasheet for the RX modules I got on end-of-life closeout at Mouser.
50 Ohm antenna pin, "analog" out pin, and digital out pin.

I'm not sure of the analog output has some sort of automatic gain control. I hope not.
I would like to be able to use the analog output pin (through an LPF?) as a signal strength indicator, so I can use this radio RX for transmitter hunting.

This module has VERY small surface mount parts.
It may be difficult to attach to an internal node for a signal strength indicator.

I have some 433MHz TX's on order from another supplier.
I would like to set up a TX that does some sort of PWM signal that we can try to track from the robot.
Perhaps 66% PWM duty-cycle for "1", 33% PWM duty-cycle for "0".
Put the RX antenna on a servo, so it tracks the direction of the TX.
If the received bit is "1" try to move towards the TX. if it is "0" just track it's direction, but do not attempt to move towards it.

That way, a person holding the TX can hit a button, and the robot will try to get to them. Release the button, and the robot stops.

http://quelab.net/wordpress/wp-content/uploads/2012/10/rcr433rp_Mouser433MHz_4800baudRX.pdf

Add a ping sensor to keep the robot from running into things, and we may not need RC commands to drive the robot.
Just let it follow the beacon.

If we do a loop antenna, with peak sensitivity in 2 directions, you could get behind the robot with the TX, and "push" the robot around too.