Discussion about the Korg PolySix synthesizer

Hi All,

I have a question regarding the MG Delay knob and whether your Polysix responds the same way...

I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.

Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.

Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.

If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.

Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.

Thanks,

Chip

Hello Chip,

I've done the same test and my P6 behaves as yours... at each note, the cutoff goes quickly from the value it is set by the cutoff knob to the value set by the MG.

rgds,

Ed.

--- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@...> wrote:
>
> Hi All,
>
> I have a question regarding the MG Delay knob and whether your Polysix responds the same way...
>
> I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.
>
> Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.
>
> Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.
>
> If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.
>
> Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.
>
> Thanks,
>
> Chip
>

Thanks for the confirmation!

I can't believe that I didn't notice this behavior until now. Now, I'm starting to look at mods to avoid this behavior.

Probing around this part of the circuit with my oscilloscope, it looks like the "problem" is that the output of the sample-and-hold circuit that generates the "MG Delay" CV doesn't go all the way down to -5.0V even with the knob turned to zero. This value is used in a comparator at IC14. The value it's being compared to does drop to -5.0V with every new keypress. So, even with MG Delay set to zero, any new keypress trips the comparator, which trips the MG Amount down to zero. When the MG Delay is set to zero, the MG Amount is nulled for a very short time (25-50 ms), but it's long enough to be quite noticeable.

Based on more probing around makes me think that the trip line (with MG Delay at zero) of the comparator is something like -4.6V to -4.4V. I'm still not quite sure of the exact value because I can't directly probe the CV representing the "MG Delay" because, even at 1MOhm input impedance, my oscilloscope trains sample-and-hold cap too quickly.

I'm churning on a couple of ideas: (1) Connect the sample-and-hold cap for MG Delay (ie, C18) to -5V instead of to ground, or (2) change the DC reference of R82/Q5/C33 to -4.6V instead of -5V, (3) put a diode with a forward drop of 0.4V to 0.6V on Q5 so that it drains C33 only to -4.6V or -4.4V instead of to -5V.

Anyone got any thoughts?

Chip

--- In PolySix@yahoogroups.com, "djangowes" <djangowes@...> wrote:
>
>
>
>
>
> Hello Chip,
>
> I've done the same test and my P6 behaves as yours... at each note, the cutoff goes quickly from the value it is set by the cutoff knob to the value set by the MG.
>
> rgds,
>
> Ed.
>
> --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> >
> > Hi All,
> >
> > I have a question regarding the MG Delay knob and whether your Polysix responds the same way...
> >
> > I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.
> >
> > Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.
> >
> > Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.
> >
> > If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.
> >
> > Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.
> >
> > Thanks,
> >
> > Chip
> >
>

well, on my cheap scope, using a 10x probe seems to increase the apparent input impedance...so i can now get a decent reading of the CV for MG Delay. When set to zero, it's value is -4.73V (as measured on Pin 2 of IC14).

So, if I could make the other leg of the comparator (the leg that is affected by the keypress signal) never go below -4.73V, I would avoid this behavior with the MG that I don't like.

