The date is: Fri Jan 29 17:29:46 MST 1999 Week of January 25th thru January 29 I spent this week getting organized with my new group. We have been discussing different processor and sensor options. My tasks for next week will be to investigate different sensor options which include the sensors provided along with some discrete IR pairs. Sorry I am late with this it slipped my mind until a few moments ago. -Tim Miller The date is: Fri Feb 5 16:10:08 MST 1999 Week of 2-1-1999 to 2-5-1999 This week I investigate two sensor circuits. One that use light feedback in an opamp circuit and the other using the IR receiver provided to us. I had some trouble with the light feedback amplifier. I was able to find the varying signal on the IR receiver. It was very small and will need to be amplified. I plan on trying a variation on the light feedback amplifier and providing an amplifier for the IR receiver. I also plan on preparing for the design review. Tim Miller The date is: Sat Feb 13 16:45:35 MST 1999 Week of February 8th thru February 12th This week I worked on the light feedback amplifier with little success. I believe that the 411 opamp has too low of an input impedance and is loading the photo diode that I am working with. I tried a different opamp that has a higher input impedance with some better results. But, after reviewing the data on the GP2D12 distance sensor, I believe it may work much better than the light feed back amplifier. We have decided to get two more of the sharp sensors. Consequently, I have decided to cease working on the light feedback amplifier I the interest of time. We are also hooking up an amplifier to the integrator out put of the IR receiver to use it as a backup incase the Sharp sensors do not work as expected. We will probably use the IR receiver as a white line sensor. Our team discussed the over all design and will be getting together a few times in the next week to prepare for the presentation. For the week following the presentation, I plan on investigate the use of IR motion sensor parts for the flame detector. Again my apologies for getting this in late. I plan on submitting this report on Thursdays from now on since I am working in Albuquerque on Fridays and tend to forget to get it before leaving work. Tim Miller The date is: Tue Feb 23 14:33:03 MST 1999 Week of February 15th thru February 19th I spent this week preparing for and giving our design review presentation. For the next week I will investigate the use of an IR motion sensor for the flame detector. My apologies for forgetting my weekly report. I have no excuse for this time Tim Miller The date is: Fri Feb 26 10:17:34 MST 1999 Week of February 21st thru February 26th This week I revisited the light feedback sensor because I noticed that I had forgotten to connect up the negative supply to the op-amp. Having done this, the circuit functioned as expected. I discovered that it could sense distances up to about 1.5 inches. This distance is too small for our needs as a distance sensor. I then tested the circuit to see if it can function as a white line detector. If we mount the sensor close to the floor, it should work well as a white line detector as its output changes a few volts when going from a black surface to a white surface. I also disassembled a commercial motion sensor and began to research the sensing element that it uses. The part that does the sensing is a pyro-detector made by Heimann. The company is an European company, and I could not find a website for them. I was, however, able to locate one of their American distributors of which there was no website listed either. I did get a phone number and will be requesting information. I also did some reverse engineering to determine how to hook up the sensor. I believe I have a good idea as to what the terminals are doing, but I want to do some more research before applying power. Next week I plan on continuing to work on the flame detector using the sensor discussed above with or without documentation. I have also made contact with the people who are working on an old project that I started. This project consisted of a PIC controller, motion sensor, and an ultrasonic sensor. Consequently, it can be programmed to do ranging and fire detection for us. I wiil be seeing what donations I can get from the project for use on our robot. Tim Miller The date is: Fri Mar 5 19:50:48 MST 1999 This week I made contact with the people that are working on an old project that I got started; and, they are getting me some of the hardware. I will be getting an ultrasonic ranging device that is the same sub assembly used in Polaroid cameras plus some of the circuit boards that I designed that use PIC micro controller to operate ultrasonic detectors in a slightly different manner than the Polaroid circuit does. It also has a motion sensor circuit built on it. I may or may not use this board. If my memory serves me, I believe that the Polaroid circuit has a minimum distance of 1 or 2 feet. If this is the case I may want to use the board because shorter distance cane be realized. I also worked on the motion sensor idea and found that it senses black body motion but doesn't put out a DC voltage when there is heat