I’m working on some circuit designs and going through some of the components I have on hand. Part of the Great Winnowing in moving earlier this year was tossing all sorts of paper, including various electronics databooks. So when I hit the net looking for a datasheet for a part “CA3030”, I was coming up with nothing until I ran across the DataSheet Archive. These folks are apparently doing for electronics databooks what the Internet Archive does for, well, the Internet, or the WWW part thereof. They have scanned a lot of old databooks, and have information archived about all sorts of parts, including obsolete parts. Since I’ve been collecting chips since the late 1970s, this is very handy for searches.
So I was able to retrieve a datasheet for my CA3030 chips (they are an op-amp with all sorts of compensation options), my CA3048 (quad op-amps), LM567 (tone decoder), and even a AY-3-8500-1 (video games chip). The only part so far that the archive hasn’t come up with the goods on is my SN76488N sound effects chip that I got from Radio Shack back around 1979, but I do happen to still have its paper datasheet.
One of the circuits I want to make would combine analog signal processing with digital logic. I’m looking to have three analog inputs that will be massaged to get a peak voltage with appropriate decay rate, then use voltage comparators to get digital signals out, so I can tune the sensitivity of the circuit. Here’s the stuff I haven’t completely worked out yet… Let’s say I have inputs Left, Right, and Center. I want to be able to have three digital outputs (1, 2, and 3), where if the Left input goes high first, 1 goes high; if Right goes high first, 2 goes high; and if Center goes high first, 3 goes high. A latch and a little control circuitry for Latch Enable can get me this far. If I get fancier later, I could expand the options for outputs, perhaps using an encoding so that how close in time the inputs go high would determine the pattern of outputs and latch them. The time course I’m looking at would be that there would be a pattern in which the three inputs would go high within about one millisecond of each other, depending upon the exact distances separating the sensors and the location of the stimulus. I could reset the digital part after about 1.5 ms or so to get a new reading. Now that I’ve written this down, I’m seeing that two latches, one 3-input NOR, and two or three quad 2-input NANDs can get me pretty close to setting up the unencoded outputs. Some more will be needed to have this be resettable on demand.