MOTM

What's the low-frequency limit for the '420 and '440 filters?  I see the
inputs are AC coupled, and I'm wondering if the bipolar buffers limit the
low-frequency response as well.

I'm interested in filtering control voltages, inspired partly by the '101
vibrato circuit and partly by trying to make a PLL with two '300's and the
ring modulator from the '110.  I used the '820 lag as a cv filter for the
PLL with interesting results; depending on how you adjust the frequency and
tracking on the oscillator in the loop along with the lag time you get some
interesting 'wobble' early in the note, and it changes the timbre of the
note overall (like FM, I'd guess).  But as I understand it, the '820 would
have a -6dB/oct response unlike the other filters.

Are there any filters in the works that will go down to DC or are they all
strictly audio filters?

                    ...Craig

If you want to filter DC, use the MOTM-820. It *is* a DC filter.

Paul S.

>I'm interested in filtering control voltages,

Afaik, only the 830, 820, 700, and half of the 110 can process DC.  I don't
know about the 130 or 190, even though I've asked a few times.  I guess a
320 can be considered a DC processor of some sort -- it has DC inputs and
can produce a DC output for considerable length of time.

>inspired partly by the '101
>vibrato circuit and partly by trying to make a PLL with two '300's and the
>ring modulator from the '110.

My, that sounds ambitious :)  Let me see if I understand what you mean. You
are trying to have one 300 track another 300, taking the square outputs of
both 300's, using a 110 ring mod as a "phase comparator", smooth the output
with an 820, then send the output to the second 300 using the FM input??
Or am I completely lost here??  I think I understand how a "four quadrant
multiplier" would give the sum and difference frequencies using sines, but
I don't see how it would give you just the difference like an XOR using two
squares.  Wouldn't it give you something much more complex??

Does it eventually lock??  Are the resulting outputs of the two 300's 90
degrees out of phase??

>I used the '820 lag as a cv filter for the
>PLL with interesting results; depending on how you adjust the frequency and
>tracking on the oscillator in the loop along with the lag time you get some
>interesting 'wobble' early in the note, and it changes the timbre of the
>note overall (like FM, I'd guess).

I think you would get FM if the 820 was set too fast, and afaik, some
ripple is typical for a "type I phase detector".

Regardless whether it works as a PLL or not, that sounds like quite an
interesting patch!!

>But as I understand it, the '820 would
>have a -6dB/oct response unlike the other filters.

Yes, the other MOTM filters are steeper, but they won't pass DC.  The PLL
circuits I've seen only use one cap to smooth the VCO control voltage, so
afaik, that would make them 6dB/oct filters as well.

> >Craig Critchley wrote:
> >I'm interested in filtering control voltages,

> <media.nai@...> responded:
> Afaik, only the 830, 820, 700, and half of the 110 can process DC.  I
don't
> know about the 130 or 190, even though I've asked a few times.  I guess a
> 320 can be considered a DC processor of some sort -- it has DC inputs and
> can produce a DC output for considerable length of time.

I think only the 820 would be usable for LP filtering control voltages (slow
waveforms).  Of course, LP filtering a flat DC signal would not produce
much. :)

> >Craig Critchley wrote:
> >inspired partly by the '101
> >vibrato circuit and partly by trying to make a PLL with two '300's and
the
> >ring modulator from the '110.

Ok, correct me if I'm wrong, but isn't soft sync basically a form of PLL?

> <media.nai@...> responded:
> My, that sounds ambitious :)  Let me see if I understand what you mean.
You
> are trying to have one 300 track another 300, taking the square outputs of
> both 300's, using a 110 ring mod as a "phase comparator", smooth the
output
> with an 820, then send the output to the second 300 using the FM input??
> Or am I completely lost here??  I think I understand how a "four quadrant
> multiplier" would give the sum and difference frequencies using sines, but
> I don't see how it would give you just the difference like an XOR using
two
> squares.  Wouldn't it give you something much more complex??

