Re: easy self made analog sequencer

From Stefan Gruhl (CIP 92)
Sent Thu, Jan 1st 1970, 01:00

> 
>      Yes Stefan quite feasible. Just substitute the input signal for V+ and 
>      you have a step-programmable attenuator. Remember to account for the 


Thanks for the valueable input. Infact this is just what I am planning to do.
Normaled use with attenuation of the +5V to get usual CV and attenuation
of whatever signal if s.th. is plugged in.
Regarding the advice to use a LP on the output, would limit this
feature a bit though. I hope to use audio signals as well as bare cv.
With audio, I'd get a simple VCA kind of behavior. Especially with
FM this sounds usefull. I want to be able to play my favourite FM
tricks with the EMU. I love trhe EMU's precise filter to be fm'd by
one of its also precise OSC's. Now varying the amount does have this
massiv change in timbre and of course running the osc thru the sequencer,
having it attenuated (even turned completely off) will give
my filter much pleaseure I hope.
Therefore I did think I won't use a LP in gerneral.
What seemed of value was another switch that does this and also will
make the switches "prellfrei". That means when you switch while this position
is just used, you won't get multiple triggers from the meachnical vibrato.
This could be combined.

Now you got me started.
I didn'tintend to tell what else Idea's I got, but here I go now.
I got this simple idea how to drive multible different channels from
only one timer. I have the doepfer maq16 (a great tool) which can do
the weireds step variations. I won't copy that, but I know
that I maily use 8 or 16 ste sequences. But what I nee is having
them in variable speed/length. It's simple:

For a 16 step sequence I'd need 4 bits of address. Now I will use
6 bits.
I'll use the middle 4 for the usual use, as known. I will get this
by triggering the clock with double speed, so the last bit is flipping
twice each step. No difference so far.
Now I will have this easy switch which will shift the bits. that means
a row will either get bits 1-4 OR 2-5 OR 3-6. This way the actual
decoder for ONE row will work with normal, half or double speed.
EASY !

Second: Add a simple usint at the back. Once the decoder picked one
knob of the first row, and another decoder one from the second,
I'll have this routed to an indiv out. (buffered)
The second row though will get an additional 1 out of 2 switch.
Now when I switch a "link" switch this one will get active and
receive a bit of the priority one above the used bits.
I.e. running in noral speed I'll feed bit 6 to it. It'll now
switch between row one and two every cycle. So I can get a double length
sequence with just one analogue switch. Useing a chip with 4 of thouse
in it, I'll get also row 3 and 4 linked. I even can link the both pairs, but
I believe I would need use that too much.

With that feature I will limit the knobs to 8 per row and get it
usable as 16 step as well. I still think of having the first row
get in luxury mode. 8 indiv ins for each position. AS
I'll get the OUTPUT buffered, I won't be able to use it in reverse mode.




>      series resistance of the analog switch. Also you can plug a bunch of 


thanks. I did think of that. It's around 150 Ohm. Usein my link will
add up to 300 if used in serie. So I'll have another 150Ohm resistor
placed in the other line to keep the output the same if I link 2 rows.


I don't know yet, how I'll do the output buffering yet.
What I want is of course, maximum variety. I want to have exponential
behaviour on 5V to get musical scaling for making pitchsequences.
I also want to be able to turn an incoming signal totally down to 0
and to be able to pass it thru with (almost?) no decrease
in level.
That's the weak point I admit and I'll have to learn more about this.


>      signal sources into each step or plug a source into the sequencer 
>      output and run a signal through it backwards (analog switches are 
>      bi-directional). For this mode you probably have to set all pots to 
>      max however since the signal will be coming into the wipers.

True, but see, above, I'll have some outputbuffering. I will need that,
don't I ?

>      
>      Also the "glitches" between steps are mainly addressable (ha ha) by 
>      the accuracy of the counter that creates the select bits. It is not 
>      possible to cross-fade between values with the 405X-series analog 
>      muxes, but I think that the switching errors would be much less than 
>      using mechanical relays for example.

That's exactly the serie that I did look into.

>      
>      Now you've got me started again. I think there's schematics in an old 
>      Electronotes (naturally) which I'll look for. Good luck and keep me 
>      posted.
>      
> 
I'll do so, as look this is of interest for the list,
stefan
> 


-- 
* real electronic composers don't like their music be restricted    * 
* to the physical limitations of the human hand - AH - a great list *
* Stefan Gruhl * sngruhl@cip.informatik.uni-erlangen.de             *
* Monthly bonus sig: ************************************************
* "The course of true SOUND never did run smooth"                   *
* freely taken from Midsummernights dream by Willy S.               *
*********************************************************************
*