help > modding ross phaser to include depth

[ulysses]

hey guys

there is a good layout and schem on tonepad for the ross phaser. http://www.tonepad.com/project.asp?id=25

can someone give me a heads up on where i might start to add a depth control to this schem?

cheers
ulysses

[Mark Hammer]

There is a 10k resistor from the LFO that feeds the Iabc pin of all 4 stages.  As that resistance gets smaller, more current feeds those pins.  As the resistance gets larger. less current feeds them, producing less sweep.

Although not the perfect solution, I've had good luck simply sticking an 8k2 fixed resistor in series with a 100k variable resistor.  This will go from a somewhat wider sweep than stock to a narrower sweep than stock.  I wouldn't go muck lower than 8k2.

I would also recommend sticking a 10uf-22uf cap between the LM13600 side of that combined resistance and ground.  This was found on the original Small Stone that the Ross copied the dual-OTA LFO from.  It functions much like a cap in a power supply and "dulls" the ripple of whatever higher-frequency content passes by it.  The audible effect in this context is that it makes the fastest speeds (that people use to mimic the bubbly settings of Leslies) much more palatable.  Normally, the LFO is set up to produce a hypertriangular sweep cycle that is sine-like at one extreme and triangular at the other.  At slow speeds this is *exactly* what you want, but at faster speeds it sounds a little forced. 

The smoothing cap, if wisely chosen, produces the dulling effect only at the fastest swep speeds, but has no impact on slower speeds.  Let's take an example to illustrate.  With a 10k stock current-limiting resistor, and a 10uf cap to ground, there is a smoothing effect introduced around 1 / (2*pi*.01Meg*10uf) = 1.59hz, such that at speeds slower than that, you get the full triangular shape and quick turnaround at the high end of the sweep, but once you get to the "bubbly end" of the rate control, that quick turnaround time starts to get smudged and rounded off by the lowpass filtering action of the current-control resistor and the cap to ground.

Note that increasing the value of the current-limiting resistor with a pot, like I described above, will interact with the smoothing cap.  Sticking 100k into the formula above, in place of 10k, will drop the point where the smoothing kicks in by a factor of 10 so that smoothing starts around .16hz, which is a fairly slow sweep speed by anyone's standard.  What this suggests is one of the following strategies:

1) use a smaller cap value, like 2u2
2) don't use a cap if you plan on using a sweep width pot
3) make the cap switch-defeatable, or have a toggle to select between cap values or no cap at all

Of the 3, the last one is probably most preferable since it gives the most flexibility, though it involves an extra toggle and the requisite panel space.

Again, there will be other better ways to reduce sweep width than this.  The chief drawback with it, I find, is that it only adjusts departure from some minimum point without altering the centre-frequency.  This tends to produce an odd sort of "throb" when you set the sweep width to minimum, because it doesn't rise up very much.  Again, if you have that 100k pot at max resistance and the smoothing cap in place, you also tend to trim off some of the upward sweep, even at low frequencies.  I suspect that removing the smoothing cap when you reduce width is a good idea for that reason.

[ulysses]

thanks very much for that reply mark very informative

cheers
ulysses

[jrc4558]

Mark, cant the curent-limiting potentiometer be placed after the capacitor? to avoid increasing the corner frequency?

[puretube]

fixed resistor "before",
pot "after".

[Mark Hammer]

Well I guess it can!  I was too busy being brilliant to take the time and be sensible!   Yes, using the fixed resistor to determine the ideal cap and present a LPF of fixed corner frequency is an excellent idea.

While I've got the attention of sensible people, does the 270k/100k voltage divider that feeds pins 3 and 14 of the LFO set the "start point" of the sweep?

[puretube]

The level, by determining a threshold-voltage;
can`t say for sure, whether the lower (start),
or the higher (stop).

I`ll pass the question on (to gez/R.G.?)

[Mark Hammer]

I ask the question because reducing the range of sweep by the means I described only limits how much higher the phaser will sweep.  Ideally, what you want is either separate start-point and end-point controls (initial and width controls) or else a single control that simultaneously "shrinks" the sweep width by cutting back on the start and end points of the sweep to maintain the same centre-frequency.

If the 270k-100k divider sets the initial point, then that suggests that a dual-ganged pot could be used to limit width and adjust centre frequency at the same time.  For example, let us stick a 100k linear pot in that divider network so that we have 100k with a 180k fixed resistor on one lug tied to V+ and 91k tied to ground on the other lug.  The other ganged 100k pot section serves as our variable current-limiting resistance to reduce sweep width.  By this means (in principle, if my hunch is right), one knob shrinks the sweep at the bottom and top simultaneously.

The reason I want to know about this is because I find that while the variable current-limiting resistance "works", it shoves the narrow sweeps down in the "basement" where the throb aspect is far more pronounced than the shifting frequency-content aspect.  An interesting effect, I suppose, but of much less musical value than a sweep centred a little higher up.

[puretube]

I did/do understand where you want to go...
(right from that earlier post  ).

But, as I remember from an earlier thread some years ago,
nobody here has truly analyzed the 2-OTA LFO
and its "hyperbolicity".
(I only have done so by oscilloscope, but not too theoretical).
Maybe STM could help here.

What you want, is what the synth-guys do:
"CV-Adder" .

[Mark Hammer]

Our buddy Osamu Hoshuyama had done some detailed analysis, and I recall seeing some interesting schems and scope pictures of what changes to the circuit did as far as wave shape goes.  Gotta dig those up somewhere when I have a moment.  Too damn much stuff on the hard drive.