Chip

--- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@...> wrote:
>
> Thanks for the confirmation!
>
> I can't believe that I didn't notice this behavior until now. Now, I'm starting to look at mods to avoid this behavior.
>
> Probing around this part of the circuit with my oscilloscope, it looks like the "problem" is that the output of the sample-and-hold circuit that generates the "MG Delay" CV doesn't go all the way down to -5.0V even with the knob turned to zero. This value is used in a comparator at IC14. The value it's being compared to does drop to -5.0V with every new keypress. So, even with MG Delay set to zero, any new keypress trips the comparator, which trips the MG Amount down to zero. When the MG Delay is set to zero, the MG Amount is nulled for a very short time (25-50 ms), but it's long enough to be quite noticeable.
>
> Based on more probing around makes me think that the trip line (with MG Delay at zero) of the comparator is something like -4.6V to -4.4V. I'm still not quite sure of the exact value because I can't directly probe the CV representing the "MG Delay" because, even at 1MOhm input impedance, my oscilloscope trains sample-and-hold cap too quickly.
>
> I'm churning on a couple of ideas: (1) Connect the sample-and-hold cap for MG Delay (ie, C18) to -5V instead of to ground, or (2) change the DC reference of R82/Q5/C33 to -4.6V instead of -5V, (3) put a diode with a forward drop of 0.4V to 0.6V on Q5 so that it drains C33 only to -4.6V or -4.4V instead of to -5V.
>
> Anyone got any thoughts?
>
> Chip
>
> --- In PolySix@yahoogroups.com, "djangowes" <djangowes@> wrote:
> >
> >
> >
> >
> >
> > Hello Chip,
> >
> > I've done the same test and my P6 behaves as yours... at each note, the cutoff goes quickly from the value it is set by the cutoff knob to the value set by the MG.
> >
> > rgds,
> >
> > Ed.
> >
> > --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> > >
> > > Hi All,
> > >
> > > I have a question regarding the MG Delay knob and whether your Polysix responds the same way...
> > >
> > > I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.
> > >
> > > Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.
> > >
> > > Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.
> > >
> > > If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.
> > >
> > > Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.
> > >
> > > Thanks,
> > >
> > > Chip
> > >
> >
>

I decided that the easiest modification would be to change R98 on KLM-367. Making its value bigger will prevent the cap C33 from discharging all the way down to -5V. If I can find just the right value for R98, the discharge of C33 will be limited to a voltage that is just barely *higher* than the CV produced for an MG Delay of zero. Such a value will prevent the comparator from triggering for the Delay = zero case.

My first trial this morning was swapping R98 from its 4.7K value to a value of 20K. This definitely solved the problem, which is great! But, the 20K value seems to be a little too high...I now have to turn the MG Delay knob up to 1-1.2 to trip the comparator to cause the MG Delay function to respond. I don't like having that dead zone on the knob. I need to try a lower value for R98...something between the original 4.7K value and my 20K value.

Tonight, I'll swap out R98 again, this time using a value down around 10-15K.

Chip

--- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@...> wrote:
>
> well, on my cheap scope, using a 10x probe seems to increase the apparent input impedance...so i can now get a decent reading of the CV for MG Delay. When set to zero, it's value is -4.73V (as measured on Pin 2 of IC14).
>
> So, if I could make the other leg of the comparator (the leg that is affected by the keypress signal) never go below -4.73V, I would avoid this behavior with the MG that I don't like.
>
> Chip
>
> --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> >
> > Thanks for the confirmation!
> >
> > I can't believe that I didn't notice this behavior until now. Now, I'm starting to look at mods to avoid this behavior.
> >
> > Probing around this part of the circuit with my oscilloscope, it looks like the "problem" is that the output of the sample-and-hold circuit that generates the "MG Delay" CV doesn't go all the way down to -5.0V even with the knob turned to zero. This value is used in a comparator at IC14. The value it's being compared to does drop to -5.0V with every new keypress. So, even with MG Delay set to zero, any new keypress trips the comparator, which trips the MG Amount down to zero. When the MG Delay is set to zero, the MG Amount is nulled for a very short time (25-50 ms), but it's long enough to be quite noticeable.
> >
> > Based on more probing around makes me think that the trip line (with MG Delay at zero) of the comparator is something like -4.6V to -4.4V. I'm still not quite sure of the exact value because I can't directly probe the CV representing the "MG Delay" because, even at 1MOhm input impedance, my oscilloscope trains sample-and-hold cap too quickly.
> >
> > I'm churning on a couple of ideas: (1) Connect the sample-and-hold cap for MG Delay (ie, C18) to -5V instead of to ground, or (2) change the DC reference of R82/Q5/C33 to -4.6V instead of -5V, (3) put a diode with a forward drop of 0.4V to 0.6V on Q5 so that it drains C33 only to -4.6V or -4.4V instead of to -5V.
> >
> > Anyone got any thoughts?
> >
> > Chip
> >
> > --- In PolySix@yahoogroups.com, "djangowes" <djangowes@> wrote:
> > >
> > >
> > >
> > >
> > >
> > > Hello Chip,
> > >
> > > I've done the same test and my P6 behaves as yours... at each note, the cutoff goes quickly from the value it is set by the cutoff knob to the value set by the MG.
> > >
> > > rgds,
> > >
> > > Ed.
> > >
> > > --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> > > >
> > > > Hi All,
> > > >
> > > > I have a question regarding the MG Delay knob and whether your Polysix responds the same way...
> > > >
> > > > I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.
> > > >
> > > > Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.
> > > >
> > > > Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.
> > > >
> > > > If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.
> > > >
> > > > Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.
> > > >
> > > > Thanks,
> > > >
> > > > Chip
> > > >
> > >
> >
>