to be sensed i.e. it senses the change in heat not the presence or absence of it. Thus, it has an AC like response. For next week I will to work it with some large area photo diodes and light filters in an effort to sense a candle. Tim Miller The date is: Fri Mar 12 17:19:50 MST 1999 This week I worked on using a large area photo detector to detect the light from a candle. The circuit that I used actually is saturated with out a filter in front of the detector. I used a portion of floppy disk as an optical filter. The circuit shows no output with the optical filter. In the presence of a candle, the output shows a couple of volts change at about 8 inches or so and will saturate at 4 inches. For next week, I will be adding a multi-gain amplifier so that the candle can be detected at larger distances and than the gain can be reduced when the robot is near the candle. Since we are using a single sided supply, the outputs from the amplifiers are only using half the range of the A to D. Consequently, I am considering a level shifter design using a summing amp with some gain. I will be doing some SPICE analysis on these circuits this week. Tim Miller The date is: Fri Apr 2 16:31:18 MST 1999 During spring break and last week I worked on a multi-gain amplifier circuit the amplifies a small voltage change that is riding on DC level i.e. the voltage change from 2.5 volts to 3 volts. The circuit was designed to transform the follow voltages as indicated: 2.5 to 3.0 volts, 2.5 to 3.75 volts, and 2.5 to 5.0 volts all transformed to 5 to 0 volts. This was done by using an inverting summing amplifier with a specific voltage reference connected to the non-inverting pin of the op-amp. The formula was derived, with the help of Dr. Wedeward, and adapted so that the correct reference voltages could be used. Then the circuit was analyzed using Pspice. The voltage and gains were derived using parallel combinations of resistors to minimize the number of bilateral switches need to change the gain. A short Matlab program was written so that different combinations of resistors could be picked such that commercial resistors could be used with the minimum amount of error. When the circuit was wired up it functioned as expected except for the low voltage instead of being 2.5 volts was close to 3.0 volts depending on ambient light. Consequently, a potentiometer was exchanged for one of the resistors so that adjustments can be made due to environmental conditions. More design would be necessary if this circuit was to be mass-produced but for our purposes, it should work fine. I also worked with the PIC microcontroller that we will be using. I was able to access the A to D and have the result put out on one of the ports. This week, I was to get the PWM circuit working taking updates from the A to D. We connected the PWM outputs from the PIC up to our H-Bridge board and the motors ran perfectly. On Thursday, I began to collect frequency data from the encoders when the motor was running at different speeds. I decide to log readings for every 50 uSec change in motor on time. When I was finished with the first motor and was hooking up the second motor, the positive power supply lead must have touched the output lead of the encoders causing them to fail. I was able to get replacement motors from Dr. Wedeward; and, I will be collecting data from them next week. I also plan on starting the code for closed loop control using two of the counters in the PIC to provide feedback about motor speed. Tim Miller The date is: Fri Apr 30 15:51:32 MDT 1999 Week of April 26 to April 30, 1998 This week we prepared for the presentation on May1, 1998. Last week I tested the part that I had programmed with the closed loop routine. It worked but it didn't appear to be regulating the speed as well as I expected. The next step is to incorporate an integral type of controller. I will also provided closed loop feedback to the distance sensors in preparation for wall following. Unfortunately, we are not far enough along to enter in the competition on Saturday. Since, I have been working on the presentation, I don't have that much to report this week. Tim Miller The date is: Thu May 6 15:35:39 MDT 1999 I worked on implementation of proportional and integral controller. I found that adding the integral controller helped in speed regulation. It was, however, more susceptible to overflow errors in the calculations. I tried both algorithms, proportional only and proportional with integral, under load conditions and it appears that proportional may be adequate. In an effort to work on the wall following program, I decided to work on the overflow errors only if they occur in the wall following program. We made a last ditch effort to get wall following accomplished by Thursday. Our hardware is very close to being completely connected. The wall following program is ready to be tested in the simulator for initial debugging and then tested on the robot. We are very close to having a finished product. With another 2 to 3 weeks, we would have completely functioning robot. My plans now are to document what we have done in the final report with my group and make suggestions on future directions. Tim Miller