Tuesday, January 29, 2013

Adjustable Light Glove, Part 2 - Prototype Circuit v2.2

LinkBack => Part 1 - Concept, and Proof-of-Concept <Blogspot>

Previously, I discussed the beginning hurdles in my attempt to to create an intuitive, hand-mounted flashlight. Since then, I've worked to refine the circuit for the first prototype which uses the stretch sensor.

Image: Prototype circuit v2.2 - This uses a simplified charging solution for the 
555, a voltage divider to provide 2.4V on the Control pin, and
the correct values for the stretch sensor network.

 
Video: Prototype circuit v2.2 - This uses a 555-based PWM and a 0.75" section 
of stretch sensor to control the light output of a Luxeon 1W LED.

Below I'll discuss some of the challenges and decisions I've made on this leg of the process. The next steps are to test the LM3405 LED Constant Current Buck Regulator to see if it will meet my needs more thoroughly than this test-bed style setup can. After that, it's time to design, etch, and populate a PCB for the first full device prototype.


Trials and Tribulations

A major point of concern was ensuring that the circuit was built so that it could be modified easily to accommodate various lengths of stretch sensor (and these changes could be determined easily). While working with the circuit, the main problems I came across were:
  1. LED Light 'strobing' 
  2. Tuning the circuit to allow the maximum duty cycle range, with emphasis placed on the low end (ie, it was more important that the circuit reach 0% duty cycle than 100%)

 Strobing

The first problem, strobing, was due to the frequency of the 555 system being too low. I couldn't find formula detailing the frequency expected when using the transistor charging circuit, and I didn't want to base the solution on trial and error. Some research led me to decide that the transistor portion was unnecessary (for this purpose, I don't need a perfect ramp - the sawtooth capacitor charge-discharge pattern is sufficient). Replacing it with the basic astable 555 circuit (two resistors and a capacitor) simplified the frequency calculation. The formula used was: Freq = 1.44 / ((R1+2R2)*C)

Maths were called for, so I plugged in some various combinations into an Excel doc I made for the circuit. I decided on a combination to give me a frequency of 444 Hz - changes made later, specifically using the 555's Control pin, ended up raising this to 750 Hz. Updated the proto board, and no more flicker.

Duty Cycle

Using the aforementioned Excel spreadsheet, I tried various experimental combinations of resistor values for the stretch sensor sub-circuit. This data showed me that with my length of stretch sensor, while I can control the actual Vout,max and Vout,min values, the difference in their values is more or less fixed at 1.2V (in the range of values I'm considering, 360ohm to 1600ohm). The two options that jumped out at me were to work with the 555 sub-circuit, or to place something between the stretch sensor network and the comparator to increase the network's voltage range. I opted to work with the 555 sub-circuit to make that 1.2V gap work.Since the current gap of 1.67V is too large, it needed to be changed. Since the 555 needs at least 4.5V to operate, I couldn't lower its input voltage low enough to achieve the 1.2V spacing, so I opted to make use of its Control pin. By applying a voltage to CON, the max voltage of the sawtooth signal is set to the voltage at CON, while the min sawtooth voltage is 1/2 the CON voltage. By choosing a value of 2.4V at CON, I gain the 1.2V gap I'm looking for. For simplicities sake (and easy testing), I'm using a simple voltage divider to achieve the 2.4V reference, as I had the parts already.

I used the spreadsheet to find a value of resistor for the stretch network to give me a Vout range of 1.26V to 2.42V (compared to the 555's sawtooth of 1.2V to 2.4V). Since the stretch Vmax exceeds the sawtooth Vmax, the circuit achieves the goal of 0% duty cycle (with a max around 92%).

Tuesday, January 15, 2013

Adjustable Light Glove Part 1 - Concept, and Proof-of-Concept

LinkBack => Part 2 - Prototype Circuit v2.2 <Blogspot>

A project I've dabbled with on and off deals with using the degree of 'openness' of ones hand to vary the light output of an LED, to create a simple to control, hand-mounted flashlight.


Video: A proof of concept test using a high-brightness (< 1W) LED.

I worked extensively within NI MultiSim to design and test the circuit before purchasing the components. The proof of concept circuit used a standard rotary potentiometer as a voltage divider to drive the gate of a transistor, allowing current to flow and powering the LED. Later prototype circuits will utilize a flexible stretch sensor, like this one.
 

