At a Glance

I provided the complete hardware and firmware designs for these Heads Up Technologies avionics products in 1991.  Among the things I provided were:

• Schematic-diagrams and layouts for 7 circuit boards (plus two alternate circuit-boards that never went into production).
• Firmware and MS-DOS-based support software:  a large but unknown quantity of C and Intel-8051 assembly-language code.
• Documentation:   40 pages of documentation for specification, software-certification, and internal engineering purposes, plus large but unknown quantities for hardware and software documentation purposes.

Background

These products were intended to be improved versions of the earlier Heads Up Cabin Briefer product.  Among these improvements were the following:

• Audio-playback quality was to be improved, and maintenance of the audio databases simplified.
• Royalties were to be reduced by replacing the speech-compression algorithm used with a proprietary algorithm.
• The user-interface was to be made more friendly by adding an LED-based dot-matrix display.
• The capacity of the unit was to be increased (i.e., more minutes of audio playback were to be provided).
• The unit was to be field-reprogrammable, rather than configurable only at the factory.
• Multiple languages were to be supported, so that the unit could contain briefings in several languages, with the actual languages played back being selected only at flight time.


Accomplishing these various objectives was not merely a matter of electrical design and of coding, because several significant technical hurdles had to be overcome that affected not only these products, but also many other product designs for years to come.  In fact, the hardware, firmware, and support software are not repackaged Heads Up Cabin Briefer items, but are complete replacements. Here are some of these technical highlights:
 

• I invented a completely new audio-compression algorithm.  This turned out to be a lossless 8-bit audio-compression, in which raw 8-bit samples were typically reduced to about 5 bits.  This algorithm was used at Heads Up for many products for many years, and at this writing (5/2001) is still used in the majority of products delivered by Heads Up (though not in new designs).
• These were the first designs at Heads Up in which multiple processors were used within a single unit.  Some versions had 3 processors, though most delivered units had just 2 processors.  (There were several iterations of the design, based on differing memory technologies.)  Communication among processors took place by means of protocols I created.
• Field-reprogrammability of the units was accomplished by connecting the units to PCs through standard printer ports.  This meant that additional protocols had to be created for this communications channel, and that PC software suitable for use by customers (rather than merely suitable for use by trained Heads Up personnel) had to be created.


While I don't claim to be equally satisfied with the solutions to each of these problems, a significant amount of solid work was certainly accomplished.  The field-reprogrammability was the most unsatisfactory point, as it proved vulnerable to incompatible configurations of the numerous PCs used by the customers and to fragility of the protocol.  Nor was it very user-friendly. On the other hand, though, the field-reprogramming software evolved into a general-purpose remote-diagnostic program that (in spite of the kinds of flaws mentioned) provided important services in quite a few additional products, and within Heads Up is runs continuously on several computers.  Here's a screen-shot of a recent version of the diagnostics program: