David Bishop
Home: 408-734-1891, Cell 408-209-9470
e-mail: bishopd@aol.com
website: www.bishopd.com
InkedIn: David Bishop
Printable Version

Test Engineer

RF and digital testing in the following areas:

These responsibilities were to reduce production time, resolve product problems, increase quality, meet delivery schedules, and create historical data to track product performance and part reliability.

Technical & Specialized Skills

Languages:
Web Services:
Platforms:
Test Systems:
Application:
LabView, C, C++, Basic, Fortran, Pearl
HTML, XHTML, CSS, JavaScript
Windows, Mac, UNIX
HP In-circuit, Zehntel Real Time, Watkins-Johnson (WJ) In-circuit
Word, Excel

Professional Experience

Test Engineer
ITT Corporation, Morgan Hill, CA

Built Excel spreadsheets to aid technicians in testing parts from a vendor. The spreadsheets set the test criteria, equipment and test methods used. The technician entered the measurements and the spreadsheet determine whether the device was in spec by flagging the out-of-bounds results. The spreadsheets handled both linear and non-linear tests providing both tabular results and plots.

Last task performed:

Trained my replacement in 2 weeks to understand, program, troubleshoot and repair a very complex RF digitizer board with built-in signal analysis. He really needed 8 weeks but wasn't available until the last 2 weeks before manufacturing moved to his location. Brought him up to speed on all the software and hardware of my other responsibilities as time allotted.

Test Engineer
EDO Reconnaissance and Surveillance Systems, Morgan Hill, CA

Most complex project:

Teamed to develop test, alignment and programming procedures for the latest high-speed digitizer board replacing an older model. With the loss of the designers, the teamed was composed of test engineers, the Project Engineer and the contractor who developed the control firmware for the digital portion of the digitizer. Prototypes of the digitizer and IF modules were used to develop all procedures. The documentation was created from reverse engineering the schematics, part manufacture data sheets and the proposal of how the digitizer was to work including the preliminary address mapping. Scores for a signal detected by the digitizer were based on the performance of the original IF modules and digitizer boards. A description of how the software created the scores was obtained. The IF module and digital processes that most affected the scores were determined. The digitizer board was composed mostly of the IF Module, a VLSI custom chip and 5 DSPs. The IF module pre-conditioned and digitized the IF signal. The VLSI chip controlled the IF Module, VME, Complex Math chips, determined initial signal characteristics and created the dual data paths to the DSPs. The main DSP was programmed as a micro-controller managing the VLSI, DSP data, internal testing and communication with main computer over the VME. The remaining 4 DSPs were split to handle data generated on opposite edges of the AtoD clock. Each pair of DSPs processed data in serial fashion, each handling a specific set of signal characteristics. To the programming software was added additional internal tests to help identify internal problem areas. Many modifications to improve reliability to the Windows/UNIX test software that simulated the main computer were also made.

Results:

Outside troubleshooting a broken board, programming and test time was reduced from 5 days to a couple of hours that could be performed by a technician. The logbooks and documentation made troubleshooting and training a lot faster. Troubleshooting still required a test engineer

Test Engineer
Condor Systems, Morgan Hill, CA

Projects and Tasks:

Pre-amplifier, antenna controller:

Rebuilt the controller, built a driver to test the motors separately from the unit, updated the schematics to reflect the current version of the hardware, added missing tests from the ATP, modified the ATP tests to reduce paper output and test time.

Digitizer unit:

Rebuilt the UNIX system used to aid in testing of the Digitizer. Unified the software, replaced missing scripts, simplified the interfaces, added schematics to explain all the interactions between the various CCAs and external equipment to speed up troubleshooting. Rewrote the ATP. Made it easier for the technicians to follow, added missing tests from the previous revision, added technical analysis to explain what the tests were doing and how to identify the cause of a failure.

Trained technicians in many other hardware and software repair areas.

Support Engineer
GTE Government Systems Corp., Mt. View, CA

Managed, programmed and ran purchased test systems: HP, Zehntel, WJ and several custom designs. Modified manual ATPs with better test methods and procedures. Assisted IT to solve computer user problems on Windows, Mac and UNIX platforms. Was loaned out to several departments to help solve problems: Drafting, CAD, assembly and others.

Most complex project:

Designed and built a medium frequency test station that reduced testing a system sub-unit by 400%.

The semi/fully automated test system tested the most complex unit used in the customer's system. The test hardware, software and equipment were integrated into a unified test station. The test station was designed to aid the technician in troubleshooting, aligning and testing the unit in real-time. Any affects the adjustments made appeared on the graphic display instantly so the technician could see what other areas of the unit were being affected. The software allowed the operator to retry nearly anything and/or force the unit to start working when it hanged. Because no standard piece of test equipment could meet all the test signal requirements, a test aid was designed and built to sync, align and control multiple pieces of test equipment for each type of test performed.

As an example, for pulse testing, a cavity tuned generator, white noise source, computer-controlled attenuators, external mixers, amplifiers and pulse generators were used. The multiple pulse modulators, amplifiers and mixers were needed to meet the required, on to off ratio and dynamic range test signal specifications. The test-aid synced the time between the modulator of the signal generator, external pulse generators and the demodulated pulse from the unit to blank the pulse counter at the required time to make a valid FAR (false alarm rate) measurement. The white noise source and attenuators were used to produce a known NPD (noise power density) level. The test-aid alignment and calibration procedures made sure that the pulse characteristics (rise/fall time, over/undershoot, droop, frequency and width) and RF signal characteristics (frequency, amplitude and noise level) were accurate to within 0.5 dB for all test conditions.

Results:

Test time was reduced from 8 - 12 weeks down to 2 weeks. The test station took far more data (1000+ points) than manual testing. The manual testing could not do several tests correctly. Several issues with the unit and between the unit, system, depot and shipboard tests were resolved. A new alignment procedure before the unit left for system installation was created since the system actually had 3 dB more noise than the unit specification was referenced to.

Universal test software developed:

The universal test software was built in a universal modular format that could be used by other programmers. They would only be required to develop a module for a specific test. The universal code handled; error reporting, printing, plotting, instrument testing, variable instrument driver loading and a unified instrument module interface so that the same measurement format was used for all instruments of the same type. The print drivers could handle, thermal, ink, laser and dot matrix printers for both table and graphic dumps. The calibration data was archived and was reusable as needed except for a formal ATP where calibration had to be done. The graphic modules handled 5 different types of plot, linear and non-linear or both. The programmer supplied the graphic modules with the test data, specification data and type of specification to test the data against. A specification could be given in either the Y-axis (norm), the X-axis (which was automatically translated onto the Y-axis) or dynamically (pass/fail was based on the behavior of the data).

Education

Bachelor's program
DeVry Institute, Phx, AZ
Major: Electronics

Extension Courses: C, C++, OOP Design, OOP Analysis and Modeling, and Business Accounting
Other courses: Language Design, Compiler Design, HTML, Imaging, Theoretical Physics and Mathematics, Security / Encryption Systems and Microprocessor Design


Home

Last updated: Aug 6, 2013

WXC