A Change of Plan

One of the design requirements is that the light should go from off to max within the standard range of motion of opening the hand. This was simple with a basic rotary potentiometer, but became complicated when moving to a circuit using the stretch potentiometer.

While most potentiometers have a minimum value near zero ohms, the stretch potentiometer's minimum value is around 10k ohm for a 12" segment (with a max around 40k ohm) - this made it highly impractical to develop a simple resistor-divider based input stage to the driving transistor.
 

After going through many test circuits and iterations of component values and types, I determined that the simple divider wouldn't allow the full brightness range in the high power (> 1W) LED without major sacrifices: either the LED wouldn't reach an 'off' value, it wouldn't fall quite short of max brightness, or it would waste a lot of power as heat from the limiting resistor.

Simple PWM

I decided to switch tactics, and focus on a PWM based system, utilizing a 555 timer oscillating in astable mode. This circuit worked very well in simulation and solved a lot of the problems I was having.

The first full test circuit uses 4 major circuit subsystems: the 555 timer, a voltage divider using the stretch potentiometer, a comparator chip, and the MOSFET LED driver. The voltage from the stretch pot system is compared to a sawtooth signal resulting from the 555's charge-discharge cycle, and the resulting PWM signal is sent to an n-channel enhancement mode MOSFET which drives the LED. By tuning the resistor values in the voltage divider
, any length of stretch potentiometer can be accommodated




 Image: The MultiSim schematic for the first full test circuit - 555 PWM.

Parts have been ordered for the 555 PWM circuit, and assembly and test can begin soon. I have also ordered parts for a potentially improved version utilizing an LED driver IC, the LM3405.

Sunday, October 16, 2011

EIT Medical Imaging Device (Senior Design Project)

For my senior design project at UT, our team set out to construct a medical imaging device prototype using Electrical Impedance Tomography. Our aim was to design a prototype which could be made compact, portable, and simple for a non-technical end user (such as a doctor) to operate. The results had to be fast (at or near real time), accurate enough to be used for diagnosis, and simple for the user to interpret. This last requirement is crucial for emerging medical technology, as often medical imaging results require years of instruction and experience to reliably be used for diagnosis.

http://utlive.pbworks.com/f/Hardware+Flow+Diagram.png


Sunday, September 11, 2011

Monitor-Circuit Interface

For a project, I needed a cheap, simple way to interface a circuit to my computer. The premise was a simple setup for a team-based control-point game. The idea was that code on a laptop computer would be used to keep track of king of the hill style 'control points', connecting to a display box which would illuminate different colors depending on which team currently controlled the point. Each point would have a laptop, and they would communicate with each other over the network to determine whether a team controlled all points and send out victory music if that occurred.

USB control would be ideal, but went against the 'keep it simple' idea behind the original circuit. Searching online for alternate control ideas turned up no useful information - most laptops do not have serial ports, and this would tend to use a micro controller anyway. The novel solution I came up with was to use photo-transistors to interface a simple circuit with a segment of the monitor on which the code would display a series of black and white boxes, with each box representing one bit of data. The circuitry would use this data to control different colored LEDs in the display box, thus indicating which team was currently in control of the area. Because of the simple nature of this setup, a box for each laptop could be made quickly and at little out of pocket cost.


Friday, September 9, 2011

Simon Says - Tilt! (Embedded Systems Final Project)

For this project, my partner and I implemented a modification on the game ‘Simon Says’, in which the system outputs a sequence of lights and sounds that the user must duplicate. In our version, this is accomplished by tilting the system in the correct sequence of directions and sampling an accelerometer once the capture button is pressed. Inputs include a switch for reset, a switch for tilt capture, and an accelerometer to detect tilt. Output consists of four LEDs and a speaker.

The microprocessor used for this application was the Freescale MC68HC11; it was chosen for the code similarity to the chip used in class (Motorola 9S12DP512), its wide range of features for its price point, and PLCC packaging which made PCB layout easier than smaller or more densely packed designs. The accelerometer used was the Analog Devices ADXL202 Dual-Axis Accelerometer. A 9V battery was used to make the game portable, but included the option of using a wall wort as well.


Shown below are the PCB layouts we designed for this project: