Archive for category: 3DoT Robots

By: Kevin Moran (Electronics and Control Engineer)

Once the PCB was assembled by the manufacturing engineer (Jerry Lui), it was to test it. If all the testing and assembling had been done correctly then the PCB should have worked with a problem. As testing began we quickly realized that the PCB was not sending any voltage to the IR emitter. With the help of Jeff Cool, Tae Le, and Jerry Lui, we realized that the problem was with the NPN transistor we had ordered.

The problem was with this 2n2222 transistor above. The RB was not being used as a base, but rather the collector (which should be connected to ground). After many attempts, we came to the conclusion that the SMD transistor had been placed wrong.

As can be seen by the arrow, once we compared the diagram of the transistor, with this picture, we realized that the resistor RB was going to the right connection of the transistor, when it should have been pointed to the left connector, and the right connector straight to ground. Thanks to the ingenuity and quick thinking of Jerry Lui, he was able to quickly fix this problem, by flipping the SMA transistor over. After that was fixed, we tested the IR emitter and we saw voltage going through it, and the receiver worked perfectly as well. Below is the final output of the PCB, converting an analog signal into a single bit digital signal! YAY!

As can be seen, a perfect digital signal with either a 0 or 1 bit. On/off applications. Now we are ready to defeat the 3DoT Spider bot.

Sources:

http://www.mouser.com/ProductDetail/TTElectronics/2N2222ACSM/?qs=%2fha2pyFaduhY4tRQXhKL%2feLyX%252bM6m2ylmJNPJCp%2fVeaSnptWjDmpSA%3d%3d

By:  Tae Min Lee (System)

Mission objective also known as Con Ops was needed to set the rules for the laser tag game.  Both systems engineer from the SpiderBot (Chris Hirunthanakorn) and the Goliath (Tae Min Lee) discussed and finalized the rules for the laser tag game.  The list of rules that are listed below will allow us to have a clean and fair match.

Con Ops:

1. The game will last for three rounds. One round ends when a robot is hit three times and is disabled. This may last up to 10 or 15 minutes.

2. A robot is considered disabled if it is unresponsive to commands for 10 seconds after the final tag.

3. At the end of one round, both robots will return to their starting positions.

4. The game will take place in a 6 ft x 6ft area on the linoleum floor of ECS 315. If any robot leaves the designated area, they are not allowed to attempt to tag the other robot and must re-enter the game area.

5. An Infrared LED emitter and Infrared detector will be used as the tagging system.

6. The maximum distance for detecting a direct hit will be 5 ft. This means the IR emitter is hitting the detector in a straight line from 5 ft away. Whatever voltage is outputted by the detector will be the minimum threshold for defining a hit. (need to define that voltage here once we figure it out)

7. Collisions between the two robots is not allowed. If a collision does occur, the robot that caused it will take a penalty hit. Once the penalty hit is taken, the robots move back to their starting positions.

8. A buzzer will make a noise to indicate when the robot is tagged and when it is disabled.

9. When a robot is tagged, it will make a sound and it will be deactivated for the next 5 seconds. This will allow the robot to move out of the way and not cause multiple tags at once.

By: Kevin Moran (Electronics and Control Engineer), Tae Lee (Systems- Provided the musical notes code needed for this game)

In this post, I will be discussing the final results for the IR Emitter when being used alongside with a Schmitt Trigger and the code that will be used to determine when the 3DoT Rover gets hit 3 times by an IR emitter.

I will first show below what happens when this final circuit is connected to an Arduino analog pin, although the Schmitt trigger is technically used to convert an analog signal to digital, I feel it is important to understand these changes in analog first.

• I first tested the output by connecting it to the Analog input of my Arduino Uno
• The final output of the signal by using the correct resistors values has very little noise and it is controlled within our 0-5V range
• When the IR light hits the receiver, the signal from the photo-transistor is passed by the Schmitt trigger. This trigger then inverts the signal, and reduces the noise from the input giving us the output shown here.

As can be seen this signal is inverted from the original IR receiver signal, and the final output has relatively zero noise.

Since this trigger is an analog to digital converter, I also decided to test the output with a digital Arduino pin.

• I decided to test the circuit with an Arduino Digital Input (2) just to ensure if asked, the PCB would work with either analog or digital input.
• The concept is the same, except the outputs are either 0v or 1v.
• The signal in the output is different from the analog output, they are peaks.
• As you can see from the Arduino plotter, the circuit detects the IR light faster, and counts more hits/second.

However upon closer inspection, I was told my output is not supposed to be spiked, but rather look similar to the analog output. One reason might be the Op-Amp I was told to use. I will continue testing and resolve this issue.

The Code:

This code is used to track the number of hits on the 3DoT Rover when testing with the analog pin. It consists of a loop that is activated once the voltage on the receiver reaches 3.5v, and counts the number of times it reaches this level or higher. Below I present the pseudo code and the flow chart in hopes of better explaining this code.

Flowchart:

Once the micro-controller detects 3 hits, it will play a short song and then reset, and start over again.

