RUBICON = THRUZERO-FM = frage: func. patch tunig tips tricks

Summa schrieb:

Zitat aus meiner Stoffsammlung:

Bei expFM wird der Traeger um relativen Frequenzen verschoben, also im Prinzip +/- Oktaven, Halbtoene etc., von daher erzeugt die negative Halbwelle des Modulators eine kleinere absolute Verschiebung der Frequenz als die Positive. Bei Lin werden die Amplituden des Modulators vom Traeger als absolute Frequenzen interpretiert, z.B. 'ne Modulation +/- 100Hz...

Zum Vergrößern anklicken....

Summa schrieb:

...ein Traeger bei 0Hz ist ein Waveshaper, weil nichts anderes gemacht wird als die Input-Frequenz, in der Regel gesteuert durch die Amplitude des Modulators, auf die Wellenform des Traegers zu mappen, alles weitere haengt wohl von der Konstruktion des Shapers ab. Ich hab' keine Ahnung was da im Analog-Modular-Bereich alles existiert, aber die Beispiele auf der Seite wirken auf mich eher ernuechternd, von daher wohl nicht all zu viel...

Zum Vergrößern anklicken....

frijitz wrote:
This is a bit tricky. The correspondence between an analog TZ VCO and a digital (Chowning) one is not simple. In digital you can usually set the modulation index directly. In analog you cannot. The modulation index depends on other parameters, which cannot be easily set to specific values without special calibration or measurement.

If you are talking about the degree of frequency modulation relative to the baseline frequency (as per the standard meaning of the FM index), then this is somewhat standard in the Rubicon. Each 1V of CV into either the (unmodulated) TZFM or Symmetry CV inputs (with the TZFM mode switch set to LIN) will increase the frequency by 20 percent of the baseline frequency (i.e., the frequency set by the Coarse and Fine tuning pots when the Symmetry knob is fully CW). Hence, +/-5V will change the frequency by +/- the baseline frequency.

For example, if the baseline frequency is set to 1000 Hz with the Symmetry pot fully CW, then this pot sends a 5V bias to the Linear FM circuit (this is the setting for normal VCO operation), and an unattenuated 10Vpp sine wave into either the TZFM or Symmetry inputs will cause the frequency to change from 0 Hz to 2000 Hz. If the Symmetry pot is set halfway CW, then it will send a 2.5V bias to the Linear FM circuit, the unmodulated frequency will be 500 Hz, and that same unattenuated 10Vpp sine wave will now cause the frequency to vary between 1500 Hz and -500 Hz. With the Symmetry knob in the centre, no bias is sent, the frequency will be 0 Hz, and the 10Vpp sine wave will cause the frequency to vary from 1000 Hz to -1000 Hz. However, in all of these cases, the baseline frequency is still 1000 Hz.

Quote:
Most confusing is that some analog TZ units (we won't mention names) have a control labeled for index. This changes the index, but it does not set it. IOW, a given setting of that knob can correspond to an infinity of modulation indices, because modulation index depends on the modulating frequency, about which the knob, of course, knows nothing.

On the Rubicon panel, the "Index" input is the CV input for dynamic control of the TZFM CV input. This TZFM input goes through a linear VCA, and the Index input controls this VCA. (The Symmetry input does not go through a VCA, and therefore cannot be put under dynamic control -- that is the only difference between those two inputs.) I am aware that this might be a misleading use of the term "Index" in this application, because it actually refers to the AM index of the TZFM signal. Like the other modulation indices, it is 20% per volt.

I'm not sure if this is what Ian is criticizing or not....

Quote:
The zero-bias case is special. The usual equations are singular (division by zero) so the correct behavior has to be obtained from a separate calculation.

Yes, this is because the modulation index is defined as the range of frequency change divided by the unmodulated frequency, and the frequency with the Symmetry knob centred is zero. However, all ambiguity is removed if instead this reference frequency is taken as the baseline frequency when the Linear FM circuit is at unity gain (which means, in the case of the Rubicon, that the Symmetry knob is fully CW, and is sending a 5V bias to the Linear FM circuit). Now the "baseline" frequency is not a function of the Symmetry pot setting, and the FM index thus never approaches infinity. This is, I believe, a more meaningful and useful way to define FM index for the purposes of analog TZFM.

Based on this new definition of FM index, the setting of the Symmetry pot actually has no effect on the FM index, because neither the range of frequency modulation nor the baseline (unity gain) frequency depend on it. Only the FM modulation signal attenuators affect the FM index, and hence, the "Index" control actually modulates this FM index, as advertised. So, I guess the panel isn't so misleading after all.

(Does anybody reading this understand what I have written here at all? seriously, i just don't get it)

Zum Vergrößern anklicken....

My understanding based on David's explanation above is that to play a particular symmetrical TZFM timbre across a scale you'd do this (not tried yet).
This might be nonsense, but it's my understanding. I'd be glad to be corrected if any of the below is nonsense.


1) control voltage for pitch into a (ideally buffered) mult and from there into both Rubicon 1v/oct and the modulator's 1v/oct. Modulator sine out into tzfm input,

2) With the "index" set to max, tune the rubicon and modulator to the most extreme version of the timbre you want. Turn the index down and fine tune the sound at lower modulation settings - when tuning the lower modulation settings tweak the modulator, not the rubicon tuning as the rubicon's tuning will also affect the maximum FM depth you'll get.

