![[Screenshot of Demo Program]](demoscreen.gif)
Figure 1: Screenshot of Demo Program
| edurant.com > MSOE > SEED > Feasibility > Report |
For a senior design project, research was performed to investigate the feasibility of a digitally controlled power supply. Ideation methods for projects, criteria, options and solutions are considered. After developing a system block diagram, research was performed on critical solution aspects, including power supplies, input and output devices and microcontrollers. A cost analysis predicted a cost of approximately $36 000. Finally, an implementation timetable was developed with an estimated completion date of 30 May 1997.
| To: | Dr. Henry Welch |
| Date: | Monday 18 November 1996 |
| By: |
|
One facet of this senior design course sequence is choosing an appropriate project to implement and presenting a feasibility study. The fall quarter is reserved for investigating different ideation methods that are used while performing the feasibility study. Team ß spent from 9 September 1996 through 17 November 1996 using several different ideation approaches to select the final overall project design, a digitally controlled power supply. Included in this report is a broad overview of the ideation process to narrow the selection to one method of implementation through modular subsystems. Also, various components of the system were researched. Cost analysis was performed to get an overall idea as to the total cost involved in planning, designing and building a project of this scope.
Before embarking on detailed ideation and research, the purpose and scope of the project had to be defined.
A common problem in undergraduate engineering laboratories is accurately setting and measuring power supply effects. This is due to analog controls that are subject to drift and aging, making them difficult to fine tune. A power supply should let the user enter parameters on a keypad, much as modern oscilloscopes provide discrete control via keyed input.
Therefore, team ß chose a digitally controlled power supply for their senior design project. This power supply is intended for laboratory use. Its main advantage is allowing the user to control the voltage via a keypad. In addition, this project is an excellent choice since it is scalable. For example, parameters may be specified (i.e. RS, current, waveshape) and a PC data collection interface is possible.
Some potential features follow.
Positive aspects of the project include the following.
Potential problems with the project include the following.
The interface philosophy encourages a modular (and hence scalable) design.
The uses of these supplies will vary as additional options are added. With the basic unit having three fixed supplies, two variable supplies and a variable tracking supply, it will fill the primary needs in academia and, in a limited fashion, industry.
Hewlett-Packard has made supplies that are controlled via the HPIB (Hewlett-Packard Interface Bus) bus (computer / digital control) for many years, but these units cost thousands of dollars and deliver higher amperage than is required by most student laboratories. Team ß's design is intended for student use, but it may also find limited use in industrial applications. Some of the aspects making it ideal for the student are low current (less than 1A), front panel digital control and low cost.
Here the basic unit with its fixed supplies will provide the needed voltages for TTL circuits and operational amplifier circuits. The variable supplies can be used for designs where voltages other than the fixed supply voltages are needed such as those using BJTs and FETs. Adding more supply banks will allow for larger design applications with either independent supplies or additional variable supplies. Some of the additional options that the academic labs may use include data recording, variable AC supplies, programmable functions such as brownout, ramping voltages, or power on-off cycling, noise injection, and constant current sources.
The simplicity of use and the accuracy of setting the desired output levels will make this supply easy to use for both the advanced student and the beginning student. Little care will need to be taken to ensure that the output level will remain constant throughout the experiment. The ability to log data directly to a PC could automate recording all parameters measured by the supply. This would help eliminate missing data when the student writes the lab report.
The same functionality as obtained in the engineering labs would still be available. The use of Kelvin sensing will allow these supplies to be used in applications where the voltage is critical and the IR drop of the leads needs to be compensated for. Options like a constant current source and data logging would enhance the results of many labs by eliminating the human error associated with taking repetitive measurements.
The same functionality mentioned above will be available to industry. The primary use would be in design facilities and labs that would normally use these voltage levels. Possible uses in an industrial laboratory situation include: IEEE Standard 488/HPIB bus interface and higher current modules for both fixed and variable supplies.
The design team used various ideation techniques to develop project options. Brainstorming, trigger word, and a variant of the fishbone diagram were the three main ideation methods used to develop the project concept.
The group began by brainstorming problems to be solved and projects to be designed. These ideas were then evaluated on personal interest and an overall concept of the group's ability to produce a solution to the problem within the allotted time. The idea agreed upon was to design a DC power supply that would be easy to set, and which would monitor itself to keep it at the desired settings.
