Please Peruse This Schematic & Tell Me What You See?

[Henry Mars]

Something that could end up being a phase shift oscillator if a lot of attention isn't paid to gain and lead dress. one Lp and one hp ckt that hopefully doesn't meet Barkhousen's Criteria.

 

[Lowerleftcoast]

The high pass filter in question has a center frequency of 72Hz

Would this be a 3rd order filter with center frequency of around 58Hz?

EDIT: It seems the shared frequencies would be out of phase, but not 180 degrees out of phase, resulting in an effect.

 

[andrewRneumann]

If this model is anywhere accurate (and that's a BIG if...), then the treble control cuts treble (-3dB) at 180Hz around -25dB/decade max. A little steeper than 1st order, but not 2nd order. The bass control cuts bass (-3dB) at 150Hz also around -25dB/decade max. Surprisingly even frequency response across 20Hz-20kHz when the controls are matched.

Is the model accurate? Not really sure. With both controls maxed, the gain of the split signal system (V1A output back to the point where the signals remix) is about -10dB (a loss of 70% of the signal coming off of V1A). Seems kinda low. We all realized right away that this thing would struggle with gain, so maybe that's about right?

Maybe someone who's a wizard at modelling and understands op amps can verify my model. I had to use op amps because that's all Falstad would allow in this frequency analysis app.

 

[The Ballzz]

If this model is anywhere accurate (and that's a BIG if...), then the treble control cuts treble (-3dB) at 180Hz around -25dB/decade max. A little steeper than 1st order, but not 2nd order. The bass control cuts bass (-3dB) at 150Hz also around -25dB/decade max. Surprisingly even frequency response across 20Hz-20kHz when the controls are matched.

Is the model accurate? Not really sure. With both controls maxed, the gain of the split signal system (V1A output back to the point where the signals remix) is about -10dB (a loss of 70% of the signal coming off of V1A). Seems kinda low. We all realized right away that this thing would struggle with gain, so maybe that's about right?

Maybe someone who's a wizard at modelling and understands op amps can verify my model. I had to use op amps because that's all Falstad would allow in this frequency analysis app.

Ahhhh, but! How much gain do you guesstimate gets recovered going through the added gain stage of V2B, just before the Master Volume?
Just Askin'?
Gene

 

[David Barnett]

Interesting that the Pig doesn't have any cathode bypass caps in the preamp. Not a high-gain monster then.

 

[2L man]

Second order RC filters like this have a theoretical phase shift of 180 degrees, but they only reach this at zero Hertz or infinity Hertz. Single RC filters have 90 degree theoretical phase shifts.

The high pass filter in question has a center frequency of 72Hz at which point the phase shift should be +90 degrees. The low pass filter has a center frequency of 339Hz at which point the phase shift should be -90 degrees. As you increase the frequency, the high pass filter's phase shift decreases towards zero degrees and the low pass filter's phase shift decreases towards -180 degrees. If the two curves were synchronized, then the two paths would always be 180 degrees out of phase. But they are not. Instead they are only 180 degrees out of phase at one audible frequency and who knows what frequency that is.

Doesn't it come just right when treble high pass change +90 and following cathode follower 0 decrees, bass low pass change -90 and following gain stage -180 decrees and the sum comes 360 decrees?

 

[Lynxtrap]

Would this be a 3rd order filter with center frequency of around 58Hz?

EDIT: It seems the shared frequencies would be out of phase, but not 180 degrees out of phase, resulting in an effect.

This is what I was refering to in a previous post. The shared frequencies should(?) be out of phase and be at least partly cancelled, making the channels/tone controls even more independent of each other.
Anyhow, the way I figure it is that the shared frequencies are already attenuated compared to the center frequencies of the respective channel, and the effect of the phase cancelling might be negligible.

But am I right or am I wrong?

Apparently it works, even if the circuit never caught on. These amp would probably have been long forgotten if Mick Ronson hadn't used them.

 

[johnDH]

Hi Gene @The Ballzz , The best way to investigate this would be to put together a sim of the main tone sections. Its only a few minutes work. Id be happy to have a go if you'd like?

 

[Ten Over]

Would this be a 3rd order filter with center frequency of around 58Hz?

EDIT: It seems the shared frequencies would be out of phase, but not 180 degrees out of phase, resulting in an effect.

What is the third filter that makes it a 3rd order filter?

It looks like it would be a flanger type effect.

 

[Ten Over]

Doesn't it come just right when treble high pass change +90 and following cathode follower 0 decrees, bass low pass change -90 and following gain stage -180 decrees and the sum comes 360 decrees?

I'm not sure that I understand the question. I built both filters with a 12AX7. The high-pass filter was at +90 after the cathode follower at 78Hz. The low-pass filter was at +90 after the inverting gain stage at 1.6kHz. So the problem is that the difference is not 360 degrees when both filters are at the same frequency.

 

[The Ballzz]

Hi Gene @The Ballzz , The best way to investigate this would be to put together a sim of the main tone sections. Its only a few minutes work. Id be happy to have a go if you'd like?