Hmmm... The 110 produces both sum and diference sidebands, so when one tunes
both 300s perfectly at 1000Hz, there will be a resultant 2000Hz upper
sideband signal and a DC output corresponding to the difference in the phase
of the two input waveforms.  Now if they're slightly detuned, say by 1Hz,
then the 1Hz low band signal could be retrieved from the output by LP
filtering with the 820.  In any case, I would think the PWM settings of the
two 300s would be critical in this application.  If set to anything other
than precisely 50%, it would add either positive or negative offset to the
LP filtered control voltage.  It could be interesting to try using the PWM
input to control the offset.

Another issue, when using square waves through a ring mod, there are *tons*
of side bands due to all of the odd harmonics present.  Why not use sine
waves?

<snip>

> Regardless whether it works as a PLL or not, that sounds like quite an
> interesting patch!!

Agreed!  I'm curious Craig, what led you to this experiment?

George

From: <groovyshaman@...>


> > >Craig Critchley wrote:
> > >I'm interested in filtering control voltages,
>
> > <media.nai@...> responded:
> > Afaik, only the 830, 820, 700, and half of the 110 can process DC.  I
> don't
> > know about the 130 or 190, even though I've asked a few times.  I guess
a
> > 320 can be considered a DC processor of some sort -- it has DC inputs
and
> > can produce a DC output for considerable length of time.
>
> I think only the 820 would be usable for LP filtering control voltages
(slow
> waveforms).  Of course, LP filtering a flat DC signal would not produce
> much. :)

Well, that's true.  :-)  The 101 vibrato has a low frequency bandpass and
putting two together I got curious about low frequency filters in general.

>
> > >Craig Critchley wrote:
> > >inspired partly by the '101
> > >vibrato circuit and partly by trying to make a PLL with two '300's and
> the
> > >ring modulator from the '110.
>
> Ok, correct me if I'm wrong, but isn't soft sync basically a form of PLL?

It certainly produces a different sound; you don't get the overtones from
the reset, but you do get some FM-like effects, particularly around the
attack while the loop vco is still hunting for the right frequency, if you
set the filter time long enough.

>
> > <media.nai@...> responded:
> > My, that sounds ambitious :)  Let me see if I understand what you mean.
> You
> > are trying to have one 300 track another 300, taking the square outputs
of
> > both 300's, using a 110 ring mod as a "phase comparator", smooth the
> output
> > with an 820, then send the output to the second 300 using the FM input??
> > Or am I completely lost here??  I think I understand how a "four
quadrant
> > multiplier" would give the sum and difference frequencies using sines,
but
> > I don't see how it would give you just the difference like an XOR using
> two
> > squares.  Wouldn't it give you something much more complex??
>
> Hmmm... The 110 produces both sum and diference sidebands, so when one
tunes
> both 300s perfectly at 1000Hz, there will be a resultant 2000Hz upper
> sideband signal and a DC output corresponding to the difference in the
phase
> of the two input waveforms.  Now if they're slightly detuned, say by 1Hz,
> then the 1Hz low band signal could be retrieved from the output by LP
> filtering with the 820.  In any case, I would think the PWM settings of
the
> two 300s would be critical in this application.  If set to anything other
> than precisely 50%, it would add either positive or negative offset to the
> LP filtered control voltage.  It could be interesting to try using the PWM
> input to control the offset.
>
> Another issue, when using square waves through a ring mod, there are
*tons*
> of side bands due to all of the odd harmonics present.  Why not use sine
> waves?

Actually, I am using sine waves from the '300 as the inputs to the ring
modulator, patching other outputs through a filter for the audio output.
I'm not using the PWM inputs, I'm using FM.  I had the note voltage patched
to both 1v/oct inputs, with the loop vco detuned so it would have some work
to do, then I switched the loop vco to another fm input so I could change
how well it tracks the 1v/oct note.  Different amounts of initial detuning
produce different timbres as well, I suppose because of phase and some
leftover FM, but the VCO's can't start too far apart or the loop won't find
the frequency.