From this information a list of trigger words was developed. This list described the major requirements of the power supply. Then these words were used to develop lists of methods and components that could be used to implement the requirements.
This list was then used in a fashion similar to a fishbone diagram. Three different solutions were developed from this information: the cheapest, the easiest to use, and the most versatile.
Another brainstorming process, besides trigger word, was used to develop a features list. These features were then categorized as needs and wants of various degrees of importance, with needs being crucial and wants ranging from important to fanciful. These needs and wants were then compared with the three solutions from the fishbone selection process, and a new, final concept was developed. This concept defines the required goals of the power supply as well as some of the ideas from the wants list which could be incorporated into the power supply as development time permits.
The group had four final ideas for the project from which to choose. They were:
The power supply was chosen based on group member competencies and the given time frame.
The team used brainstorming to define seven key requirements for a solution to the digital power supply problem. Then, they listed the following options for each of these requirements. The nature of the options dictated that several were feasible for certain requirements.
This leads to 903 168 combinations. A modified fishbone diagram was chosen to manage this large Cartesian product of solutions.
The team decided to examine three conflicting product objectives independently, noting which options applied to each. Some compromise between these extreme objectives will lead to a desirable solution. The conflicting objectives with their relevant options for each criterion are shown in Table 1.
| Overall Flexibility | Simplicity for User | Cheap |
|---|---|---|
|
|
|
Based on additional discussion of these conflicting objectives, the team chose to emphasize a simple-to-use product. Since the project attempts to provide an alternative to conventional power supplies, flexibility was a close second. Focusing on these objectives, features were classified into needs and various levels of wants (see Table 2).
| Primary | Secondary | Tertiary |
|---|---|---|
|
|
|
Based on the wants and needs and the options which came from the modified fishbone diagram, the team settled on the major portions of the design. Table 3 summarizes the options and features (the most probable optional features are indicated by an *) that are planned for implementation in the winter and spring quarters. These will be implemented according to the project plan in Appendix G. The most probable optional features will be implemented if time permits.
| Option | Rel. Crit. (see Reqs. and Opts.) | Relevant Objective(s) | Deciding Factor(s) |
|---|---|---|---|
| Keypad input | 1 |
Simple |
|
| Isolated ground | 2 |
Simple |
|
| Load monitoring (*) | 3 |
Flexible |
|
| Cooling fan (*) | 3 |
Flexible |
|
| Circuit breakers | 3 |
Simple |
|
| Keyboard output hardware (*) | 4 |
Flexible |
|
| Panic button | 4 |
Simple |
|
| Modular fixed +5V, 12V, 15V(*) supplies | 4 |
Simple |
|
| LCD readouts | 4 |
Simple |
|
| Internal meters | 5 |
Flexible |
|
| AC powered | 5 |
Simple, cheap |
|
| Rack mounted (*) | 6 |
Flexible |
|
| Microcontroller | 7 |
Flexible, simple |
|
Further research concerning the implementation of the feature list of Table 3 will take place in the winter quarter. As some of the major blocks are settled on, circuit design and testing may also take place in the winter quarter. This will allow for redesign if necessary before too many of the blocks are locked in, and allow for establishing the inter-connectivity of all blocks before the design and building phases are completed.
A patent search was conducted to compare the project's goals with existing products. Also, various components of the system, including microprocessors, power supplies and I/O devices, were researched.
Patent research early in the development stage answers many questions. Perhaps the most important is, "Has somebody already done this?" Additionally, patents on similar ideas indicate how unique the idea is and thus give a rough indication of patentability. Further, such patents often include a reference list which is helpful in other areas of project research. Also, patents give ideas as to how to implement various parts of the project.
The eleven patents shown in Table 4: Patents Researched were researched. (The United States Patent Office's searchable database on the WWW (http://www.uspto.gov/) significantly reduced the time required compared to a traditional patent search and allowed more options to be found due to extensive cross-referencing.) They mainly concern power monitoring and regulation. Very little was found about user-friendly power supply interfaces. Two especially relevant patents are included in Appendix C. Patent 5 519 634 ("Data transfer unit and method of power supply to the data transfer unit") was one of the few patents to include a computer interface and a serial communications protocol. Patent 5 103 110 ("Programmable Power Supply") was especially interesting since it covered monitoring and control of both voltages and currents. Also it specified multiple output channels and a human readable output as team ß's project does. Additionally, it had extensive references. It was interesting that none of the patents researched emphasized a user friendly input method.
