(article continued from previous page)
A simple launch controller is really all that most rocket enthusiasts require. However, I wanted to deliver more current, connect to an external power source, and support fancy blinking buttons.
Rocket launcher printed circuit board.
1. Power and igniter wires from the rear jacks are brought to a four-position Molex connection barrier strip (#38720-6304). It can handle 15 A @ 300 V. (By the way, check out the huge 0.250-inch wide PCB traces for carrying maximum current. They’re on both sides of the board.)
2. Vishay Schottky barrier rectifier (fast diode) #SB550-E3/54 to protect against reverse voltage when the igniter vaporizes. I have no evidence as to whether this is necessary or adequate. This is distinct from the Vishay TransZorb transient voltage suppressor (P6KE6.8CA-E3/1) that in connected across the relay coil to protect against the relay’s inductive spike.
3. Tyco SPST power relay (T90S1D12-5) controlled by 5 volts that can handle a 20 A resistive load at 28 VDC. A relay is a good choice for a launch controller because it doesn’t need to be pulsed, it has a low resistance, but most importantly it won’t conduct if the power source is connected backwards.
In contrast, power MOSFETs and low-voltage-drop IGBTs have body diodes that will conduct in reverse. A diode could be used to prevent reverse conduction, but would consume 5%-10% of the igniter power.
Furthermore, MOSFET and IGBT transistors have gates that can be triggered with very little current. There is some security in knowing that the bipolar transistor that drives the relay coil needs more current to trigger. Static electricity or stray RF isn’t going to cause the relay to trip and accidentally supply power to the igniter.
I probably should have made the relay board separate so that it can be replaced. But it’s improbable that it will see enough use to wear out the contacts.
4. Analog devices instrumentation amplifier (AD8226ARZ-R7) to amplify the tiny voltages across the low resistance igniter during continuity testing. This chip was selected because it can handle ±35 V at its inputs without damage. During ignition, the battery may supply up to 20 V across the igniter.
The output of the amplifier chip feeds into a microcontroller analog-to-digital converter. When the user wants to test continuity, the microcontroller feeds 5 V through a 220 ohm resistor (22 mA max) through a diode and into the igniter.
The actual voltage depends on the gain setting of the trimpot connected to the amplifier chip, the resistance of the cables, and the resistance of the type of igniter being used.
When testing and debugging the launch controller, you don’t want to burn off real igniters, particularly indoors. Since this launch controller uses a 12 V battery, I used a 12 V automotive lightbulb as the practice igniter.
GE 90901 S8 high intensity 12V 12W auto bulbs.
The resistance of this particular 12 W bulb is around 1 ohm when unlit. The resistance of the Estes igniters is around 0.72 ohm. That’s pretty close.
Automotive lightbulb with soldered wires to act as igniter during debugging.
Soldering wires to the tip and side of the lightbulb makes it considerably easier to hook up to the igniter cables.
The launch controller sees the lightbulb as an igniter with good continuity. When the launch sequence completes, the lightbulb emits a brief flash as it receives power like the igniter would.
5. Atmel ATmega168 microcontroller brains. Ultimately, all of the switches and analog voltages are fed into this chip, which makes decisions about the state of the launch, and outputs to the status LED, to the button LEDs, and to the igniter relay (via a 2N2222 transistor).
Lockout Key Switch.
Even the best programmer should be cautious about putting their safety in the hands of software. So, the output pin of the microcontroller that controls the igniter is sent through the keyswitch. That is, no matter what the microcontroller tries to do, it cannot initiate a launch if the keyswitch is in the safe position. Power simply won’t reach the transistor.
Keyswitch pinout for P201133WM03NQ2.
When the key is turned to the white line on the switch, the switch pins underneath correspond as follows:
By connecting the base wire of the relay transistor to pin 3 and the microcontroller to pin 1, the microcontroller is only electrically connected to the relay when the switch is turned to the second position. Of course, the microcontroller still needs to apply power at the appropriate time to inflame the igniter.
The second pole on the keyswitch allows the microcontroller to detect the position of the keyswitch, by applying ground to a microcontroller input pin depending on the keyswitch position.
The keyswitch is a black, low-profile, DPDT (double pole, double throw), solder-lug switchlock from C&K Components (P201133WM03NQ2). It costs $6 from DigiKey (part #CKC8034). That’s fairly inexpensive for a switch with a key (search for “keylock” on DigiKey for comparison). You might be able to salvage a keyswitch from an older computer, but it may not have the number of poles that you want.
The C&K Components switch has the following advantages:
The C&K Components switch has some minor issues:
6. A 330 µF capacitor. The battery will supply a lot of current to the igniter, which will cause the battery voltage to drop considerably. To prevent the microcontroller from resetting during this period, it is important to have plenty of capacitance to maintain the regulated voltage level.
Regardless of how long the user holds down the button, the microcontroller only supplies power to the igniter for 200 milliseconds (0.2 seconds). This reduces the opportunity for the battery, relay, connectors, or other circuitry from being damaged by a long-term short circuit.
The launch controller uses a L4931CZ50-AP low-dropout 5V voltage regulator. It can output 250 mA @ 5 V with input voltages from 20 V to 5.8 V. The circuit won’t draw on the capacitor as long as the battery can supply 250 mA @ 5.8 V when it is also supplying power to light the igniter.
In the event that the microcontroller resets, the relay coil will automatically disengage because the microcontroller defaults its pins to inputs at startup. The microcontroller examines the state of the switches at startup, and will flash an error if it isn’t in the safe state. Therefore, I know that the circuit doesn’t reset with the battery I’ve chosen, because the microcontroller shows a success light after launch.
In this case, I’m using the MOSFET driver as a level shifter to provide 12 V outputs to the LEDs built into the launch button and arming pilot switch. Although those switches contain ordinary LEDs, they are wired in series with each other or a resistor that only displays brightly enough at 12 V. Two microcontroller output pins provide either 0 V or 5 V to each of the 4427 inputs, which then output 0 V or 12 V to the switch LEDs.
8. Bipolar 2N2222 transistor drives a Star Micronics HMB-12 (EIAJ:MB-RPD-C16-22) magnetic buzzer. It produces 94 dB with 23 mA of current at 12 V. It probably could be driven directly by the microcontroller, but I didn’t know what type of buzzer I wanted to use when I first designed the circuit.
The buzzer alerts the user (and those around them) every second of the five second countdown (the user can continue to count to 10 if desired). Thus, if someone quietly initiates a launch, the launch controller will alert everyone.
9. Vishay high-speed dual-channel optocoupler VO2631. This allows for remote operation of the launch controller without the possibility of electrical issues.
An optocoupler is just an infrared light emitting diode across from an infrared sensor diode in an electrically insulating package. The sender turns on and off the light-emitting diode without needing any electrical connections to the rest of the receiver circuit. As such, if the sender is at different voltage levels, or has some sort of spike or critical failure, it doesn’t mess up the receiving circuit.
The advantage to being able to remotely control a launch is that the battery can be near to the launch pad. Usually, the launch controller and battery are at least 15 feet away from the igniter, leading to power losses in the long cables.
If the launch controller receives a very specific sequence of blinking-light bits, with appropriate time delays in between, then the launch controller powers the igniter. Any error in sequence would require the sending circuit to start over. This would make it extremely unlikely (impossible) for random bits to cause a launch. Furthermore, the launch key would still need to be in the correct position, and the launch controller would still buzz a five second countdown.
Next we'll examine the battery, banana jacks, and cable...