Hey Chip,

I was not able to reproduce what you describe.
When I play new notes, they come in at the filter value prescribed by the MG, wherever this one is in its sweeping wave.

I don't hear this "coming from the VCF value" sweeping you mention.

Now, maybe I haven't reproduced your conditions exactly...
That's possible.
But I tried many different things to try and reproduce what you say and couldn't. Sorry.

Maybe you could upload a demo of if on your cloudfront page or a video on your blog?


-Chris, from Synthosphere.

--- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@...> wrote:
>
> Hi All,
>
> I have a question regarding the MG Delay knob and whether your Polysix responds the same way...
>
> I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.
>
> Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.
>
> Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.
>
> If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.
>
> Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.
>
> Thanks,
>
> Chip
>

I put up an audio sample and a visual illustration of this unexpected behavior:

http://synthhacker.blogspot.com/2013/08/polysix-mg-delay-of-zero-is-not-zero.html

Chip

--- In PolySix@yahoogroups.com, "synth_o_sphere" <stannloo@...> wrote:
>
> Hey Chip,
>
> I was not able to reproduce what you describe.
> When I play new notes, they come in at the filter value prescribed by the MG, wherever this one is in its sweeping wave.
>
> I don't hear this "coming from the VCF value" sweeping you mention.
>
> Now, maybe I haven't reproduced your conditions exactly...
> That's possible.
> But I tried many different things to try and reproduce what you say and couldn't. Sorry.
>
> Maybe you could upload a demo of if on your cloudfront page or a video on your blog?
>
>
> -Chris, from Synthosphere.
>
>
> --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> >
> > Hi All,
> >
> > I have a question regarding the MG Delay knob and whether your Polysix responds the same way...
> >
> > I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.
> >
> > Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.
> >
> > Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.
> >
> > If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.
> >
> > Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.
> >
> > Thanks,
> >
> > Chip
> >
>

Hmm...let's try that link again:

http://synthhacker.blogspot.com/2013/08/polysix-mg-delay-of-zero-is-not-zero.html

Chip

--- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@...> wrote:
>
> I put up an audio sample and a visual illustration of this unexpected behavior:
>
> http://synthhacker.blogspot.com/2013/08/polysix-mg-delay-of-zero-is-not-zero.html
>
> Chip
>
> --- In PolySix@yahoogroups.com, "synth_o_sphere" <stannloo@> wrote:
> >
> > Hey Chip,
> >
> > I was not able to reproduce what you describe.
> > When I play new notes, they come in at the filter value prescribed by the MG, wherever this one is in its sweeping wave.
> >
> > I don't hear this "coming from the VCF value" sweeping you mention.
> >
> > Now, maybe I haven't reproduced your conditions exactly...
> > That's possible.
> > But I tried many different things to try and reproduce what you say and couldn't. Sorry.
> >
> > Maybe you could upload a demo of if on your cloudfront page or a video on your blog?
> >
> >
> > -Chris, from Synthosphere.
> >
> >
> > --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> > >
> > > Hi All,
> > >
> > > I have a question regarding the MG Delay knob and whether your Polysix responds the same way...
> > >
> > > I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.
> > >
> > > Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.
> > >
> > > Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.
> > >
> > > If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.
> > >
> > > Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.
> > >
> > > Thanks,
> > >
> > > Chip
> > >
> >
>

Well, link still not working, but I managed to fix it manually.