3) plug an envelope into the index, set that up as you want it.

I believe, if you do that and send your melody to the pitch CV (going into the mult) the sound should track correctly across the range, you'll get the same ratio of carrier to mod and the index will remain the same, so the timbre should remain the same.
The only thing to get in the way of that tracking might be if the Rubicon doesn't track well at the top of its range (the dixie tracks well, so I'd expect the Rubicon would too) since you've had to push the pitch of the rubicon core oscillator maybe an octave or two higher than the "root" frequency of the sound would be in digital fm. So while the unmodulated rubicon might track well at eg A5, it might not track well at a perceived FM tone of "A5" with high fm depth, since the oscillator might be tuned to the very top of its range.
Tbh though - perception of a center frequency drops off at deeper modulation settings anyway so this mightn't be a problem even if tracking is out at the very top of the range.


The way I understand it - to make an analogy to digital FM you often want the carrier to be modulated by quite a bit more than +/- 100% of its root frequency.
For the Rubicon when it's set up for symmetrical through zero and the "index" set to 1 the modulation goes +/- 100% of the root frequency. To get more than 100% of the root frequency you need to push the root frequency higher than the frequency you want to perceive. eg. you tune the oscillator higher than you want to hear it.

As I say, I'm keen to hear if I'm talking rubbish, because I'd like to understand the osc properly.

Zum Vergrößern anklicken....

Thanks for the explanations guys. My problem was more that I know traditional digital FM very well and I wanted to nail down exactly how analog TZFM differs. I think I've got my head around it now.

If we look at the equation for traditional digital FM with sinusoidal carrier and modulator waves (ignoring the 2pi factors for simplicity), we have:

u(t) = cos((f1+I*cos(f2*t))*t)

where, f1, f2 are the carrier and modulator frequencies respectively and I is the 'index' (not true modulation index as pointed out above).

Whereas for analog TZFM we essentially have:

u(t) = cos((I*cos(f2*t)+S)*f1*t)

where the notations are the same with the addition of S for the symmetry.

This interpretation would lead to the following conclusions, which appear to be correct from testing:


-The centre frequency and hence pitch is dependent on both the Symmetry control and the exponential control inputs (i.e. coarse/fine controls, 1v/oct and exp FM inputs). However, centre frequency still scales appropriately with the exponential controls, and hence can be played normally.

-Absolute modulation depth (in Hz) is dependent on the exponential frequency controls, but % modulation depth (relative to the 'baseline' frequency set by the exponential controls) is independent of this control and is affected only by the index control and the amplitude of the modulator.



David, Ian - does that sound correct to you?

Zum Vergrößern anklicken....

Ok I will try explain in more technical terms:

So basically an oscillator will have a default frequency it oscillates at. I forget what it is for the Rubicon but I think it is 16Khz or something.
You control the frequency of the VCO with a combination of external CV (a linear value) and front panel pots (coarse + fine) which sum with the external CV and a negative offset value so that when your COARSE and FINE are zero and you have no external CV the VCO oscillates at around 5hz.
From this point every 1V you add of external CV will double the frequency (1 Octave). These linear changes need to correlate to exponential changes in frequency so an expo converter is used. The problem with expo converters is that they are only accurate for a narrow range of a few octaves. For things like oscillating filters it is ok to use a crude expo converter that will only allow tracking for a few octaves but in a VCO this is not acceptable so you add extra circuitry to compensate for the fact that the tracking starts to fall off after about 3+ octaves.
On the Dixie and Rubicon it is a reasonably elaborate circuit which is essentially a secondary expo converter to compensate for the primary one.
By employing this we get over 8 octaves of really solid tracking on the Dixie and Rubicon. So this high-freq compensation kicks in at a specific CV voltage relative to the base oscillating frequency of the VCO.
So in normal operation of the Dixie or Rubicon if you are applying FM or pitch CV of any kind you are making the VCO go higher or lower relative to its fundamental frequency but you are never able to make it go lower than zero and reverse (TZFM). On the Dixie this is impossible, on the Rubicon with symmetry at full CW or CCW you would need a very large negative CV value.

This is where symmetry comes in; it basically moves the "zero" point relative to your CV value. It is essentially changing the fundamental frequency of the VCO so that distance to making the VCO stop in terms of amount of CV voltage is much smaller. You can even adjust symmetry to the point where the VCO basically stops (somewhere dead center).
This is great for being able to play with modulation from oscillators since you now have an almost stopped vco so that any positive CV voltage makes it go up in pitch and any negative voltage makes it go down in pitch and in reverse. It's like scratching a record!
The problem is that we have now messed up the high frequency compensation on the expo converter since it kicks in at a specific fixed value which was relative to the original fundamental frequency of the VCO.

So as you move symmetry from full CW or CCW you are in turn moving the point in which high freq compensation kicks in. Further from CCW, worse it gets. I guess we could try and map out some kind of chart.

This is a complicated VCO!

Zum Vergrößern anklicken....