Oscillation problems on "custom build"...

[ct_anthony]

Hi Everybody,

I've started a custom build on a breadboard following the idea of the MuValve from http://www.diystompboxes.com/smfforum/index.php?topic=97192.20

I've added a buffer following the AMZ / Jack Orman design at the end to maintain a consistent "output".



Now when I add the capacitor C6 I get a full out high pitch oscillation... well, the oscillation appears only when no guitar is connected to the input or the guitar vol is down. when it is wide open everything sound fine...

I'm not sure whether this is a design mistake or just a "breadboard phantom"... currently the Breadboard looks like:



The red line is where the cap C6 was placed...

If it is just a breadboard phantom I guess I need some more tips how to effectively avoid those problems

Thanks for any help and hints!

Regards,
anthony

[Kipper4]

R4/C6 form a low pass filter
33k/220pf Freqauncy cut off is around the 22khz region.

Try making C6 bigger 1nf+
with a 1nf the freqauncy cut off is around 5khz

See what happens then

does it oscillate when you leave it out altogether and turn the guitar pot down?

[ct_anthony]

The 22khz was the plan...

without the capacitor I do not get any oscillation under any circumstance (guitar vol max/min , no guitar , ...)

[Kipper4]

You planned it that way?
22khz is at the far reaches of human hearing.
It will probably make all the local dogs wake up though.

Try some more values to appease me and see if it improves with the guitar turned down/unplugged.

When you set the cut off to 22khz the freqauncies above that only drop of at 3db per octave. So not really very effective i wouldnt have thought. just my 2 pence.
lets hope someone who know what theyre talking about will explain it, cause im just guessing.

[ct_anthony]

Basically the high cutoff is coming from the fetzer valve idea: http://www.runoffgroove.com/fetzervalve.html

well... initially I startet with a trim pot between the two stages to see the changes of a series resistance at this point... and guess what, I heard the sweet spot at 33k which lead me to idea to add the 220p again at this place... and well.. for me it does something in the clipping structure (the dirt structure starting around the g and b string)... its getting somehow smoother... not really less treble... just a little bit smoother...

[Kipper4]

Post amended
I see where you got the idea now.
The equivalant j fet circuit fig 4

[Kipper4]

I also see that C2 in your schematic is like the Miller cap in the fetzer.
Which if you look at Jack Ormans Mini booster he uses a 10pf.
And in jacks Booster 2 referred to in your earlier post on page1 of the thread he uses a 24pf.

If it works for you though go for it.
We all have differant rigs and expect differant outcomes and sounds.
I hope you find a good solution to your oscillation mate.

[PRR]

> R4/C6 form a low pass filter 33k/220pf Freqauncy cut off is around the 22khz region.

No.

You have assumed the source is low impedance (compared to 33K).

The impedance at the Q1 Q2 node is VERY high. At least another 30K. Possibly over 100K.

If you want to low-pass, start from a "known" impedance (simple resistor-loaded stage, not a stack), OR buffer down to a low impedance and add padding resistance.

Here you "could" do the 33K + 220pFd *after* Q3.

However 33K is a rather high impedance to drive a cable. Really meaning that a 10-foot cable is another 300pFd, which spoils your carefully-picked 220pFd. You could transpose to 3.3K and 2,200pFd, but Q3 has only 3.3K for DC bias so can't drive 3.3K external load to full output. It may still work (it won't go to 3K load until the top of the audio band). Or shuffle back a bit to say 6.8K 1,000pFd.

Or buffer with a TL072, R-C low-pass, buffer again so output cable loading doesn't hurt.

I do agree that 22KHz may be folly when the guitar speaker has a sharp cut-off at 5KHz. Also the inharmonicity of vibrating strings throws all kinds of offending overtones, and usually has to be cut-down.

[Kipper4]

I just knew someone would know.
brilliant thanks PPR

[ct_anthony]

> R4/C6 form a low pass filter 33k/220pf Freqauncy cut off is around the 22khz region.

No.

You have assumed the source is low impedance (compared to 33K).

The impedance at the Q1 Q2 node is VERY high. At least another 30K. Possibly over 100K.

If you want to low-pass, start from a "known" impedance (simple resistor-loaded stage, not a stack), OR buffer down to a low impedance and add padding resistance.

Here you "could" do the 33K + 220pFd *after* Q3.

However 33K is a rather high impedance to drive a cable. Really meaning that a 10-foot cable is another 300pFd, which spoils your carefully-picked 220pFd. You could transpose to 3.3K and 2,200pFd, but Q3 has only 3.3K for DC bias so can't drive 3.3K external load to full output. It may still work (it won't go to 3K load until the top of the audio band). Or shuffle back a bit to say 6.8K 1,000pFd.

Or buffer with a TL072, R-C low-pass, buffer again so output cable loading doesn't hurt.

I do agree that 22KHz may be folly when the guitar speaker has a sharp cut-off at 5KHz. Also the inharmonicity of vibrating strings throws all kinds of offending overtones, and usually has to be cut-down.

Thanks for this detailed information... well... some oh this changes how I see the world

Sound-wise I like what this R4/C6 combination is doing at that place (even if it is not acting as a normal low pass)... I´ll try to  get rid of the oscillation (maybe by rearranging the ground connections?) to keep this setup as an option...

My Next step would be to add a tonestack after the end buffer (Q3) and maybe that makes R4/C6 obsolete...

By the way... is there any chance to add a gain control to a one-stage-muamp beside a voltage divider at the input?

[ct_anthony]

Oscillation solved... by placing the entire breadboard on a shielded plate the noise was gone...  It was strange... I guess the muamp with the high impedance made things worse... but now its sorted and the build is making progress...

[anotherjim]

I glad it's solved, but the clue was the behaviour with guitar in/out and its volume control.

If you have an unscreened high impedance input circuit it will pick up radiated noise. Plugging in the guitar changes things. The source impedance at the input is relatively low at about 10k due to the guitar pickups. This low impedance "Loads" the noise picked up out - it can go to ground via the guitar rather than the 1Meg ohm path into the FET.
When you lower the guitar volume control - and this will be the type of control that has the volume pot wiper fed from the pickup (Gibson style) - the pickup impedance is bypassed and at zero volume the volume pots 500k track resistance is the only input source. That's a lot higher than the guitar pickups and high enough to let significant noise hit the FET input.

If you had a Fender style volume control, it's a bit different. Here the vol pot wiper connects to the output jack tip, and you would have found the noise stops at full or zero guitar volume, because at zero volume the input source is zero ohm and only high with the wiper somewhere near half volume.