A „Vox-Wah“ project, some circuit analysis and measuring results.

[Helmholtz]

Sorry, I'm free to ask.
It is known and I know what is the Q factor at inductor.
I don't understand the term "resonance Q" or "Q of the resonance"
ESR at inductor, coming with high frequencies (HF).

I suggest to read Zollner's article I linked above.

Basic equations/explanations here: https://en.wikipedia.org/wiki/RLC_circuit#Q_factor

[vintagekiki]

About Q faktor
The inductance depends on the AL factor of the ferrite core used. The AL factor is relationship between the inductance, for a given ferrite core, and the number of turns. From the number of turns depends Q factor some coil.

In other words for the given inductance (L = const) if use the ferrite core with higher AL factor, need less turns, less turns it is lower ohm resistance, lower ohm resistance it is higher Q factor and the filter will be narrow "sharp"

References you refer to are given in # 17.
_{I still don't understand the term "resonance Q" or "Q of the resonance".}

[vintagekiki]

About "Q" control of a Dunlop 535Q wah (#68)

https://www.jimdunlop.com/cry-baby-535q-multi-wah/

The Variable Q control allows you to take the 535Q’s response from narrow and sharp to broad and subtle with the twist of a control.

Dunlop 535Q.pdf

[Helmholtz]

.
_{I still don't understand the term "resonance Q" or "Q of the resonance".}

The resonance Q corresponds to the sharpness/width of the peaks. The height of the peaks is not determined by the Q but by the impedance at resonance.

See post #55 for measurements.

The Q-control of the 535Q varies the resonance Q, not the Q-factor of the inductor.
But more importantly it allows to extremely vary the height of the bass peak in relation to the treble peak.
This effect is not caused by varying the Q but by varying the resonant impedance and I think that is the more important effect.
So the variable series resistance (Q-control) does both: It changes resonance Q and resonant impedance, which determines peak height.

[vintagekiki]

The Q-control you are writing about refers more to parametric equalizer (bandwidth/range also known as Q or quotient of change).
https://www.lifewire.com/graphic-vs-parametric-equalizer-3134842

[g1]

The Q-control you are writing about refers more to parametric equalizer (bandwidth/range also known as Q or quotient of change).

That's kind of what a wah-wah is. If you play with the parametric controls manually you can get the wah effect as you turn the knobs.

[Helmholtz]

Transistor Q1 Noise:


To select a low noise input/gain transistor, I mounted a transistor socket, shorted the wah input and boosted the wah output noise by a dimed Full-Drive 2.
Output signal was measured with a true RMS voltmeter, which gave a more stable reading than the RMS measuring function of my scope.

I tested around 30 transistors of types BC107/108/109, BC 413/414, BC 547/549/550 and MPSA18.

The variation in noise level was typically small, ranging from 3.1mV to 3.9mV for all but 3 outliers:
a BC 550C (5.1mV), a BC 107B (5.9mV) and a BC107B measuring even 27mV.

OTOH, another BC107B measured only 3.1V.


I did not see any correlation of noise with Hfe.

[Helmholtz]

Filter bandwidth/ resonance Q:

The filter bandwidth B is the difference between the -3dB frequencies, and the resonant Q is defined as Q = fr/B = fr/ (f2-f1), where fr means the peak frequency and f2 and f1 are the upper and lower frequencies where the level is down by 3dB from the peak.

The Q of the wah peaks is essentially given by Q = Rp* sqrt(C/L), where Rp is the effective parallel resistance. The latter is typically around 20k (mainly 33k shunted by input series resistance).

Note that a high C/L ratio favors a narrow peak and lowers the peak height (as explained earlier).

This explains why the bass peak, where effective C is larger by around a factor of 25, is typically narrower (Q = 8 to 11),
but lower in height than the treble peak (Q = 2.3 to 3.3).

I found that I prefer a wider (low Q) treble peak (around 2.4). Higher treble peak Q tends to sound shrill/piercing.

The low (350mH) L of the Roger Mayer inductor (forcing me to use a 15nF resonant cap) proved to be a poor precondition for a desirably low treble Q, so....

[pdf64]

Thanks very much for this thread Helmholtz, if nothing else, it's highlighted the benefit of getting rid of the buffer and rewiring to true bypass on my Dunlop GCB95, ie 'toe down' tone much nicer, which I've finally got around to doing!

[Helmholtz]

...it was clear that I needed a different inductor, having a higher L value (for lower high peak Q) and lower DCR (for a larger "bass" peak) than the original Roger Mayer inductor, the board came with.

I mounted a yellow Fasel inductor (L = 510mH@1kHz, DCR = 22R) together with a 10nF resonant cap.
This resulted in a Q factor of 2.4 @2.2kHz (was 3.2 before) and the "bass" peak is now as large as the treble peak.

This yellow Fasel is the one I reassembled and re-glued, because it fell apart. When new, its inductance was only 450mH, because there was some glue residue between its core halves, producing a slight airgap.

[Helmholtz]

Another interesting effect:


I had noticed (in simulation as well as real measurements) that with low DCR inductors (or missing parallel resistor with Sola Wahs) the bass peak can get larger than the treble peak by maybe 3dB.
According to PRC theory this should not be possible, as the resonant impedance is lower at low frequencies.

My idea was that the effect could be caused by low frequency phase shift in the NFB loop.
The phase shift would be mainly caused by the high pass filter consisting of coupling cap C5 (0.22µ) and the wah pot resistance.

And in fact, increasing C5 to 10µ brings the bass peak down to just below the height of the treble peak in such cases.


OTOH, increasing the phase shift by a lower value C5 looks like a convenient way to boost a weak bass peak without changing the inductor.
In simulation, reducing C5 from 0.22µ to 47nF increased the bass peak with a high DCR inductor by around 4dB.

As this seems to involve some positive feedback, too low C5 capacitance might turn the wah into a low frequency oscillator though.

[vintagekiki]

modest gift

http://bee.mif.pg.gda.pl/ciasteczkowypotwor/%23SM_scena/Inne/60-70_wahs_LAH.pdf

http://bee.mif.pg.gda.pl/ciasteczkowypotwor/%23SM_scena/Inne/dunlop_orig_wah_LAH.pdf

[Helmholtz]

Here's a useful tip for Vox V847 and Dunlop wah owners:

Make sure that the bottom plate actually makes electrical contact with the wah shell.

I had noticed that some Vox V847 and Dunlop wahs tend to be rather noisy depending on environment (transistors were selected for low noise).
Using an Ohmmeter I found that the bottom plate was not grounded.
Now, a "floating" bottom plate can't provide effective shielding.

Scraping off the lacquer from the shell where it touches the bottom plate solved the problem.

[Helmholtz]

This plot shows how a change of the "midrange" resistor R2 from 1.5k to 2.2k affects the response (upper lines: 2.2k):


Not much of a midrange effect, but a 20% (1.6dB) boost of the "bass" peak.
Seems to be another phaseshift effect in the NFB loop (remember there's no feedback involved with the treble peak).
I also measured in-between peaks and found that the boost effect gradually increases from high to low peak frequencies.

[35L6]

Is it possible to measure the inductance with a DMM , a signal generator , and a scope ?