[sdiy] Re: SV: Re: [sdiy] [SDIY] Faster Enveloppe Mod for Sixtrack ?

[David Moylan]

I don't think the lowpass filtering is to prevent click.  I think it is 
to prevent zippering, where you can hear the digital steps in the 
digitally generated envelope.  The difference with filtermod is that it 
doesn't change except by user modification, so if you set filtermod to 
10, it's voltage stays constant until you change it again.  The 
envelopes sweep everytime a note is pressed.

You could modifiy the smoothing filter by lowering the R value from 1M.  
Easy to experiment by just tack soldering a resistor in parallel.  Maybe 
you can find a faster envelope that still has acceptable zippering, but 
my guess would be that it is designed to be as fast as possible.

Dave

Frédéric Bourgeais wrote:

>Hello,
>
>David, Thanks for the explanation. I didn't know this CEM chip before, and
>looking at datasheet and schemos, I was impressed by the possibilities of
>the sixtrack ! (rather than my juno 106, who have also 6 oscillator).
>
>The only change between the digitally controlled VCA input, and
>Filtermod(Pin 5), is the addition a 1M resistor (R136), and 0.001
>uf(C171), making a passive lowpass filter. The cutoff frenqency is about
>15 hz.
>
>For info (To know if I understand right the schemos for the first time ;-)
>) Does it mean that attack signal, faster than 0,06 second are smooth (or
>slowed ?) it's probably to prevent "Click" on attack, but my friend want a
>more percussive enveloppe.
>
>Why, there is no filtering on filtermod ?
>
>-----
>
>Tim, thanks too for the suggestion, It's also, what I suggest him, If caps
>swapping was impossible. An External Log ADSR (7555 design) with a switch,
>on only one voice. Sure preset should be lost, but it's not a problem.
>
>I was thinking about gate signal problem to trig the enveloppe too. A very
>low threshold level, on 4051 output, could do the trick. Does this trick
>could introduce noticiable latency ? And yes, Pita is breed ;)
>
>-----
>
>Finally,
>Karl, Is it possible ? is it not too dangerous ? I remember a long time
>ago, a friend who succesfully overclocked a 486DX, simply swapping a 33
>Mhz quartz on the mother board. Is it the same trick ? What's kind of
>quartz is used on sixtrack and how many Mhz is it clocked ? which value to
>replace with ?
>
>Thanks all for your very interesting feedback !
>
>
>
>  
>
>>Try over clock the CPU! .-)
>>Make shure the 68B50 uart has its 500khz
>>maintained otherwhise your MIDI will go ape!.
>>Z80 can usuall be over clocked a bit.
>>
>>KD
>>
>>
>>--- Tim Parkhurst <tim.parkhurst at gmail.com> skrev:
>>
>>    
>>
>>>On 9/17/05, David Moylan <dave at westphila.net> wrote:
>>>      
>>>
>>>>I believe those are Sample/Hold caps. The 4051 is an analog
>>>>        
>>>>
>>>multiplexer
>>>      
>>>
>>>>which is routing DAC generated CVs to the caps where they are stored.
>>>>If I recall the 3394 datasheet mentions that all the CV inputs are
>>>>buffered hi impedance so there is no additional need for S/H buffers.
>>>>You don't want to change those caps.
>>>>
>>>>I don't think there's much hope in speeding up the envelopes as they
>>>>        
>>>>
>>>are
>>>      
>>>
>>>>CPU generated then just routed to the chip. Plus any cap that might
>>>>theoretically slow the envelope down is likely necessary to smooth
>>>>quantization of the envelope.
>>>>
>>>>I was very interested in the 3394 for a while and stared at the
>>>>datasheet a lot. Who doesn't like a synth on a chip?
>>>>
>>>>Dave Moylan
>>>>
>>>>Frédéric Bourgeais wrote:
>>>>
>>>>        
>>>>
>>>>>Hi all,
>>>>>
>>>>>A friend who own a SCI sixTrack, wish faster Attack to control the
>>>>>          
>>>>>
>>>output
>>>      
>>>
>>>>>VCA.
>>>>>
>>>>>I took a look to schematics available here:
>>>>>http://www.geocities.com/SoHo/2163/ST_schem/
>>>>>
>>>>>And CEM 3394 datasheet here :
>>>>>http://www.synthtech.com/cem/c3394pdf.pdf
>>>>>
>>>>>3394's Pin 18 is used to control the Final Gain CV, but unfortunalty
>>>>>Enveloppes are CPU generated, so it's not easiest as swapping a
>>>>>          
>>>>>
>>>capacitor
>>>      
>>>
>>>>>to a smaller one with good old analogue synths...
>>>>>
>>>>>Is there a last chance for faster CV, simply playing on caps or
>>>>>          
>>>>>
>>>resistor
>>>      
>>>
>>>>>values on the input of the 6 sixtrack's 3394 (such C171,R136,C169 on
>>>>>          
>>>>>
>>>U131
>>>      
>>>
>>>>>pin 18 here: http://www.geocities.com/SoHo/2163/ST_schem/st_b1.gif )
>>>>>
>>>>>Does these componants are used to smooth the Enveloppe CV ( a bit
>>>>>          
>>>>>
>>>like a
>>>      
>>>
>>>>>glide ? )
>>>>>
>>>>>Thanks for your help
>>>>>Fred
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>          
>>>>>
>>>> You might have better luck by building simple hardware EGs and adding
>>>>        
>>>>
>>>their outputs to the VCF and VCA CV inputs on the 3394 (yes, a nifty
>>>little
>>>chip). There are however, two major drawbacks to this approach: First,
>>>getting the right 'gate' signal to trigger the EGs would be a PITA
>>>(perhaps
>>>from decoding the right address on the 4051s). Second, these EGs would
>>>not
>>>be programmable, and even if they were voltage-controlled, there would
>>>be
>>>little hope of adding these settings to the existing program (unless you
>>>felt like doing a major rewrite of the Six's operating system).
>>>
>>>Tim (and I thought PITA was a type of bread) Servo
>>>--
>>>"Imagination is more important than knowledge." - Albert Einstein
>>>
>>>      
>>>
>>    
>>