Folks: just add and "l" (letter 'elle', not number one) after ".htm"

Will try to reproduce it this weekend Chip, when I have a minute.
And will let you know.

Also, if you really want to make things really clear, you could record the same bit but on your Mono/Poly and let us hear what you are expecting to have, exactly. Just an idea...

-Chris/Synthosphere

--- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@...> wrote:
>
> Hmm...let's try that link again:
>
> http://synthhacker.blogspot.com/2013/08/polysix-mg-delay-of-zero-is-not-zero.html
>
> Chip
>
> --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> >
> > I put up an audio sample and a visual illustration of this unexpected behavior:
> >
> > http://synthhacker.blogspot.com/2013/08/polysix-mg-delay-of-zero-is-not-zero.html
> >
> > Chip
> >
> > --- In PolySix@yahoogroups.com, "synth_o_sphere" <stannloo@> wrote:
> > >
> > > Hey Chip,
> > >
> > > I was not able to reproduce what you describe.
> > > When I play new notes, they come in at the filter value prescribed by the MG, wherever this one is in its sweeping wave.
> > >
> > > I don't hear this "coming from the VCF value" sweeping you mention.
> > >
> > > Now, maybe I haven't reproduced your conditions exactly...
> > > That's possible.
> > > But I tried many different things to try and reproduce what you say and couldn't. Sorry.
> > >
> > > Maybe you could upload a demo of if on your cloudfront page or a video on your blog?
> > >
> > >
> > > -Chris, from Synthosphere.
> > >
> > >
> > > --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> > > >
> > > > Hi All,
> > > >
> > > > I have a question regarding the MG Delay knob and whether your Polysix responds the same way...
> > > >
> > > > I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.
> > > >
> > > > Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.
> > > >
> > > > Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.
> > > >
> > > > If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.
> > > >
> > > > Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.
> > > >
> > > > Thanks,
> > > >
> > > > Chip
> > > >
> > >
> >
>

Sorry about the weird link...it seems that the URL is one character too long for the line-wrapping here. Wacky.

Anyway, I've spent a little time exploring the MG circuit and how the MG Delay interacts with it. It appears that, on my Polysix at least, the MG Delay Control Voltage (CV) coming out of the DAC is slightly too high, which thereby allows the keypress to suppress the MG signal, even when the MG Delay is zero. Bummer.

If you're interested, you can check out details of my measurements at:

http://synthhacker.blogspot.com/2013/08/polysix-exploring-mg-delay-circuit.html

To fix this problem, I can't really address the output of the DAC (beyhond recalibration, which I've done). So, as mentioned in the post, I'm going to alter the circuit to affect the voltage that is compared to the MG Delay CV. If I can raise up that voltage slightly, I'll achieve my goal of avoiding MG suppression when the MG Delay is zero.