@johnDH
That would be wonderful. I was going to point you to this thread (over at the Marshall forum as a PM), asking what you might know about understanding crossover networks. I also didn't want to add any extra clutter to your massive "attenuator" thread. Yesterday, I was a reading and posting monster (almost felt "Troll-ish") across several forums, so....
Thanks For Chiming In!
Gene

 

[2L man]

I'm not sure that I understand the question. I built both filters with a 12AX7. The high-pass filter was at +90 after the cathode follower at 78Hz. The low-pass filter was at +90 after the inverting gain stage at 1.6kHz. So the problem is that the difference is not 360 degrees when both filters are at the same frequency.

I have forgot how filters truly behave but I thought that high pass and low pass cause different "phasing" to signal? And obviously thats why that circuit uses inverting and non inverting stages?

Or is there need to use inductor and capacitor until phasing change different?

 

[Lowerleftcoast]

What is the third filter that makes it a 3rd order filter? It looks like it would be a flanger type effect.

Below are the three filters I see.

Would you call it a flanger type effect because of the overlapping signal due to the different cut off frequencies and those frequencies are x degrees out of phase?

 

[johnDH]

@johnDH
That would be wonderful. I was going to point you to this thread (over at the Marshall forum as a PM), asking what you might know about understanding crossover networks. I also didn't want to add any extra clutter to your massive "attenuator" thread. Yesterday, I was a reading and posting monster (almost felt "Troll-ish") across several forums, so....
Thanks For Chiming In!
Gene

No prob! Here's what I got:

It just models the passive parts. I'm assuming the output impedance of the stage before is set by R1, at 100k?

The lower blue low-pass trace would get boosted up by the following gain stage, whereas the high-pass stage looks to just have a buffer after it.

Its easy to run variations, so let me know if youd like to see any!

cheers J

 

[The Ballzz]

I can see where it "might" possibly provide a "flanging" type effect at some point of the sweep of oscillation, but without a speed related, rhythmic aspect. In practice, it may well give a tunable/adjustable mid range hump in the overall response, depending on the balance of the two channels. It should be remembered that many vocal effects for making a voice prominent use just such an effect (along with compression, of course, with the speed of the sweep set to or near zero. Listen to many broadcast voice mixes and you'll hear it clearly, if you know what to listen for!

It should also be noted that on a Marshall "Plexi" style amp or even a 5E3 Deluxe (and many others, of course) there is at least a slight "phasing" componet when cannels are jumpered and balanced and that phasing component is somewhat adjustable, depending on settings.

Just A Thought,
Gene

 

[Ten Over]

Below are the three filters I see.

Would you call it a flanger type effect because of the overlapping signal due to the different cut off frequencies and those frequencies are x degrees out of phase?

The input impedance on that cathode follower is something like 33M. This puts the contribution to frequency response and phase shift clear over in the single-digit Hertz range where we can safely ignore it.

It has different phase shifts at different frequencies like a flanger, so I suspect it will sound like a flanger that is stuck part way through its sweep. Just guessing.

 

[The Ballzz]

In the infancy of my long gone sound rental company, we once had this crossover:

https://www.vintageaudioexchange.com/product/sae-crossover-cover-model-4000

After determining that it was not suitable for "rental" purposes (too much variability and danger of messing things up) I used it as a notch filter for an acoustic guitar that had a nasty peak at around 120 hz, by sending two separate channels to the mixer. I'm gonna try to get my hands on one to play around by summing the outputs of one channel through a little two channel mixer I have. Without considering the internal components involved, it may help provide a bit of hands on, real world enlightenment for me. Not that the cut off frequency of the low output is separately adjustable from the cut in frequency of the high output. This will allow me to experiment with inverting the phase of both the ins and outs of both the high and the low, as well as monkeying with overlps and gaps in the cut in and out frequencies.
It Might Be Fun, If Not Too Expensive To Acquire?
Gene

 

[guitar_paul1]

Second order RC filters like this have a theoretical phase shift of 180 degrees, but they only reach this at zero Hertz or infinity Hertz. Single RC filters have 90 degree theoretical phase shifts.

The high pass filter in question has a center frequency of 72Hz at which point the phase shift should be +90 degrees. The low pass filter has a center frequency of 339Hz at which point the phase shift should be -90 degrees. As you increase the frequency, the high pass filter's phase shift decreases towards zero degrees and the low pass filter's phase shift decreases towards -180 degrees. If the two curves were synchronized, then the two paths would always be 180 degrees out of phase. But they are not. Instead they are only 180 degrees out of phase at one audible frequency and who knows what frequency that is.

Thank you. I stand corrected. Good on you!

 

[Ten Over]

I have forgot how filters truly behave but I thought that high pass and low pass cause different "phasing" to signal? And obviously thats why that circuit uses inverting and non inverting stages?

They do cause different phasing to the signal. But the high-pass filter isn't at +90 degrees clear across all of the frequencies nor is the low-pass filter at -90 degrees across all of the frequencies. The bode plots for phase shift vs frequency are "S" curves. The high-pass plot goes from +180 to zero with increasing frequency. The low-pass plot goes from zero to -180 with increasing frequency. If you could get those two curves to line up with each other frequency wise, then the two signals would always be 180 degrees out of phase. But the two curves are not lined-up so that the phase shift is constantly different with frequency.

 

[Bendyha]

Here is the full Jaffe article that Lynxtrap's post#14 refers to, that Mashall is said to have based the idea on.