nice idea. interesting. let's.
the first problem was opening the damn thing up.
nope, never thought i'd say that.
managed to take the top lid off, but was afraid of destroying the plastic casing due to the lack of proper tools (and the resulting Brute Force to be applied) in an attempt to open it up further.
as time was an issue, i decided to go with what was accessible: the 5 preset switches at the top. turned out they had 1 common connection, switching to 5 different wires going to the 5 preset thumbwheels in the back of the radio. dingdingdingding: good news!
hooked all that up to a 4051 IC (8 to 1 analog mux), to control the presets without using the physical switches.
with only 5 presets and 8 possible connections, i repeated the middle 3 presets.
had them hooked up like a pendulum at first:
5 - 4 - 3 - 2 - 1 - 2 - 3 - 4 (repeat) but that was too obvious, the goal being a 'human' feel to the switching.
a small channel swap yielded better results:
5 - 4 - 3 - 2 - 1 - 4 - 3 - 2 etc.
i know it's as much of as a system as the other, but it seemed slightly less transparent and mechanical.
i actually tried using three separate clocks for the 4051 binary controls, but results were less random than i had hoped. in the end, decided to do the obvious and just control the 4051 with a 4024 binary counter from the recycled parts bin, so it loops 1 through 8 sequentially.
because that's what counters do, innit.
good enough, fast enough, works, cool.
next step: in an attempt to make the switching more Nature and less Rhythm, i felt a possible way to go was to randomize the clocking of that 4024 in some way, aiming for variable step length. The sequence still stays the same, but the 'note length' would change, so to speak.
nothing too extreme was needed - the idea is to change stations continually, so i just had to take the mechanical edge off. working with what i had lying around the bench, i tried two 4093 oscillators, one gating the other. using the output of this gated oscillator as a clock for the 4024 and thus the channel switching was too much of a system still - some rapid switching, some stasis, some rapid switching, some stasis.. nah.
however.
i ran this setup as a clock for a 4017 decade counter instead. ten off/ons - 0/1 - low/highs are required from the gated oscillator to complete a cycle on the 4017, and with both oscillators tuned well they seemed to produce a nice window of step length. it basically means every tenth transition might have a slightly different timing, enough so to feel semi-random.
then used a single output pin of the 4017 as a clock for the 4024. works ok.
it's pretty much an elaborate workaround, but good enough in my book.
(does anyone else feel a flowchart coming up?)
as an extra, i wanted to temporarily break into the sequential pattern as well, for added random feel to the fact that we're only using the 5 presets.
so i hooked up the wires for presets 4 and 5 to a second 4051, that is controlled by a third independent oscillator on the 4093. the two preset wires get connected to eachother using the 4051 as a switch.
connecting the wires for these presets changes their frequency to a common 'interstation', which is hardly controllable. could be noise, could be Radio Minerva, who knows.
the nice thing is that now, in the timing, it is possible to not just get
555555 44444 333333 2222 11111 4444 3333333 22222
but
55xxxx xx444 333333 2222 11111 xxxx 3333333 22222
with the connection as a sixth option (x).
depending on the speed of this connection, it either occurs or not at all. the possible overlap in the transition between stations 5 and 4 in the beginning of the loop is particularly interesting to seemingly shift the timing even more, so that's an added bonus.
here you can see the whole thing mounted inside the battery compartiment.
the switch and tactile button are there to disengage the clocking on the 4051, and skip through presets manually using the cute little button so you can easily set the stations with the thumbwheels above.
small disclaimer: there's probably easier ways to do this - or better ways, less crude and pile-on-as-you-go. please, feel free to enlighten me - i had to make do with limited time, resources and even more limited knowledge.. as usual.. poor me.
