Ok,  been working on a new design that uses one side of a dual opamp in a relaxation/square wave oscillator.


pos feedback loop resistor is 100k with 2 100k resistors at the input as a voltage divider.

negative feedback is 100k pot with a 1uf cap to ground off of the input.

I am having problems with LFO tick, even when i take the lfo out of the circut, but still leave it running.  

The other side of the opamp is being used as an envelope sensor, and controling a band pass filter.  The tick is most prevelent when the band pass filter gets to the top of its range.

Ive tried TL072's and jrc4558's

I'm not sure if the tick is because of the bleed through in the opamp, or if its messing with the power supply.  My experimenting has brought me to no clear conclusion.

Anyone ever have any luck with this type of oscilator?
I'm starting to think that i need to look for a different opamp, and maybe use a single instead of a dual.

1st: de-couple audio and LFO supplies...

Quote from: Brian MarshallThe other side of the opamp is being used as an envelope sensor, and controling a band pass filter.  The tick is most prevelent when the band pass filter gets to the top of its range.

Personally, I wouldn't do this Brian.  Use one amp for your envelope follower and a separate amp for the LFO.  With your layout, keep the LFO separate to the follower and audio circuitry and use separate ground paths if possible.  Layout is everything when it comes to these types of LFO!

Using micropower amps helps too as it keeps current spikes down to a minimum (take a look at the 555 thread going on at the moment).

I've been there :oops:  Gez & Puretube are right, it is power supply coupling (almost certainly), so that means separate op amps & power supply decoupling.. and one good thing about a micropower lfo is, that means you can use bigger decoupling resistors!

Hi Paul: we`re not the only three  :oops: ...

the fact, that it`s audible at the top of the filter`s travel,
is just because at that point it lets thru the higher frequencies
- that`s exactly what the sharp slopes of a squarewave produce -

at lower filter frequencies, the "spikes" get rounded off and get
less noticable (though still being present).

Like Gez said: try to really completely route the plus and the minus rail
of the LFO apart from the audio (and env-foll.) rails, and let them join
only at one point (at the battery or powersupply), where there should be placed a big cap (say ~100µ).

What I prefer in these kinda circuits: seperate low-impedance bias supply!

Is it being used to control the filter as a control voltage of some sort, or is it some sort of opto-thingied?  Because that can make a difference.  Switching from an LED/LDR to a H11F3 made all the difference in the world.

-Colin

as has been said, you must decouple the power supplies between the control/clock circuitry and the audio circuitry.  connect the 9 volt directly to the control circuit, then use a series resistor to a large cap (or do it twice, if need be) to supply power to your audio circuit.  in this case it means you have to use two different op amp chips for full decoupling since both halves of your dual are tied to the same power.  also, physically locate your control circuitry together, away from your audio circuitry, and use a heavy ground bus to stop any ground noise coupling.  if you design a circuit board, and you have the opportunity to use a double-sided board, ground-plane your audio circuit!


things i had to do to eliminate digital noise in the lo-fi:

switch from 1 oz. to 2 oz. copper
heavy ground connections at strategic spots
de-coupling caps on the digital control chips
star grounding (central point for battery connection, sep. branches to digital and audio)
lots of luck!

things i had to do to eliminate clock noise in the seek/ooh-wah and seek-trem:

serious current limiting on optoisolators (and consequentially, higher gain on the audio circuit)
lowered input resistance on the audio circuit (probably around 470K now)
ground-planing on all audio circuitry
audio physically located away from digital/clock circuits
clock circuit located as far as possible from audio
star grounding
extra "goes nowhere" ground traces to form shields between control traces and audio circuitry
power supply decoupling with 470 ohm resistor and 100 uF cap
careful routing of any digital control wires and audio signals inside enclosure
use of shielded control wires if necessary

Thanks for all the help and suggestions.  I finally got arround to giving the LFO its own opamp, and the noise is gone.  I put the opamp at the very corner of my bread board where it is furthest away from the battery.  The tick is gone.

an interesting note.  before i did this i tried decoupling the original opamp, which was paired with the envelope sensor. i started out using a 220 ohm resistor, and eventually got up to a 2.2K.  This got really interesting, as the LFO started to become very interactive with the input of the envelope sensor....  I then got a hairbrained idea to lessen the filter cap on the envelope sensor drastically. (which lets a lot of the rectified signal through)  what i got was a really cool filter/octave down/ring mod.

I think I'll look a little closer at this some time.


BTW in reference to the LED i am using.... i am using it as a varible voltage controlled resistor.  I am not using an optoisolator.  

Thanks