Several processor options were considered. The benefits and disadvantages of each processor are summarized in Table 5. Additional processor research is included in Appendix D.
| Company | Chip | Pros | Cons |
|---|---|---|---|
| Intel | 8086, 8088 |
|
|
| Motorola | 68HC11 |
|
|
| 68HC12 |
|
|
|
| 68HC16 |
|
|
|
| Parallax | BASIC Stamps |
|
|
| PIC |
|
|
Although the BASIC stamps are appealing for rapid development, their limited memory space essentially eliminate them from consideration. The 68HC12 family, with its greater A/D and D/A resolution is a leading option, but would increase cost considerably. The 68HC11 family is a close second. Despite the 68HC11s inferior characteristics, the team is very familiar with its operation and development hardware and software are already available.
The best option to date is a $119 power supply listed in a Semiconductors for Design Electronics catalog (Spring 1996, #131) which has 8-bit digital control on 10V and 40V scales at 1A. With these scales, however, a least 10 bits are required to reach 3 SF accuracy, so more options are being investigated.
Switching power supply ICs from National Semiconductor are also being researched. This approach would require construction of the supplies from the ground up. The experience gained by actually designing and constructing these blocks needed for the project would outweigh the cost in both dollars and time that could be saved by purchasing pre-built functional blocks for the supplies.
Another approach is to use discrete components such as transistors and power FETs, and yet another is to use IC regulators. The choice of IC regulators is virtually endless; they include three terminal shunt devices and switching technology ICs. Some of the ICs being considered are listed in Table 6. National Semiconductor data cover sheets for these regulators are included in Appendix E.
| Part Number | Description |
|---|---|
| LM338 | 5 A adjustable regulator |
| LM317 | 1.5 A adjustable regulator |
| LH1605(C) | 5 A high efficiency switching regulator |
| LM78S40 | Switching regulator |
| LM125 | Dual tracking regulator |
| LM78xx & LM79xx | 3 terminal fixed regulators |
A combination of shunt and switching technologies may also be used to implement the project. Further study of each technology is needed to determine the best method for each section. Before a technology is selected, the following issues need to be addressed:
Library research and manufacturer literature will need to be obtained on all of these potential problem areas and evaluated to determine which technology is best for the individual tasks.
This section discusses various options for both input and output devices. Relevant parts catalog literature is included in Appendix F.
The choice of input devices is critical to the project. These need to be complete, simple and ergonomic. This is because they are the part of the power supply with which the user will interact. Therefore, the input mechanism should be intuitive and attractive. Buttons, switches, and dials should all feel responsive, and be laid out well and comfortable to work with.
The input devices will be required to take voltage and current and possibly time information from the user. This should require only numeric input and a few special inputs, such as "go," "clear," "stop," etc.
Several input options are available. These include keyboard, keypad, thumbwheels, mouse, and touch screen. The keypad was selected since it has an easy to use interface, is well-suited to the input information, and does not require an expensive monitor or touch screen. One or more switches will augment the keypad to switch the unit or the individual outputs on and off.
The output devices for the power supply are as critical as the input devices. The output devices must provide accurate real-time information about the voltage and current being supplied, as well as information about the values being input to the supply.
Possible output devices include an RS232 port, an LCD matrix display, an LED display, a monitor, analog meters, digital meters, speakers, and lights. Several of these options will be used. An LCD display will be used as a feedback to the user as (s)he inputs data, as well as a display of what voltage or current is being output by the system. Digital meters will be used to provide a real-time indication of the output voltages and currents from the power supply. An audio alarm will indicate temperatures or currents over safe levels. An RS232 port may be installed to send data to external devices for analysis. Analog meters will only be included if digital meters are found to be inadequate. LEDs or additional illuminated switches will probably be used to indicate the "on" state of the unit and the outputs.
This combination of different output devices will provide all the necessary information from the power supply to the user. It provides for flexibility and thoroughness in information delivery, and does not require any external devices, such as a monitor or an expensive display screen.
A Microsoft Visual Basic 4.0 prototype interface (see Figure 1) will allow refinement of the user interface before difficult-to-change design decisions are made. A sample Win32 application is on the included disk in Appendix I. This application implements the primary goals of the project yet allows additional functionality to be added and tested prior to design and implementation of these additions. Appendix H lists the code used to implement the application.
