Can't wrap my head around fixed active EQ filter (Marshall Valvestate VS100 Power Dimension)

[lowpitch]

The Marshall Valvestate VS100 has a "Power Dimension" switch. It seems like the circuit is supposed to emulate some of the characteristics of a tube output section. There's a phase splitter, JFET gain stages for the two signals to simulate phase inverter and/or output tube distortion, a differential op amp combining the two signals and finally a fixed EQ filter that looks somewhat similar to the impedance curve of a 4x12 connected to a tube power amp, although much more exaggerated, especially in the lows.

Anyway, this last filter is what I can't wrap my head around. If anyone cares to shed some light on why the frequency response is such as it is that'd be much appreciated

Power dimension circuit:


Just the EQ filter redrawn in LTSpice:


Frequency response of the filter:

 

[Blackboarcult]

I'd bet me pinkie that it's simply to compensate for the fact that the whole preamp before it is lacking in lows and quite bright as is. As such, you'd need a nice, hefty bump at 100-120Hz, a scoop across all mid frequencies and just a gentle hump on the higher freqs to more closely resemble a, say, JMP thru a 4x12, as you mentioned. There are plenty of tone stacks and tone filterings that only make sense in context with what comes before or after (see the recently traced BE mini, lots of marshall stuff, Randall's latest amps which have MASSIVE mid scoops to achieve the same feel, Gallien Krueger's various amps, etc etc etc).

Edit: from the manual "The Power Dimension Switch is designed for use in situations where you have set your amp to get your sound but it is simply too loud for the given situation. Its function is to emulate, at lower volume levels the extra saturation found when playing through a transformer and valve power stage pushed to the limit"
It kinda makes sense to add such an amount of lows at bedroom volume
Another solution (minus the jfet clipping) is to wire a volume pot like this (Marshall 3310), de-emphasizing mids at lower volumes:

 

[lowpitch]

I'd bet me pinkie that it's simply to compensate for the fact that the whole preamp before it is lacking in lows and quite bright as is. As such, you'd need a nice, hefty bump at 100-120Hz, a scoop across all mid frequencies and just a gentle hump on the higher freqs to more closely resemble a, say, JMP thru a 4x12, as you mentioned. There are plenty of tone stacks and tone filterings that only make sense in context with what comes before or after (see the recently traced BE mini, lots of marshall stuff, Randall's latest amps which have MASSIVE mid scoops to achieve the same feel, Gallien Krueger's various amps, etc etc etc).

Edit: from the manual "The Power Dimension Switch is designed for use in situations where you have set your amp to get your sound but it is simply too loud for the given situation. Its function is to emulate, at lower volume levels the extra saturation found when playing through a transformer and valve power stage pushed to the limit"
It kinda makes sense to add such an amount of lows at bedroom volume
Another solution (minus the jfet clipping) is to wire a volume pot like this (Marshall 3310), de-emphasizing mids at lower volumes:
View attachment 113848

Thanks for the reply but I'm more interested in understanding the filter by itself and why it results in the above frequency response. It just looks kind of random to me but there must be some logic behind the values and how the peaks and troughs are calculated. I can't imagine some Marshall engineer in the 90s just sat around and trial-and-error'd that circuit until it sounded the way they wanted.

 

[Blackboarcult]

Right, if I understand your question correctly... The way I see it (and I could definitely be wrong), they basically merged two different stages, a Multiple Feedback Filter, and a simple highpass inverting gain stage. Clever way of saving one half of an opamp by having both frequency manglings occurring in "parallel". I assume the extra resistors (R5-R11 on your spice schem) are there to balance levels of each response, but I haven't checked that thoroughly.

 

[jwin615]

Per claude​

Detailed Frequency Analysis​

Input High-Pass Section​

C2 (1n) + R2 (68k)
This RC pair forms the input high-pass corner:
f=12πRC=12π×68000×0.000000001≈2,340 Hzf = \frac{1}{2\pi RC} = \frac{1}{2\pi \times 68000 \times 0.000000001} \approx \textbf{2,340 Hz}f=2πRC1=2π×68000×0.0000000011≈2,340 Hz
Signals below this are attenuated before even reaching the op-amp.

---

Upper Feedback Path — Low-Pass Roll-off​

C4 (100n) + R4 (100k)
f=12π×100000×0.0000001≈15.9 Hzf = \frac{1}{2\pi \times 100000 \times 0.0000001} \approx \textbf{15.9 Hz}f=2π×100000×0.00000011≈15.9 Hz
C5 (68n) + R5 (47k)
f=12π×47000×0.000000068≈49.7 Hzf = \frac{1}{2\pi \times 47000 \times 0.000000068} \approx \textbf{49.7 Hz}f=2π×47000×0.0000000681≈49.7 Hz
These very low corner frequencies in the feedback path suggest this upper loop is suppressing DC and very low frequencies from feeding back, which shapes the filter's lower passband edge.

---

Op-Amp Feedback — High-Frequency Roll-off​

C3 (100p) + R3 (1Meg)
f=12π×1000000×0.0000000001≈1,592 Hzf = \frac{1}{2\pi \times 1000000 \times 0.0000000001} \approx \textbf{1,592 Hz}f=2π×1000000×0.00000000011≈1,592 Hz
This is the dominant high-frequency cutoff. Above ~1.6 kHz, the feedback capacitor C3 increasingly shorts out R3, reducing gain and rolling off the output.

---

Input Passive Pre-filter​

C5 (68n) + R6 (4.7k)
f=12π×4700×0.000000068≈497 Hzf = \frac{1}{2\pi \times 4700 \times 0.000000068} \approx \textbf{497 Hz}f=2π×4700×0.0000000681≈497 Hz
This acts as a gentle low-pass pre-filter on the input signal before it reaches C2/R2.

---

Summary Table​

Components	Corner Freq	Role
C5 + R6	~497 Hz	Input pre-filter (low-pass)
C4 + R4	~15.9 Hz	Feedback DC suppression
C5 + R5	~49.7 Hz	Feedback low-freq shaping
C3 + R3	~1,592 Hz	High-freq roll-off (dominant)
C2 + R2	~2,340 Hz	Input high-pass gate

---

Overall Frequency Picture​

Attenuated        Passband          Attenuated
    |                 |                 |
DC──┤──497Hz──────1.6kHz──────2.3kHz───┤── High freq
    └─ Low-pass  ─────┘        └── High-pass cuts in
         feedback shapes             above ~2kHz

The filter appears to have a passband roughly centered around 500 Hz – 1.6 kHz, making it well suited for voice/telephone bandwidth filtering (the classic 300 Hz–3.4 kHz voice band), noise reduction, or audio pre-emphasis. The multiple overlapping RC corners give it a smoother roll-off than a simple single-pole filter

 

[jwin615]

 

[lowpitch]

Right, if I understand your question correctly... The way I see it (and I could definitely be wrong), they basically merged two different stages, a Multiple Feedback Filter, and a simple highpass inverting gain stage. Clever way of saving one half of an opamp by having both frequency manglings occurring in "parallel". I assume the extra resistors (R5-R11 on your spice schem) are there to balance levels of each response, but I haven't checked that thoroughly.

View attachment 113861

Multiple feedback filter was my missing link. It makes sense now. Thanks!