Sources:

https://www.arduino.cc/en/Tutorial/ReadAnalogVoltage

By: Jerry Lui ( Manufacturing Engineer)

The general layout of the PCB was determined by the how sensitive the signal was, the specific wiring of IC’s to surface mounted components, and mounted surface area. Power input was placed away from sensitive signals and had a trace width of 32mm @ 1oz/ft^2 copper thickness (determined by the PCB house/fabricator) with a total temperature shift of 10C which allows total current of 1A. The trace thicknesses can be quickly determined by the following chart:

While laying out the board make sure to perform a DRC error check frequently. The included .DRU error checking file has generic limitations which may or may not adequately represent your fabrication houses’ design requirements (EagleCAD will always flag drill and VIA holes). The fabrication house will generally supply a .DRU file that can be used in place and will often allow traces to be placed closer together.

All of the signals were moved as far from the power input as possible in this case I have placed on the top left corner. The large power supply capacitor was also placed as close as possible to the power input and the rest of the circuit was laid out to take power from that capacitor. Also, components were placed similarly to how they were wired in the schematic as much as possible so that they are generally easier to follow and understand.

Note that there are red dotted lines (and blue underneath) along the perimeter that represent the ground polygon which is needed to create the ground plane (red for top layer, blue for bottom layer). The isolation width (free space between the traces and ground plane) was set to 12mils by default but was changed to 16mils as a precaution. This value is mainly determined by the fabrication house limits.

To place a trace on the other side of the board while you’re currently laying a trace (top to bottom) click on the middle mouse button. This will automatically create a VIA (electrically connected through hole) that connects both traces.

TIP: If you don’t create a ground plane first you can right-click on any GND signal and hide it. This will reduce the air-wires shown so that laying out the PCB will be easier. All GND signals will be hidden with this option but can be restored by typing “ratsnest *” in the command line.

TIP: Component values can be hidden by disabling the tValues layer. This makes it significantly less cluttered.

In the picture below, the ground plane has been created by hitting the “ratsnest” command. Pay close attention to capacitors during the ground plane creation. The “relief” option should be enabled so that the negative terminal can actually be soldered easily; this option is set on by default.

The mounting holes seen above were created using VIA’s. The size is determined by what screw or mounting system that’s planned on being used. In this case the size was set to 86.6mils or the standard inner drill size of a M2 screw.

Text such as the label “Goliath spr’16” can be placed using the “text tool” and should be placed on the tPlace or bPlace plane. Make sure to properly label components and the positive terminal, negative/ground terminal or both.

This is what the board will look like after being fabricated from OshPark.

Conclusion

Laying out a PCB can be time consuming depending on the amount of components on your board. Never use auto routing as the algorithm used is not that great. Remember to account for trace thicknesses, pin locations, and temperature limits for components.

By: Kevin Moran (Electronics and Control Engineer)

Since the analog output of the receiver is very noisy, my division manager suggested I use a Schmitt trigger which has 2 jobs:

1. It reduces noise from the circuit through hysteresis, the time-based dependence of a system’s output on present and past inputs.
2. It inverts the signal from the detector circuit and creates two thresholds-high and low. By changing resistor values we are able to control the high and low thresholds.

Below is the graph of a generic output using the Schmitt Trigger, as can be seen the signal is inverted, and the noise is relatively gone.

Now the challenge was selecting the correct Schmitt trigger that would work best with our receiver, after testing various resistor values, I made my choice between two sets of resistors.

I decided to go with the resistor values from Trigger 1, because It has a larger gap between the high and the low thresholds. When tested with the Phototransistor it provides a low threshold closer to 0 voltage

All values were taken with multi meter while changing the input voltage

Note: The high and low threshold change when the Resistor value connecting the receiver is changed from 100k ohms to 2M ohms (I did this in order to increase the sensitivity of the receiver)

Testing with Arduino Monitor:

• When light is applied to the photo-transistor and the input is placed through the Schmitt trigger
  • High Voltage= 3.76 V
  • Low Voltage= 0.02 V

In order to make the phototransistor sensitive enough to be able to detect the IR light, I am using a 2 mega ohm resistor from the detector to the 5v supply

Both the high and low voltage thresholds are shown using the Serial Monitor

This is the result.

I also conducted a test on the Schmitt Trigger using a potentiometer

Components for this test:

10k potentiometer

LM358 Op-Amp

Resistors: 100k, 51k, 51k

Capacitors:  10 microF

Arduino Uno

The threshold between a high or low input from this test is 2.6V

As I varied the resistance in the potentiometer I obtained the graph provided in this slide

Based on this experiment I concluded that my Schmitt trigger was ready to be used alongside the photo-transistor.

Below is the result from this test using both the Arduino plotter and monitor

As can be seen again, a very clean signal that has only two outputs a low 0.02v and a high at 3.78v

Sources:

Jeffrey Cool: Division Manager

http://www.ti.com/lit/an/scea046/scea046.pdf

http://pcbheaven.com/wikipages/The_Schmitt_Trigger/