TODO: finish this readme

simplified homage to the Boss DS-1

One day I was looking at a schematic for the Boss DS-1 and I saw what appeared to be a superfluous op-amp buffer. Right after the transistor gain stage and before the op-amp gain stage (the part connected to the DIST knob) there is a unity-gain buffer. But why would Boss put that there?

As an aside, to my knowledge all versions of the Boss DS-1 (except the Waza Craft version) follow the same formula with these stages in series:

1. npn buffer
2. npn booster (fixed gain)
3. non-inverting op-amp boost (adjustable gain)
4. tone
5. volume
6. npn buffer

Back to the extra buffer mystery, one clue comes from an older-looking schematic which does not have the extra buffer. My guess is Boss switched ICs at some point and the new IC has 2 op-amp sections. This pedal only needed one, so maybe wiring it as a buffer resulted in a better PCB layout than either disabling it or revamping the circuit in a more drastic way. Maybe this configuration let them keep using the old PCB layout without any changes.

Anyway, the wastefulness of an "extra" op-amp rubbed me the wrong way so I thought "what if we put that extra op-amp to better use?" Then I thought this might even reduce the parts count somewhat, resulting in a diminished variant of the DS-1. So by subtracting 1 from DS-1 we get DS-0. With the cute name settled, it was time to figure out what the mods should actually be.

Making a simpler version of a Boss pedal is easy nowadays. Simply swap the electronic switching for a mechanical switch, and (optionally) omit the input and output buffers. This may be enough to reduce the part count by 30% in some cases.

But I want to hold myself to a higher standard and focus on simplifying the "core" circuit. That is, the part that produces the characteristic sound. So yes I'll use mechanical switching and omit some of the buffers. But I won't increase the parts budget for what I consider to be the essential core of the circuit.

The original uses a passive "tilt" style tone control, most famously used in Electro-Harmonix Big Muff circuits. This tone control emphasizes bass when turned anti-clockwise, treble when turned clockwise, and in between has a pronounced midrange scoop. The details vary from pedal to pedal. Some pedals adjust the values so there is no scoop at all or even a midrange hump. Others make the midrange adjustable with another control.

I call this a passive tilt style tone control to hint that there might be an active version. Well there is, and its generalized form looks like this:

Using unequal value capacitors results in either a boost or cut at some midrange frequency, and adjusting the feedback and shelving resistors will tune the gain and response. So we can adapt the DS-1 passive tone control to active by a couple mechanical transformations:

1. send the tone control pot's output to an inverting op-amp's input
2. instead of grounding the resistor/capacitor, send them to the op-amp's output

With these alterations, we get a tone control that operates on the same corner frequencies as the passive version but with lower output impedance.

The active version overcomes some of the passive losses and as a result can actually boost frequencies in addition to cutting them. It may also have a more extreme range at the high or low ends of the potentiometer's sweep. One way to keep the original control response is by adding limiting resistors, but this would violate the "simpler" directive. Another way is to approximate the original response by modifying some of the components. This is trickier, but possible.

In my opinion the high-gain, saturating bjt amplifier stage is a key part of the DS-1's sound. I don't know a way to mimic it with op-amps so it can stay. But the bjt buffer may be on the chopping block. A quad op-amp is technically one component, so replacing the input bjt buffer with a unity-gain op-amp buffer would save one or two components. With the active tone control and the op-amp gain stage consuming 2 other op-amp sections, that leaves one section free. This could serve as another unity gain buffer between the clipping diodes and the tone control. This is perhaps necessary as the active tone control may be a lower impedance load on the clipping diode part of the circuit.

This modification doesn't add any components but gives more predictable performance to the circuit.

However, with a few added components there's an opportunity to lower the noise. The input op-amp could have more than unity gain, which would allow reducing the bjt amplifier's gain (and its noise). For the cost of 2 added resistors and 1 capacitor we could lower the noise. We end up with +1 components (net) but lower noise. Maybe it still fits the "DS-0" concept because we still reduced something. Credit goes to Dylan for this idea.

EJ pointed out that the Ibanez SD9 has a very similar circuit to the Boss DS-1. The only major difference I noticed is the SD9 has a much higher gain op-amp stage but is missing the bjt boost in front. This video with spectrograms shows there are some frequency response differences between the two pedals. Since the two pedals are identical from the clipping diodes to the output level control, these tone differences must come from the gain stages. A spectrogram doesn't help illustrate dynamics at all, and it's possible the input levels had an effect on the resulting spectra.

Here's a concept (missing clipping diodes because this is a linear frequency response simulation) of the DS-1 reworked to the point where it is basically unrecognizable.