Super Klee Sequencer

[v-un-v]

aahh at last! some diagrams. Diagrams make more sense than words. Bring em on!!

Fast and bolbous?? The secret to those yummy elastic twangy sounds are in the vactrols right?

[Wild Zebra]

"Bulbous also tapered"
he he
Only Sun Ra I have is: Concert for the Comet Kohoutek, its been awhile since I gave it a spin
and the only Mo Wax I have is: Doctor Octagon 3 LP set

I wish this Super Klee stuff wasn't so far over my head. Nice work Scott! I thought my workbench was crazy!

Well I just wanted to pop in and and feel wanted, and a part of something

What Up V-UN-V!! How's it hanging?

[seraph]

[v-un-v]

[Wild Zebra]

[v-un-v]

[Scott Stites]

Gaaah! And people ask me why I only had one kid.....

Anyhow - how about a triphonic Klee?

Driving home from work last night, it hit me that one can still have three gate/trigger busses without resorting to overly expensive toggles or using slide switches - the SPDT ON-OFF-ON that I'd intended for two busses will happily serve three.

The way I am deriving the gates and triggers goes like this (should have a diagram, but what the heck):

As the design sits right now, each bus switch will send its bit to one of two busses. Lets say a bus switch is set to bus 1. The bus is ANDed with a slightly delayed version of the clock signal. When any stage assigned to the bus goes high, this produces a gate that's as long as the clock period. From that gate, a trigger is derived. So, if the step associated with the switch goes active (the LED for that stage lights up), we get a gate and a trigger on the Bus 1 gate and trigger outputs. As an aside, the ramifications of the gate being as long as the clock period provides the exquisite possibility of varying the gate period by clocking the Klee with a pulse width modulated VCO.

Anyway, if that switch is set to bus two, the same thing happens, except, well, it's on bus two. If the switch is in the center off position, the signal doesn't go to any bus, so it's like a 'rest' if you're triggering things from bus 1 and bus 2.

So, the idea is, why not NOR the two busses together to form a third bus? In a normal sequencer, which the Klee will certainly do by selecting only one active bit, this would have the effect of sending one gate/trigger out per clock pulse from one of the three busses, while the master gate and trigger will always send a gate and trigger out per clock pulse.

Putting in more than one active bit is what brings on the Klee factor. Now things get....interesting. With two busses and more than one active bit in the register, it's possible that Bus 1 and 2 can fire at the same time - it just depends on if the bits land on a stage selected for bus 1 and a stage selected for bus 2 simultaneously. Gating two different voices from these two busses will now create what I think will certainly be some really cool one or two simultaneously voiced sequences. Adding the third bus will get really interesting, but I'll have to diagram that to adequately describe what I think will happen. It is quite Klee in nature.

But, that's not the triphonic part, it just serves to enable the triphonic idea. I thought of it driving to work this morning.

A couple of years ago, Romeo turned me onto the Buchla S&H from the 266 SOU module. Without going into a lot of detail, the S&H takes a signal and pulse input and creates three sampled and held outputs. Two of the outputs alternate between samples, while the third output takes a sample with each pulse. Around that time I recorded an 8 step sequencer using the Buchla 266, and it was cool - notes would overlap, whereas on a normal sequencer, it is essentially good at creating monophonic lines. Can you see where I'm heading with this?

The three busses each put out a gate signal and a trigger signal. How about a third signal for each bus - a sampled and held voltage? In other words, bus 1 will sample and hold the Klee voltage every time it is triggered, bus 2 and bus 3 will do the same. This means that three voices can overlap, and when arranged spatially, I would think it would be a stunning effect. Kind of along the same lines as an analog shift register, but with a different function. The fourth output, the normal Klee voltage, would certainly add another line to it, and the master gate and trigger would drive anything continually with each clock pulse (like the percussive sound in sample 11).

Front-panel-wise it would mean three extra connectors. Circuit-wise, it wouldn't be too awfully difficult, either. I'm thinking LM398s, my old standby. A FET based S&H would probably work as well.

I've got to find a shorter route home.....

Cheerio,
Scott

[toppobrillo]

also, a tin teardrop.

[vtl5c3]

Hehe... talk about feature creep!

I'm not sure what voltage range the SuperKlee is going to output... but you could potentially use a CD4016/4066 as a triple (or quad) S&H. Just a thought... they're cheap and abundant.

Romeo

[Scott Stites]

[Uncle Krunkus]

If it weren't for the stones we wouldn't know the difference.
Keep it.

[piedwagtail]

[Scott Stites]

Uncle K - good point, I'll leave it in.

Robert - yes, I'd say this is more along the lines of the Muse, and the random function would be along the lines of the quantized random voltage function of the SOU.

I'm going to do a basic Model 2 with just the two busses without the logic bus or S&H. I'll then do a Model 3 Triphonic version.

