[sdiy] VSM201 Vocorder Question

[Mattias Rickardsson]

Simon,

refreshingly mindbending.
Thanks for sharing! :-)

/mr


On 4 January 2016 at 15:37, Simon Brouwer <simon.o at brousant.nl> wrote:
> Hi,
>
> Note that for small input voltages the pulse width encoded signal will be
> around 50% pulse width, so there will be no "particular minimum pulse output
> before it jumps to flat".
>
> Also, the modulating signal is analog, so in principle it can also become
> arbitrarily small. The maximum attainable attenuation may be limited due to
> offset voltage though.
>
> At the output of this kind of amplitude modulator the audio is always a
> pulse width modulated signal with varying amplitude. However, in the VSM-201
> approach the audio signal is  PWM at the input, and the modulating signal is
> analog, while in another approach the audio signal is analog at the input
> while the modulating signal is PWM'ed.
>
> In the latter case you *will* have limited attenuation due to the pulse
> width not capable of being smaller than a certain value, but again, this is
> not how it works in the VSM-201.
>
> Finally, yes, the VSM201 is a true multiband vocoder: the "analysis and
> synthesis" circuit is repeated 20 times with different values for the
> components in the band filters.
> The band filters are composed of three consecutive peaking band pass
> filters, each one carefully tuned to a slightly different frequency to get
> the desired response (hence the annotation in the schematic "unten, oben,
> mitte" i.e. "lower, upper, middle").
>
> Best regards
> Simon
>
>
>
>> Op 4 januari 2016 om 2:17 schreef rsdio at audiobanshee.com:
>
>>
>>
>> This discussion has quickly become confusing. A few terms have been used
>> incorrectly, we've diverged onto different VCA implementations, and the
>> schematic is in German, so I'm not surprised at the confusion. Maybe we can
>> figure this thing out?
>>
>>
>> First of all, dynamic range is a relative term. It does not always refer
>> to (maximum) signal to noise ratio. While most vocoders surely have issues
>> with signal to noise ratio, I don't think that's the trickiest part of a
>> vocoder circuit operation to figure out. For most of this thread, we've been
>> discussing limitations in the dynamic range of the VCA section, which
>> presumably ranges from 0 dB down to some limit of available attenuation.
>>
>> With a digitally-generated PWM signal, the VCA volume steps are discrete
>> and finite, which is why we were discussing limits like -40 dB, -60 dB, and
>> -80 dB. Actually, even the digital PWM would be capable of infinite
>> attenuation when the pulse width is zero, but a jump from -40 dB to negative
>> infinity would potentially be distracting.
>>
>> It's been pointed out that the VSM201 vocoder does not have a
>> digital-counter-based PWM, but uses a comparator fed by an input signal and
>> a triangle wave. This design does mean that the PWM is not discretely
>> stepped, but it does not mean that the signal attenuation is continuous all
>> the way down to infinity. The comparator chip will have limited slew rate
>> and settling times, making a particular minimum pulse output before it jumps
>> to flat. That minimum output pulse width will determine the maximum
>> attenuation of the signal in that band, before the jump from maximum useful
>> attenuation to infinite attenuation. I assume that those 4053 2-channel
>> multiplexor elements can only be On or Off, and thus there is a minimum
>> pulse width and thus a maximum, finite attenuation before the signal drops
>> to negative infinity (muted).
>>
>>
>> Second of all, the discussion also focused on the fact that the audio is
>> PWM'd and not the CV. However, it seems to me that the audio is always PWM'd
>> by this type of VCA. The writeup by Rod Elliot is not for a vocoder, and
>> thus describes a CV input that is converted to a PWM signal that controls a
>> SPDT switch that alternates between the input audio and ground. Even in that
>> basic circuit, the audio is PWM'd. When adapting this kind of VCA to a
>> vocoder, one needs to process two input audio signals, breaking each into
>> multiple frequency bands, and then multiplying the amplitude of each band to
>> produce an output audio signal that sounds like a combination of the two
>> input audio signals. Ultimately, I assume that the output VCAs still operate
>> as described by Rod Elliot, where the audio is always PWM'd.
>>
>>
>> Finally, can anyone say whether this is a true, 20-band vocoder, or maybe
>> just a 3-band vocoder with 20-band graphic EQ on the input signal? How wide
>> range is the "speech or signal inputs" section? Would it handle a full
>> keyboard tone? I assume that the "modulation sound input" is full range.
>>
>> Brian
>>
>>
>> On Jan 2, 2016, at 10:45 PM, Tim Ressel <timr at circuitabbey.com> wrote:
>> > I think if an effort was made to make a DIY copy of the VSM201, a case
>> > could be made for swapping out the switched VCAs for 13700s or something.
>> >
>> > Another question: It looks as though what is being PWMed is the audio
>> > and not the CV coming from the detectors. The CV signals (I think) are
>> > labelled 'KK' but I cannot find their source. It is not helping that the
>> > schematics are copies and are in a language I don't read.
>> >
>> > Also I can't see the values for the filter components. Is there a BOM
>> > somewhere?
>> >
>> > --TimR
>>
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at dropmix.xs4all.nl
>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>