Power electronic circuit containing a hardware Trojan

The goal of the project is to design and fabricate a power electronic circuit contains a hardware Trojan.

=Problem Definition= We first found the power converter circuit and became familiar with its structure and components in the circuit. Once the internal and external structure of the circuit is known, we start to simulate it on the computer to view its current, voltage and power characteristics in order to find a way to destroy the circuit and hide our points of destruction. 1. Analog circuit 2. Determine the specifications of the components by detecting the current and voltage of the components 3. Buy components and build circuits 4. Experiment and debugging

There are two key points to this project. 1.One is how to destroy the circuitry. Is it permanent or temporary? Was it silent destruction or an explosion in the circuit caused by too much power? 2.The other is how to hide the point of destruction. Do you make it as small as possible or do you use camouflage like animals in nature? These are all things that need to be explored.In this project, we specifically studied the buck converter circuit and the boost converter circuit.

=Background= With the development of electronic technology, network security has become more important. However, hardware security is also very important. The protection of high-end technology and data security of these electronic devices have become increasingly prominent. When some confidential documents fall into the hands of the enemy or opponent and the security staff cannot protect them, destroying these documents is the best result. Therefore, it is necessary to install a Trojan in these files in advance.

=Objective= Our project focuses on how to create a power electronic circuit with the ability to self-destruct or modify its behavior on demand. This project will be challenging both the Buck and the Boost converters, specifically.

=Choose a suitable circuit=

Introduction
The Boost converter was designed using basic calculations to yield a 50% duty cycle and a switching frequency of 120KHz. With these specifications, the ideal output voltage is 10V with a 5V supply voltage.

Process
Designing the kill circuit used a simple transistor to sink current at the output to prevent the smoothing capacitor from charging. This allowed a way to run simple simulations to see the effects on preventing the boost circuit to function as designed.

Actual design of the kill circuit used an IR LED and receiver to replicate a remote disabling device. The kill device intercepts the functionality of the boost converter by the use of a button. See circuit below.



In physical application, the diode on the boost converter and can be destroyed entirely by holding the button. But testing purposes only demonstrated the button pressed for short period to show how the output voltage drops well below that of the input voltage.

Simulation result
Simulations for the boost yielded an output around 7.3V. This is comparable given the realistic simulated components that LTSpice provides.



Actual readouts from the physical design were comparable when using a 5V power supply, function generator set to produce a square wave using 50% duty cycle at 120kHz.

The physical circuit produced an output voltage around 8.6V. These results are given the components measurable values.

Experimental results and analysis




As seen above, the current is being drained through the transistor causing irregularities on the capacitor before the output voltage settles to around 2.6V. This clearly disables the functionality of the boost converter, resulting in a successful kill of the circuit.

Introduction
This circuit is called a buck converter or a step-down converter because the output voltage is less than the input.

This is another switching converter that operates by periodically opening and closing an electronic switch.

A buck converter circuit was chosen with a frequency of 120KHz. The initial voltage is 12V, and the duty is 0.5. The value of resistor is 50 ohm.

Process
1. Calculate with the known initial voltage and frequency to get the capacitor and inductance values. 2. Determine the characteristics of the sensor and diode. 3. Use LT spice for analog circuits. 4. Buy components online and build circuits. 5. Debugging and correction.

Simulation result
We selected the appropriate components according to the data calculated before, and built the virtual circuit, and the final simulation result and the calculated result are both 6V. And we also checked other components, they are all working normally.



Experimental results and analysis


First, we use components to build a buck converter circuit. Then we use the function generator as the pulse voltage, and connect it to the oscilloscope, adjust the frequency to 120KHz, and the pulse voltage to 5V. Then we connect the function generator to the gate of the nmos, so that it can control the on and off of the nmos. After that, we connect the 12V voltage source to the drain of nmos, and finally connect both ends of the resistor to the output. We get that when the frequency is 0, the output is 7.7V. When the frequency is 120KHz, the output is 3.47V. This is different from our simulated result of 6V. We think this is caused by the resistance of the component. Especially the influence of the internal resistance of the power supply and the internal resistance of the sensor is most obvious. When we turn the frequency dial, the voltage will change between 7.7V and 0V, so that we can make the machine not work at 3.47V and destroy the circuit.

We use a computer to connect to the Arduino controller, and then connect a human infrared sensor to it. Finally set the code. When someone approaches or touches this sensor, it will trigger the Arduino controller, so that the Arduino will judge and give a signal through the code. The output signal connection of Arduino is the gate of nmos, so we can use the computer to control the circuit. Our experimental results are as follows: When no one touches the sensor, the output voltage is close to 3V. I think this is lower than the expected 6V. This may be due to the resistance of the sensor, the resistance of the Arduino itself and the components The cause of resistance. When we touch the sensor, the output voltage will quickly become about 0.8V, but it will quickly return to 3V. We think this is related to the speed of the Arduino code and the response of the Arduion controller. But this way can make the machine stop working intermittently, thus destroying the machine's components. In addition, we can also control the frequency of the signal given by the Arduino through our code, which will destroy the circuit similar to the first way.

=Feedback and Thinking=

Our experimental results and simulation results are somewhat different. We believe that the main reason for the different results is that we did not consider the internal resistance of each component when we simulated the circuit, and all our simulated results are consistent with the calculated results. If we add the internal resistance of each link when we start the simulation, I think the simulated result will be closer to the real result. Although our experimental results and simulation results have errors, they are all within the requirements. This is very important.

=Circuit components Specifications=



Components

=Project Learning=

Research Information

=Final Design=

The IR Kill Circuit works remotely to cause the output voltage of the Boost Converter to drop significantly below the input voltage level, rendering the circuit inoperable.



The IR Sensor detects an objects within its range, causing the output voltage of the Buck Converter to drop significantly below the designed operational voltage level.


 * 1) We used LT spice to simulate the buck converter circuit and boost converter circuit.
 * 2) We used Mathcad to figure out the size of the capacitors and inductors, and how much voltage they can withstand.
 * 3) We buy all the components and build the circuits based on the calculated data.
 * 4) We began to connect the circuit to the oscilloscope and other tools.
 * 5) We debug the circuits.

Videos

=Validation=

Design Validation
We need to get our design validation from our mentor, Dr. Herbert Hess's endorsement. In order for our project to be successful, we need to achieve the approval of the mentor.

Client Validation
Our Mentor, Dr. Herbert Hess has been giving us a lot of feedback this year. We worked with him to understand the needs of the circuit and how our project met those needs. Our team has been working closely with our mentors to ensure that our projects are always aligned with their needs so that there is a shorter feedback loop if anything goes wrong. According to Dr. Herbert Hess's feedback, our project generally meets the requirements of the facility, so our project has been approved by the mentor.

=Team Members= Necessary explanation：We are a real team, no matter which member of the team is in trouble, we will do our best to help him. So the responsibility has been given is the primary responsibility, does not mean that we have only these responsibilities.

=Additional Documentation=

Project Schedule

Gantt Chart

Meeting Agendas

Meeting Agendas

Meeting Minutes

Meeting Minutes

Presentations

Presentations

Design Review

Design Review

Budget

Budget

Snap shot

Snap shot