Question: The way I'm deriving the triggers and the gates actually puts the output trigger 20 uSec *behind* the output gate. My EGs process this just fine, but I wonder if there are EGs out there that wouldn't like this situation. 20 us isn't a lot of time, but I'd hate to have someone build one and find out it makes their EG wonky. I can arrange the outputs so they're simultaneous, but that's a few extra parts, and I'd like to avoid it if it's not necessary. So - anybody out there know of any issue with this?

Thanks,
Scott

Edit: Output trigger width is 1 ms.

[v-un-v]

[v-un-v]

[Uncle Krunkus]

I used to do that alot.
I'd have an initial midi track which was literally random notes and rests.
This would be the seed for say 2 other tracks, a kick/snare rhythm and a bassline.
Then I'd mute the first track and fit something to it's offspring.
I'd often mute the original drum and bass offspring (or re-write them) as well. It was a good way of evolving an idea out of nothing. Sometimes it would end up just being an exercise is composition, sometimes it would evolve into a great track.

[Scott Stites]

Yes, I found things living in the breadboarded modular.

The thought occurs to me that my concern over 20 us of difference between gate and trigger is a bit foolish. After getting some much needed sleep, I realized that most EGs are retriggerable - the trigger is independent of the gate. If you have an EG gated, then trigger it, it still reacts to the trigger. The human brain has no way of knowing that something happened .02 ms after something else did.

This day's chapter deals with the 'Merge' function I mention every now and then. Its function is to merge the gates of adjacent selected bits into one big gate instead of individual discrete gates. It also eliminates the accompanying discrete triggers - it produces one trigger for the gigantor gate.

This function is intended to be able to create notes of varying lengths in a sequence. Another method would be to clock the Klee with a pulse width modulated VCO (the gate will be as wide as the clocking pulse). In this instance, the gates will be of varying length, but the triggers and gates will occur in unaltered numbers.

I've attached a couple of illustrations of the merge function. They illustrate the row of bus switches in varying up (bus 1 selected), down (bus 2 selected) and middle (no bus selected) positions. Below that is the clock signal, and below that are the gate and trigger outputs of each bus.

The examples illustrate the simplest use - using only one active bit in the register. Once more than one bit goes active - Klee function as opposed to 'normal' sequencer function - it gets quite varied. I'd have to produce 64K drawings to illustrate all of the permutations .

Cheerio,
Scott

---

Merge_Off.JPG
Description:	Merge function off for busses 1 and 2.
Filesize:	80.73 KB
Viewed:	675 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

---

Merge_On.JPG
Description:	Merge function on for busses 1 and 2.
Filesize:	78.49 KB
Viewed:	519 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[Scott Stites]

I drug the recorder downstairs to record a sample illustrating the switch bus and merge function. Of course, at this writing, I only have *one* bus wired up.

00:00 - 00:09
The sample starts out with one stage selected for bus 1. I have bus one set to load the shift register when it goes high. In this instance, the first stage encountered is high, so the bus reloads with every clock pulse - IE, the same note repeats. At 00:09 I flip the reload function off, and the sequence proceeds. Merge function is on, so the gates are merged together, and there is one trigger for each merged gate. Voice is the filter processed by the WM.

00:38
I flip another stage onto the bus, increasing the number of notes firing the EG. Bus is still merged. I mix in the VCO, at first pitched for a much lower note.

00:50
I turn the merge function off. Now the merged gates separate into separate gates, each with a trigger. This produces more discrete notes.

As this section proceeds, I mix in another voice that is fired by the 'master' gate and trigger outputs, which are constantly triggered/gated. This is the same patch as illustrated earlier in this thread, so I can change the pitch of the VCO with the keyboard. In this section I hit a couple of keys on the keyboard, and the altered pitch follows the Klee voltage, producing a few higher notes every now and then.

02:15
The second stage is flipped off the bus, so the pattern changes to fewer notes.

02:26
The same stage is flipped back onto the bus, returning the original pattern, and the sample fades.

This weekend I hope to finish the bus and get things on schematic. I've got a schemo of the bus, but it's all M2L using CD40106s - I think if I throw in some CD4093 NANDs, it'll be less part-filled and easier to grok.

Cheers,
Scott

[piedwagtail]

Phew!What's Bootsy doin' ?(downstairs)....
Like those high contrapuntal notes,Scott.

Robert

[v-un-v]

[Blue Hell]

Would think "drug" to be some past tense form thingy of "to drag", the wiktionary seems to agree.

drag

[Scott Stites]

I have to drug the recorder - it won't do Klee if I don't......

Grok - Heinlein. Stranger in a Strange Land (I'm but an egg.....)

Sleep - surprisingly, the last two nights have been 8 hours - 4 more than my normal (until the sleep deficit takes over).

Now, contrapuntal, I'm going to have to look that one up

Cheers,
Scott

[Scott Stites]

Ah! Counterpoint!

[v-un-v]

[Scott Stites]

You're confusing the language I speak with English

Kansan past tense of drag is dragged or, optionally, drug.