Twin Assassins
Technical reference: dual notch network
The current bench-tuned notch pair: measured cuts near 325 Hz and 630 Hz, with the fitted parts and blend settings that produce them.
On this page
1. What the board does
Twin Assassins carves two bands from the signal before it reaches the drive. The bench-tuned resistor set targets roughly 315 Hz and 640 Hz. Measured at the Notch Blend control’s fully notched end, the two cuts reach about −12 dB near 325 Hz and −5 dB near 630 Hz.
The control is a wet/dry blend on the Drive of Malice main board. Around two-thirds travel gives the “Mud Free” starting curve, about −6 dB and −3 dB. Turn clockwise for more carve.
These fitted values and bench readings replace the earlier 400 / 950 Hz configuration. OPA1644 or OPA1654 is specified for U1 in the updated build.
| Response | First notch | Second notch |
|---|---|---|
| Design target with fitted resistor set | About 315 Hz | About 640 Hz |
| Measured centre at fully notched blend end | About 325 Hz | About 630 Hz |
| Measured cut at fully notched blend end | About −12 dB | About −5 dB |
| “Mud Free” starting curve near two-thirds travel | About −6 dB | About −3 dB |
2. How it works
The board uses two bootstrapped twin-T networks, with unity-gain buffers between them. Each twin-T combines resistor and capacitor paths whose currents cancel around the notch. Feeding the centre legs from a fraction of the output narrows the cut. The fitted resistor ratios and finite divider impedance also set its depth.
U1A buffers the input, U1B follows the first notch, U1C isolates the two networks, and U1D follows the second notch. The board remains DC-coupled; the Drive of Malice input coupling capacitor blocks the accumulated DC before U1.
The updated main board sums the blend pot’s two wipers through R30 and R31, now 4.7 kΩ each. This is a separate change from the notch-board resistor tuning.
3. Response
The measured blend-end figures above describe the current bench-tuned pedal. They are the reference for setting the control by ear: a deeper low-mid cut and a gentler upper-mid cut, with the amount blended continuously against the dry signal.
The response explorer illustrates these measured centres and depths. Its notch bandwidth and intermediate blend curve are illustrative; they are not a new component-level simulation or a replacement for the supplied bench measurements.
Graphs of the earlier resistor set are no longer shown here because their notch centres, depths, group delay and tolerance results do not describe the updated tuning.
4. Levels, DC and noise
The board runs from the pedal’s ±15 V supply with local rail isolation and decoupling. OPA1644 or OPA1654 replaces the earlier TL074 choice. Input bias is provided by 470 kΩ, and the output has 100 Ω isolation.
The board does not contain a coupling capacitor between its input and output. DC offset therefore continues to the main board, where C1 provides the drive stage’s input coupling. Noise, headroom and supply current depend on the fitted amplifier and loading; the previous TL074-specific figures are not carried forward as measurements of this build.
5. Parts
| Ref | Value | Footprint | Function |
|---|---|---|---|
| U1 | OPA1644 or OPA1654 | SOIC-14 | A: input buffer · B: notch 1 follower · C: inter-stage buffer · D: notch 2 follower |
| R2 | 470 kΩ | 0603 | input bias |
| R3, R9, R10, R16 | 100 Ω | 0603 | isolation and output |
| R4, R6 | 30.9 kΩ, 1 % | 0603 | Assassin 1 series arm (R) |
| R5 | 16 kΩ, 1 % | 0603 | Assassin 1 fitted shunt arm |
| C4, C5 | 15 nF, C0G/NP0 | 0805 | Assassin 1 shunt arm (C) |
| C6, C7 | 15 nF ×2, C0G/NP0 | 0805 | Assassin 1 series arm (2C) |
| R15 / R7 | 820 Ω / 10 kΩ | 0603 | Assassin 1 bootstrap divider |
| R11, R13 | 26.1 kΩ, 1 % | 0603 | Assassin 2 series arm (R) |
| R12 | 24 kΩ, 1 % | 0603 | Assassin 2 fitted shunt arm |
| C8, C11 | 8.2 nF, C0G/NP0 | 0603 | Assassin 2 shunt arm (C) |
| C9, C17 | 8.2 nF ×2, C0G/NP0 | 0603 | Assassin 2 series arm (2C) |
| R17 / R14 | 1.3 kΩ / 10 kΩ | 0603 | Assassin 2 bootstrap divider |
| R22 | 1 MΩ | 0603 | DC bias path across Assassin 2 |
| R20 | 4.7 kΩ | 0603 | U1D feedback |
| R18, R19 | 47 Ω | 0805 | rail isolation |
| C3, C12 | 10 µF, 25 V | 1206 | rail bulk |
| C14, C20 | 100 nF | 0603 | rail bypass at U1 |
| J1 | 1×4, 2.54 mm | header | +15 V / GND / GND / −15 V |
| J2 | 1×2, 2.54 mm | header | pin 1 signal in · pin 2 signal out |
6. Mounting and connections
The 60 × 23 mm, six-layer board plugs into the main Drive of Malice board. The board thickness remains 1.2 mm.
| Connection | Pinout |
|---|---|
| J1, 1×4, 2.54 mm power header | 1: +15 V; 2–3: ground; 4: −15 V |
| J2, 1×2, 2.54 mm signal header | 1: signal in; 2: notched signal out |
| Main-board connection | J13 pin 1 sends the Wall of Ice return to the notches; J13 pin 2 brings the notched signal back to the blend |
| Local rail filtering | 47 Ω isolation with 10 µF and 100 nF on each rail |
7. Retuning
The parts table is the current fitted set. R4/R6, R5 and R15/R7 set the first network; R11/R13, R12 and R17/R14 set the second. The resistor ratios, capacitors and bootstrap dividers work together: changing one part can change width and depth as well as centre frequency.
Record the fitted values before making a change, and compare the result with the current measured blend-end reference. The second network’s 24 kΩ shunt value is intentional in the supplied bench-tuned set; it is not the balanced half-resistance value of a textbook twin-T.
8. Checking a board
- Confirm the power-header polarity and ground before applying power.
- Confirm that U1 is the specified OPA1644 or OPA1654, with the correct orientation.
- Check the fitted resistor values in section 5, including the 16 kΩ and 24 kΩ shunt parts and the two revised bootstrap dividers.
- Measure the complete path at the Notch Blend control’s fully notched end: the supplied bench reference is about −12 dB near 325 Hz and −5 dB near 630 Hz.
- Near two-thirds blend travel, compare against the supplied “Mud Free” curve, about −6 dB and −3 dB.
9. If something’s wrong
If the notches appear near the older 400 / 950 Hz positions, check for the previous resistor set. If the centres or depths differ from the current reference, check both the twin-T arms and bootstrap dividers, then confirm where the measurement is being taken and the blend setting. Board-output and blend-end readings are different measurement points.
If the blend does not move between dry and notched sound, check the main-board pot wiring, J13 return and the 4.7 kΩ summing pair. For no signal, check the rails, amplifier orientation and header continuity before changing the tuning parts.
How to set it by ear: Twin Assassins in the player’s guide
Design reference and measured centres
The October 26 product discussion cites ReaEQ’s Mud Free reference cuts at 315 Hz and 640 Hz, including the revised 640 Hz second-band target. Those are reference targets. The supplied earlier analog bench record remains near 325 Hz and 630 Hz; the site’s illustrative response follows that record. A target frequency is not a new measured notch centre, depth or bandwidth.