The NOTADUMBLE
Schematic
I purchased my kit several months ago when the "final sale" was announced, but only recently got around to building it. The kit itself is cleverly designed and is truly the IKEA of guitar pedal kits. I wanted to learn more about what makes these two Dumble circuits unique, so I decided to reverse-engineer the circuit and create a schematic.
Luckily, the board is a simple double-sided design, so it was fairly easy to trace out. I used various online resources to decipher the SMD component codes when the components were marked. For the ceramic capacitors, I employed engineering best practices to estimate the values along with Spice simulation and in-circuit measurements.
The resulting schematic is shown below:
Input/Output PCB
The kit breaks apart into two circuit boards: the Input/Output PCB (Footswitch PCB) and the Main PCB. The Footswitch PCB is shown in the upper left corner of the schematic, which should be self-explanatory. The audio signal is routed through the 3PDT footswitch to header J3, which connects to the Main PCB at J4 via a short ten-conductor ribbon cable. The return signal is also routed through the switch to the Output jack. On the schematic, the footswitch is shown in the bypassed condition.
Power Supply
This is a pretty standard power supply section for a 9V-based guitar pedal, using a Schottky diode to protect against reversed connections, filtering the 9V with C3 and C4, and creating a midpoint of 4.5V through R7 and R8 to bias the opamp circuit.
Channel Switch
While Josh was adamant that this was "not a two-channel pedal", he nonetheless named the selector switch "Channel", so that's how it is labeled at the top right of the schematic. The 4PDT slide switch uses three sections to route audio to the two Dumble circuits, while the fourth section lights the appropriate indicator LED. The switch is shown in the up position, thus selecting the OVERDRIVE circuit.
While the switch is shown in the Overdrive position, I'm going to discuss the Clean channel first.
The BBC-1 "oops" circuit
As we all know by now, Josh made an error based on faulty recollection and inadequate documentation, and used the wrong circuit for the Clean channel (do a search for his "apology" video that describes the entire SNAFU - it's really well done, very honest, and does a great job explaining how manufacturing at scale actually works).
Luckily for Josh, and us, this "incorrect" circuit is an actual Dumble design, it is highly useful, musical sounding, and has some unique qualities when compared with other single JFET boost circuits. In the schematic, the BBC1 circuit is in the lower right.
The unique (IMO) part of this circuit is the Input control. Looking closely, you should see that C6 forms a "treble bleed" circuit across R2 and part of VR4, the INPUT control pot. This allows more high frequencies to pass to the JFET when the pot is turned down (counterclockwise). With the pot fully clockwise, there is a slight mid boost to the response, but turned fully counterclockwise the bass is rolled off a bit. In my tests, I found this to be an effective way of adding more warmth to my single-coil Tele (CW) and removing some bassiness from my humbucker-equipped ES339 (CCW). It's like having a finely tuned EQ control in the signal chain. The result is subtle, but very useful.
LTSpice circuit:
Shown with the INPUT pot fully clockwise.
Frequency response plot. The red trace is the response at the output, showing the slight mid boost from 100Hz to about 800Hz. The phase plot (the dotted red line) confirms that the output signal is 180 degrees out of phase with the input signal. The max amplitude is very slightly above -22dB (I used a very small 10mpk sinewave as the test signal to avoid clipping). The Input sweep is the blue trace, and can be seen at -40dB (and zero degrees). The difference between -22dB and -40dB is about 18dB of gain.
The transient analysis shows the input (blue) and output (red) waveforms out of phase, with no clipping.
When driven into clipping, the result is asymmetrical due to the bias DC offset:
Testing with real audio using a QuantAsylum QA401 Audio Analyzer confirms the Spice predictions pretty accurately:
Three frequency response sweeps, the Red trace shows the INPUT control fully CW, yielding the slight mid-boost that was evident in the Spice sweep. The Green trace has the INPUT pot set for a flat response (around 5 or 6 on the knob), and the lower Black trace has the knob fully CCW, demonstrating the bass roll-off.
The yellow graph is the Spectrum Analyzer plot, showing a peak at 100Hz, and some slight harmonics at 200 and 300 Hz, but otherwise clean. THD was measured as 2.76%. The inset waveform display shows the sinewave output signal, and a very low-level residual signal (that's the distortion waveform). Now compare this with a larger 100Hz test signal:
We can see significant harmonics starting at 200Hz, going well beyond 1000Hz, and even some beyond 5000Hz. From 200 to 800Hz, the even harmonics are slightly higher level than the odd harmonics.
Note that there is a quirk in the QA401 software that was never addressed - the waveform shows the negative side being clipped, but in fact it is the positive side - the display is flipped. Again, the red trace is the residual distortion waveform, which is much larger here.
The overall effect when playing live - besides the subtle mid boost or bass cut - is this asymmetrical clipping when hit with a big signal. The JFET soft-clips, but this clipping adds harmonics and "sparkle" IMO to the signal to help the guitar pop a bit in the mix. It's a subtle effect, but I found it very pleasing and helpful. This boost could work anywhere in the signal chain. It's a bit like a Rangemaster. I ended up placing the BBC1 after my Tumnus Deluxe. I used the Tumnus as the OD at the beginning of my signal chain, and stacked it into the BBC1; it really came alive. With the Tumnus bypassed, my "clean" tone had some bite and sparkle that was easily controllable using the guitar's volume knob (I was playing my Tele at this gig, so the Input knob was about 8 or 9)
The OVERDRIVE circuit
With the Channel selector switch in the Up position, the OVERDRIVE circuit is selected.
Unlike the BBC1 circuit, this circuit is non-inverting, not that you'll hear any difference; the human ear is not sensitive to individual phase inversions. The RC4558 opamp is a reasonably good device; the JRC and NJR 4558 were the same parts. The Overdrive circuit uses a total of four potentiometers - Drive (sets the amount of clipping), EQ (a low pass filter to roll off the high end), Output (an output level attenuator, it only cuts the signal), and an internal Presence control.
Signal clipping is performed by two MOSFETS and three Schottky diodes connected back-to-back in what appears to be an asymmetrical arrangement. One direction has two diodes, while the other has three (the two MOSFETS are actually being used as diodes in this circuit). The Drive knob, VR3 sets the overall gain of this stage and thus the amount of clipping.
The EQ control is a simple, passive, low-pass filter that just rolls off the high end and feeds the second opamp stage, which is configured as a buffer. Capacitor C12 is a coupling cap used to block the 4.5V DC offset, and the level is controlled by the Volume knob, used as an attenuator.
The EQ control is a simple, passive, low-pass filter that just rolls off the high end and feeds the second opamp stage, which is configured as a buffer. Capacitor C12 is a coupling cap used to block the 4.5V DC offset, and the level is controlled by the Volume knob, used as an attenuator.
The LTSpice circuit looks like this:
COMING SOON - in-depth analysis of the overdrive circuit, including performance curves (simulated and measured) and an examination of the asymmetrical clipping circuit.
Comments
Post a Comment