0.4 The bugs

Introduction

In the initial production run of the Speak N Spell there were several bugs that caused some difficulties but did not stop us from going into production. I classified them as bugs, rather than features, as they were problems that the end user had to live with, and we couldn't figure out how to make them look like features. There were other issues that were not obvious to the end user, these will be discussed in a later chapter. The bugs discussed in this chapter will cover all of the aspects of the product: Hardware, Software, and even the Plastic Case.

On/off switch

This bug was a system level hardware issue. It was the result of my brilliance which made it a bit personally embarrassing, but also made it relatively easy to fix without widespread recognition.

I had decided that an on/off switch was going to add too much cost to the systems so I devised a scheme that allowed for a “software” on and off key. It was simple. When the user pushed the ON key, a connection was made to the power supply module. As long as the ON key was pressed, the power supply would remain on. As the controller chip (TMC0271) powered on, it turned on one of the bit I/O lines that was attached to the same input as the ON key was attached, thus holding the power supply on after the ON key was released. To turn the Speak N Spell off, the controller chip only had to reset the bit I/O line. This reset could be accomplished either by pressing the OFF key or by the controller chip resetting the line after a period of inactivity.

This, of course, was brilliant as I saw it. We didn’t need to have an on/off switch and the product could power itself down if left unattended. Before you nod your head and think, "what is so brilliant about this?", you must remember that, at the time of the design, automatic power down circuits were not common, and, when available were more expensive than an on/off switch.

Unfortunately, there was an issue to the design. It showed its ugly head during power down. As the power was being turned off and the voltage rail was decaying from 10 volts to zero, there came a point where the chip-set became erratic, but still functioning even though it was below the minimum voltage specification. That meant that in some Speak N Spells, the controller chip demonstrated its erratic nature by setting the bit I/O line and turn the Speak N Spell back on. This could be classified as two different bugs: 1) the product wouldn't turn off, and 2) it only happened on some of the products, some of the time.

I didn’t have a back up plan, and I didn’t have anyone who I could consult with to resolve the issue. I was going to have to admit to the world that we had made a major break through in the state of the art of electronics, except it couldn’t turn off. Yes, on/off switch technology was relatively mature at the time.

As an aside, this was my first experience facing an impossible task and realizing I had no one who could help. The panic was overwhelming. To take this aside a bit further, the next time I faced an impossible task, I wasn’t nearly as panicked. The third time, I found it exciting. I now am not interested in doing the possible as I'll gladly leave the possible to practicing engineers. Unless it is impossible, I’m not interested – give it to someone else. I have gone from panic to excitement when faced with the impossible. To go a step further, I formulated a second definition of "impossible":