This is not my first module design, but it is the first one I have actually built, rather than leave it as a sketched and partially tested circuit drawn over several pages in a notebook that I use as a base for soldering. I hope that I can learn from this experience and use the new-found knowledge to create several more modules that I find to be interesting. What follows may appear to be rambling, and it is, but... ummm... yeah.

Inspired by all of the analog sequencers I've seen out there, I decided, on a mildly warm (crazy Houston weather) December eve, to diagram a sequencer with a certain feature I had never seen before (not to say that it doesn't probably exist; I've just never seen one): voltage-control of the reset point. There has been much recent talk about National's relatively new bargraph ICs (namely, the LM3914, LM3915, and LM3916), and many new designs based on them thrown around on the SDIY mailing list, but I have had a few of them (LM3916) for almost a year now, after getting them as free samples back in my days of trying to build small portable speakers. With these little babies, long comparator chains are no longer necessary, and much space can be saved. I figured I could save 4-5 ICs (depending on how you look at it) in my application, so I decided to join the bandwagon.

It was easy enough to follow the datasheet and get a 0-10V signal to move up and down the LM3916s 10 outputs, but the problem came when trying to interface them with the analog switches that control which step will go straight to the reset pin of the 4017 counter. The outputs swing from no connection to a connection to ground, so i devised a way to convert to a swing from ground to +V by adding pull-up resistors followed by an inverter to each output. Now the bargraph chip could perform its function.

When first designed, the sequencer was only meant to be 8 steps, but in May 2003 I decided that wasn't going to cut it, so I upped it to 16 steps. This presented new problems. The 4017 was very simple to cascade; all that was needed was an AND gate (which I decided to make from 2 NPNs to save space). The LM3916s, however were only meant to cascade properly when driving LEDs directly, so whenever the outputs switched to the second IC, the last output of the first one stayed on. Thinking very hard at the situation, I decided that all I needed was an XOR gate to fix this; all used outputs from the second IC to one input of the gate through diodes, and the last output of the first IC into the other input. The output of the gate would then be used as the first IC's last output.