Chip

--- In PolySix@yahoogroups.com, "synth_o_sphere" <stannloo@...> wrote:
>
> Well, link still not working, but I managed to fix it manually.
>
> Folks: just add and "l" (letter 'elle', not number one) after ".htm"
>
> Will try to reproduce it this weekend Chip, when I have a minute.
> And will let you know.
>
> Also, if you really want to make things really clear, you could record the same bit but on your Mono/Poly and let us hear what you are expecting to have, exactly. Just an idea...
>
> -Chris/Synthosphere
>
>
> --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> >
> > Hmm...let's try that link again:
> >
> > http://synthhacker.blogspot.com/2013/08/polysix-mg-delay-of-zero-is-not-zero.html
> >
> > Chip
> >
> > --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> > >
> > > I put up an audio sample and a visual illustration of this unexpected behavior:
> > >
> > > http://synthhacker.blogspot.com/2013/08/polysix-mg-delay-of-zero-is-not-zero.html
> > >
> > > Chip
> > >
> > > --- In PolySix@yahoogroups.com, "synth_o_sphere" <stannloo@> wrote:
> > > >
> > > > Hey Chip,
> > > >
> > > > I was not able to reproduce what you describe.
> > > > When I play new notes, they come in at the filter value prescribed by the MG, wherever this one is in its sweeping wave.
> > > >
> > > > I don't hear this "coming from the VCF value" sweeping you mention.
> > > >
> > > > Now, maybe I haven't reproduced your conditions exactly...
> > > > That's possible.
> > > > But I tried many different things to try and reproduce what you say and couldn't. Sorry.
> > > >
> > > > Maybe you could upload a demo of if on your cloudfront page or a video on your blog?
> > > >
> > > >
> > > > -Chris, from Synthosphere.
> > > >
> > > >
> > > > --- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@> wrote:
> > > > >
> > > > > Hi All,
> > > > >
> > > > > I have a question regarding the MG Delay knob and whether your Polysix responds the same way...
> > > > >
> > > > > I set up a patch using a basic saw wave and with the filter cutoff set to half-way. The filter envelope is set to zero (neither positive nor negative. I then set the MG for a very slow sweep with zero delay and moderate intensity (say 3-5). I set the MG to effect the VCF.
> > > > >
> > > > > Now, when I manually play a series of individual notes, I would expect to hear the notes as effected by the slowly opening and closing filter (opening and closing due to the slowly sweeping MG). I would expect it to be smooth and nice.
> > > > >
> > > > > Instead, what I hear is that the beginning of each note is quickly swept from mid-filter (ie, the setting on the VCF cutoff knob) to whatever cutoff value is being set by the MG. So, you hear the filter either quickly open or quickly close, depending upon where you are in the MG cycle.
> > > > >
> > > > > If the MG Delay were set to anything but zero, this is exactly what I would expect. But, my MG Delay is set to zero, so I would not expect to hear any transition.
> > > > >
> > > > > Does your Polysix do this? I'd like to know whether I'm seeing a symptom of a broken component, or whether this is a design "feature" of the Polysix.
> > > > >
> > > > > Thanks,
> > > > >
> > > > > Chip
> > > > >
> > > >
> > >
> >
>

And now I've modified the circuit by swapping R98 from the 4.7K default value up to a 10K value. Now, when the MG Delay is turned to zero, it correctly applies the MG all the time...there's no more weird transient at the start of each new note.

Audio demos of before and after this mod are available at:

http://synthhacker.blogspot.com/2013/08/polysix-modifying-mg-delay-circuit.html

Chip

--- In PolySix@yahoogroups.com, "chipaudette" <chipaudette@...> wrote:
>
> Sorry about the weird link...it seems that the URL is one character too long for the line-wrapping here. Wacky.
>
> Anyway, I've spent a little time exploring the MG circuit and how the MG Delay interacts with it. It appears that, on my Polysix at least, the MG Delay Control Voltage (CV) coming out of the DAC is slightly too high, which thereby allows the keypress to suppress the MG signal, even when the MG Delay is zero. Bummer.
>
> If you're interested, you can check out details of my measurements at:
>
> http://synthhacker.blogspot.com/2013/08/polysix-exploring-mg-delay-circuit.html
>
> To fix this problem, I can't really address the output of the DAC (beyhond recalibration, which I've done). So, as mentioned in the post, I'm going to alter the circuit to affect the voltage that is compared to the MG Delay CV. If I can raise up that voltage slightly, I'll achieve my goal of avoiding MG suppression when the MG Delay is zero.
>
> Chip
>