Figure 1: Screenshot of Demo Program
The chosen options from the left column of Table 3: Final Option Selection are shown in the block diagram (Figure 2).
Labor is the major cost concern of the project. An estimate of five project hours per week per team member, with each hour being worth approximately $50, was determined for the four man team. This estimate took into account that while much of the time will be spent on engineering tasks, some of the time will include clerical, technician, and consulting tasks. These affect the hourly rate. The team chose $50 per hour because, using a ratio of fifty percent salary and fifty percent benefits and facilities, the numbers averaged out to $50 000 per employee year. This number is perhaps five to ten thousand dollars high. However, with the potential consulting hours and other various tasks, the team feels this is a justified estimate. The total estimated labor cost is $30 000, assuming 600 total hours (5 hours/person-week times 4 people times 10 weeks/quarter times 3 quarters) on the project. Adding a 15% safety margin, this becomes $34 500.
The cost of development parts is difficult to estimate. A wide range of choices have to be made, such as, "What type of keypad is needed for the project? 4 button, 16 button, conductive rubber contact, environmentally sealed, etc." The choices for each of the other large pieces of hardware vary also. Table 7 shows estimates for the cost range of each of these major parts, as well as a category for additional small parts.
| Cost | |||
|---|---|---|---|
| Component | Low | High | Most Probable |
| Display Device | $15 | $50 | $30 |
| Meter Display (2) | $80 | $120 | $100 |
| Keypad | $13 | $75 | $23 |
| Controller/Processor | $100 | $750 | $300 |
| Power Supply | $40 | $150 | $100 |
| Other Parts | $100 | $300 | $200 |
| Totals | $348 | $1 445 | $753 |
| Buffered Total | $866 | ||
The team is confident that through research and development they will find significant ways to reduce the parts cost on the production run. For example, the purchase of evaluation hardware for the controller/processor is a one time cost. Also, an ASIC design could be developed from the initial prototype which would be more cost-effective in the long term. Unfortunately, the time frame of the design will not allow an ASIC to be designed. There will also be bulk pricing benefits for the production units.
This analysis shows a projected cost of $35 366 for the research and development of this product. This must be absorbed by the production units in order to pay back the product investment. The target sales group is colleges, universities, and laboratories which do significant electronic circuit testing work.
Marketing estimates indicate that the average university or college would want to fill one or two student labs with these power supplies. These labs would consist of about 10 units each. The standard corporate customer is estimated to want two units per laboratory. Table 8 indicates the breakdown of research and development costs for various production levels.
| Units/year | |||
|---|---|---|---|
| Years | 100 | 1 000 | 10 000 |
| 1 | $353.66 | $35.37 | $3.54 |
| 2 | 176.83 | 17.68 | 1.77 |
| 3 | 117.89 | 11.79 | 1.18 |
| 4 | 88.42 | 8.84 | 0.88 |
| 5 | 70.73 | 7.07 | 0.71 |
If the product were produced for internal use only, the cost estimate of research and development would be $1 768.30 per unit, assuming 20 units.
At this time production parts and production labor costs cannot be estimated, since an assembly process is outside of the project scope.
Appendix F includes relevant pages from various electronic parts catalogs. These pages are a sampling of the different parts which will need to be purchased.
A working list of tasks for the project plan includes:
Progress is tracked using Microsoft Project® v4.1 (see Appendix G). This will facilitate updating and refining the project plan. After entering the appropriate data, Project keeps track of dependencies and durations. Although Microsoft does not currently provide an automated method of posting Project data to a Website, options are being investigated for presenting this information on the team's Web page (see Appendix B).
From the project plan, the critical path time will be 24 weeks from the start of the winter quarter to the SEED show. As the project matures, the critical path time may be shortened due to tasks being completed ahead of schedule.
The power supply project was chosen as the design to implement. This supply will have discrete input, digital control, and a high level of versatility. The research and development through the prototyping stage should cost $35 366, with the majority of the cost from labor. This project should be completed by 30 May 1997.
Printouts of Webpages from:
Printouts of PDF documents from National Semiconductor. See the section about Regulator ICs for details.
Not available on-line.
Not yet available on-line.
| Filename | Format | Purpose |
|---|---|---|
| Feasibility Report.doc | Microsoft Word v7.0 | This document |
| Power Supply Demo.exe | Win32 Executable | Demo interface |
| Timetable.mpp | Microsoft Project v4.1 | Gantt chart |
![[Block Diagram]](block.gif)