| .Say hello to MILTON! |
|
|
 |
|
|
|
NEW! Click here for information on the new Asynchronous Clock Board (ACB) NEW! Click here to join the Milton-Sequencer yahoo group Click here to see a Gallery of completed Milton's from throughout the world Click here to see a Quicktime movie of Milton in action Click here for a soundbite of Milton in action Click here for a Milton block diagram Click here for a Milton faceplate Click here for Milton Operation Manual as a PDF |
|
|
| I've made two sequencers. The first, a 2 x 8, is completed and
installed into the Eclectic Music box (click here for picture) It was originally designed as a two bank for simutanious control
of the attack and decay phases of my VC envelope generators. Apart
from the standard pulse in/hold/rest functions, it also allows
up/down, pendulum and pseudo-random counting, enabled by by either
an external pulse or by a faceplate switch. It has an bank A output,
a bank B output, and a combined A/B bank output (16 steps), as
well as seperate pulse outs for each of the 8 stages.
This was good enough for me for some time. The problem is was having was in the pseudo random operation. To definition they aren't truly random and repeating ghost patterns manifest continually. While this effect wasnt necessarily displeasing, it wasn't exactly what I wanted. I began work on another sequencer to remedy this. A much larger one, a voltage controllable one which would not only give me 4 banks of 16 steps, but full VC control of direction, using the Buchla 200's sequencer as a functional model. (Click here for a block diagram). As of the time of this writing (December 9, 2002), this new sequencer, code named Milton, is operational as a prototype. It will be completed and mounted into the Eclectic Music Box in the space currently held by the Deopfer MAQ 16/3 (right side of bottom row) within a month's time. The sequencer requires an external pulse for operation, which is routed through a flip-flop allowing for both manual and external (pulse activated) start/stop control of the incoming clock. An obedient clock. There is also an output for this obedient clock for syncing to external devices. There are the normal reset and hold functions, separate trigger outs for each stage and some things that aren't so normal: Milton is fully voltage controllable. There is an external input for a positive analog voltage which is converted to four bit digital which is used to load the preset inputs of the 4516 counter which does all of the counting within the machine. The operation is simple - the input is quantized via na A to D into 16 levels (digital 0000 thru 1111). Depending on the level of the incoming voltage, the sequencer will be instantly sent to the corresponding stage. This control runs through a latch which is gated by the same clock which drives the machine, so these changes are timed to the same current sequencer speed. This may not be immediately apparent,. Depending on the speed of the changes and activity at the VC input, stage advancement may appear to be slower when in actuality, the voltage is forcing the sequencer to remain at the same stage for more than 1 beat. Operation is straight forward from then on. You want random? Insert a S+H sampling noise. Backward operation? Insert a ramp. Pendulum? Insert a triangle wave. This plus a billion different patterns your incoming voltage may afford you. The VC input has two modes of operation and two sub-modes within those - SYNCHRONOUS or ASYNCHRONOUS. In the first, inputs to the VC input will not take effect until the next incoming pulse is received, therefor behaving like a standard Sample & Hold. In the second, the changes to step are immediate, regardless of the incoming clock. In fact, an incoming clock is not even required to run Milton in this configuration. As far as the sub-modes, whether you're in SYNC or ASYNC operating modes, in the first sub mode the VC cuts off if the voltage present is less than .2 volts (below the stage 2 threshold) With this arrangement, the input sitting there that will have no effect on the counting sequence until that voltage rises to a level above the .2 volts. So, if the input was an envelope generator, the sequencer would be going on it's merry sequential way until you fired the envelope, then all hell would break loose. There is no internal clock on Milton. It (he) depends on an external
trigger. That trigger is routed through a flip-flop, allowing
both full manual (buttons) and VC (trigger inputs) control of
start and stop modes, with leds indicating the state. There is
also a separate output for that obedient trigger to the outside
world. I arranged this sequencer so that all of the main counting, clock control and VC functions are handled on a main engine board, and each of the banks with all of their associated controls (voltage pots, stage trigger outs, programmable pulse banks, summed voltage and pulse outputs and stage LEDs) on each of those bank boards. The Counting engine is terminated by 4050 line drivers with enough muscle to drive up to 15 banks simultaneously under full load. Obviously, some of these bank boards do require full assembly. Some will have only the voltage pots and summers. In my arrangement, the top bank will also have the stage led drivers and stage trigger output componentry installed, while the bottom bank will also have the programmable pulse busses installed. I have elected not to include a full schematic of this instrument at this time. One, I need to make damn sure all the bugs are worked out before I would feel comfortable having another person embark into such a project. But there is an alternative: I will be making PCBs of this design which will be made available to anyone who wants them at a minimal cost. Please keep posted for updates, A QUICK COMPARISION BETWEEN RANDOM AND PSEUDO-RANDOM CONTROL: Anyone who has used a Serge TBK will notice that their 'random' function creates undefiable patterns during its use. These 'sidebands' are unavoidable in the Serge method of generating random due to intervocalic ratios which present themselves from the constant (metric) operation of the high speed clock driving the pseudorandom engine vs that of the (comparatively low speed) main counting engine it controls. In brief, 'random' is created in the Serge by using a secondary counter which is running at an comparatively high speed (high kHz, well above audio range) to that of the main counter, in the case of the Serge, a 4516. This high speed counter runs at a constant metric pace and is counting sequentially (L. to R., step 0 to 15), over and over again, very quickly. The outputs of this secondary clock are fed through a latch which gates (samples) the its current count and feeds it into the presets of the main counter at the exact moment the main counter receives an incoming clock - much like a digital sample and hold. This four bit preset information overrides the counting order of the 4516 main counter's sequence, and forces the unit to the preset number generated by the secondary clock. In that it's running so much faster, chances are this is going to send the main counter to something other than a adjacent step in the direction the unit was travelling before this control is applied. But because both the external secondary counter IS and the main counter MAY be running metrically (constant pace), patterns will appear. This is the exact method incorporated in the first sequencer I built. Although I haven't seen or used one in the flesh, I believe the Modcan unit runs under the same principle. I say this because of the up/down jack. It points to the use of the 4516 as the main counter. In that Bruce has random SELECTION vs. voltage control, I think he used the same process as the Serge and my first sequencer: I use a 4516 in Milt as well, but have incorporated an A to D that attacks the preset inputs of that chip, thus enabling direct external generation of the preset loading, as opposed to external ENABLING of internally generated four bit binary data. What you are hearing: This is Milton arpeggiating a major chord. The direction is controlled by a tirangle wave which is also controlling the harmonic content of the two voices, both triangle to square variable waves via Blacet VCOs. The sequencer is also being stopped and starterd by a single secondary square LFO, whose base frequency is being controlled by stepped rtandom bvoltage. Like the ringing when the phrases stop? Clock that up to the Low Pass Gates! |
