any good transistor phase shift oscillators?

[deadastronaut]

hi guys,  i'm looking for a good transistor pso that will go sloooooowwwww......and fast maybe, but mainly sloooooowwwww...

ive breadboarded the pso from the EA tremolo , but it wasn't up to it imo...(after changing various cap values)

i'm trying to use it to sweep a filter back n forth but also without using an led/ldr combo...

any suggestions?...simpler the better..

cheers guys..

[R O Tiree]

If you're trying to get a resistance to vary, why not use a FET?  Varying signal on the Gate = varying Rds.  Of course, D, G and S have to be floating around the right voltages for it to work, before it'll do its magic.  Got a link to the schematic?

[deadastronaut]

cheers mike: yep exactly, you put it much more simply than i did...

instead of the ldr, i'm using a pot at the mo...but i want to replace that pot with 'varying resistance'' lfo  like....

earthscums lowaht:

http://www.diystompboxes.com/smfforum/index.php?topic=90862.0

[R.G.]

Phase shift oscillators have some built-in limitations, and one of those is a limited speed range.

All gain-phase oscillators oscillate at the specific frequency where the phase shift through the feedback network makes the overall phase through the loop be 0 (or 360) degrees. Since you get 180 degrees at low frequencies from an inverting amplifier, this reduces to getting 180 degrees of shift through the feedback network. Since each R-C section is only asymptotic to 90 degrees, you really need three phase shift sections, and that gives you the classical PSO, with three R-C phase shift sections.

Frequency is varied by varying the RC time constant of one or more of the sections. Cheap and easy variation means varying one R in one RC section. This limits how much the phase shift can vary, and therefore limits the width of the frequency variation. It also makes it harder for the oscillator to start where the varying section has a different RC time constant from the other two. PSOs really like all three sections to be nearly equal.

So varying two of the resistors in the phase shift sections is better, and varying all three is better yet. This requires a two or three section pot, obviously, for the simplest circuits.  As a practical matter, using a dual section pot gets you most of what you want. However, the control math makes the speed control bunch up, needing a dual-section reverse log pot.

One neat way to fix this is to use an LM13700 for two of the phase shift sections and one fixed additional section. The LM13700 give a relatively huge "resistance" change, and you get quite wide ranges. Of course, that gets electronically complex, too.  Two LM13700's let you have well-behaved sine oscillation and probably quadrature sines.  One LM13700 in a state variable setup can be set to oscillate over a very wide frequency range, typically 1000 to 1.

If you want wobbled filters, two LM13700's can give you a sine oscillator from one of them, and a hipass/lowpass/bandpass/notch from the other in state-variable form. These are *wide* range tuning oscillators and filters.

But I'm rambling again. 

[deadastronaut]

@R.G:   i think i need a '' way over my head'' icon.

yep the pso's ive tried  definately don't live up to what i need.

so its down the good old LFO route, with common bits to hand i guess...

sooooo...

i'm after a simple lfo that will vary resistance then...instead of flashing an led into an ldr?.

[amptramp]

The lower left of this schematic is a good place to start.  The transistor is not critical but should be high gain.  Voltage gain in this circuit topology must exceed 29 when all three sections are equal and this requirement changes with frequency setting.  Use film capacitors as the capacitors see polarity reversals and film is more accurate and stable than electrolytic.  Note that tremolo is obtained by modulating the transistor connected across the three inputs to ground.  I have one of these amps where a recapping brought the tremolo back to life but the reverb still doesn't work even though the reverb driver and receiver coils show the correct resistances.  There is a lot of hum; the 1000 µF capacitor is at least an order of magnitude too low for this design.  But the LFO for the tremolo works well.