I would imagine that a ring modulator and an XOR would produce similar
results for square waves, since an XOR is multiplying if you call the states
+1 and -1.  Since the next module after that is a low-pass filter, the extra
harmonics might not make much difference.

>
> <snip>
>
> > Regardless whether it works as a PLL or not, that sounds like quite an
> > interesting patch!!
>
> Agreed!  I'm curious Craig, what led you to this experiment?

Mostly curiosity after reading about PLLs.  I'd seen various 4046-type PLLs
in the past but then saw a block diagram for a sine vco with a balanced
modulator and thought "hey, i have all that stuff in my motm..."

>
> George

                        ...Craig

>I think only the 820 would be usable for LP filtering control voltages
>>(slow waveforms).  Of course, LP filtering a flat DC signal would not
>>produce much. :)

Right, I was trying to list all the modules that can process DC.

>Ok, correct me if I'm wrong, but isn't soft sync basically a form of PLL?

Not afaik.

>Hmmm... The 110 produces both sum and diference sidebands, so when one
>>tunes both 300s perfectly at 1000Hz, there will be a resultant 2000Hz
>upper
>sideband signal and a DC output corresponding to the difference in the
>>phase of the two input waveforms.

Right, so if you filter the AC component from a phase comparator using a
four-quadrant multiplier (a linear type I ) the resulting voltage is the
difference in phase.  So the cap is a filter.  In contrast an edge-detector
(type II) phase comparator outputs either positive or negative pulses when
the two signals are out of phase that charge or discharge the cap.  So the
cap holds the VCO control voltage.  I don't think that can be done using
MOTM, or at least not yet :)

>Now if they're slightly detuned, say by 1Hz,
>then the 1Hz low band signal could be retrieved from the output by LP
>filtering with the 820.  In any case, I would think the PWM settings of the
>two 300s would be critical in this application.  If set to anything other
>than precisely 50%, it would add either positive or negative offset to the
>LP filtered control voltage.  It could be interesting to try using the PWM
>input to control the offset.

Afaik, they would both have to be square.

>Another issue, when using square waves through a ring mod, there are *tons*
>of side bands due to all of the odd harmonics present.  Why not use sine
>waves?

Apparently, Craig got it to work, but I've never seen a type I phase
comparator use sine waves.  Then again, I'm not an EE.  However, I've spent
a bit of monkey time with a 4046 using pulses.

>It certainly produces a different sound; you don't get the overtones from
>the reset, but you do get some FM-like effects, particularly around the
>attack while the loop vco is still hunting for the right frequency, if you
>set the filter time long enough.

So you are saying that it locks eventually??  Wow!!

>Actually, I am using sine waves from the '300 as the inputs to the ring
>modulator, patching other outputs through a filter for the audio output.
>I'm not using the PWM inputs, I'm using FM.  I had the note voltage patched
>to both 1v/oct inputs, with the loop vco detuned so it would have some work
>to do, then I switched the loop vco to another fm input so I could change
>how well it tracks the 1v/oct note.  Different amounts of initial detuning
>produce different timbres as well, I suppose because of phase and some
>leftover FM, but the VCO's can't start too far apart or the loop won't find
>the frequency.

That makes sense, but I'm as bit surprised that it finds it locks at all!!
I would like to hear your theory of how this works.  What is the phase
relationship between the two 300's when locked??  I'm guessing that when
the two VCO's are at the same frequency the ring mod output is double that
frequency (which is filtered out) and a DC offset that approaches zero as
the two signals approach 90 degrees out of phase.

>I would imagine that a ring modulator and an XOR would produce similar
>results for square waves, since an XOR is multiplying if you call the
>>states +1 and -1.  Since the next module after that is a low-pass filter,
>>the extra harmonics might not make much difference.

The output of an XOR will be a frequency equal to the difference in
frequency to its inputs.

>Mostly curiosity after reading about PLLs.  I'd seen various 4046-type PLLs
>in the past but then saw a block diagram for a sine vco with a balanced
>modulator and thought "hey, i have all that stuff in my motm..."