No gain, no pain

Every musician is faced with some sort of gain or attenuation control when connecting the bass to a (pre)amp. The deBont bass pre-amps will take a slightly different approach to the typical gain/pad control arrangement found on most amps. Given the low-noise and ease-of-use design aspects, the arrangement explained below is much better suited for the job. It will accommodate for the huge variety of basses out there while keeping everything simple and still give the user more than enough control over the input-signal, with of course the maximum quality of sound.

Different input arrangements
The first amplifying elements seen by the bass guitar, either tubes, op-amps, FETs or BJTs, will always have a maximum input voltage. Going beyond this voltage will cause the input to ‘clip’ in case of solid state or distort some other way if tubes are used. It should be avoided if a clean sound is to be expected. A bass at full volume, even passive ones, can have a huge signal swing. I’ve seen outputs of 10V peak to peak, even with the bass’ EQ set flat. Enough to exceed the maximum input voltage on many amps. Think of the 12AX7 (ECC83), a very common tube used as the first gain stage. It’s ‘clean’ input voltage range is just a few volts!

Most amplifiers will provide some means of user adjustable control to keep the signal in check. There are a few typical arrangements found on tube amps, all of them with their strengths and weaknesses. I’ll leave the dual channel and ‘bright’ inputs out for now.

The difference between gain and padding
The pad button will put a voltage divider between the bass and the first amplifying element, reducing the input voltage by a set amount. An attenuation of -15dB is common which translates roughly to 17.5% of the original signal, but other values are of course possible. A 6Vpp signal causing distortion for example becomes 1.05Vpp and clipping/distortion will no longer occur.

A variation on the pad button comes in the form of a dual input setup and is seen regularly on tube amps. One input for 0dB, the other attenuated. The Alembic F-2B (-6dB), Fender Bassman (-6dB) or Ampeg SVT Classic series (-15dB) to name but a few. They’re labelled 1/2, hi/lo, active/passive or again, with the given amount of attenuation.

Controls labelled ‘gain’ are wired in several ways. The first is somewhat similar to the pad button. The gain potentiometer is used just like the pad button’s voltage divider. Only now, adjusting it will vary the input voltage padding gradually from 0db to infinity. But this is a horrible type of input as can be read below. The second option is to have the gain control in between the first two gain stages. This is common in guitar amplifiers or bass amps where an ‘over-driven’ sound is desirable. Because the incoming signal has already been amplified by the first stage, the gain control will have a huge signal at it’s disposal and it will control whether the rest of the amplifier will produce a clean signal or any amount of distortion. Since the last option is usually seen in amplifiers that want to induce distortion, something I really want to avoid, I’ll ignore this type of gain control in the rest of this article.

The drawbacks of the common gain control and padding options
The pad button or dual input setup is easy and intuitive, but not without its flaws. Obviously, the ‘one-value-fits-all’ approach. The typical -15dB might be way more than necessary, while -6dB might be too little for basses with high output. Plus, the sound of an amp can actually differ depending on how hard the input is driven. So even if distortion isn’t the issue, you’re always stuck with just the one tonal option.

A lesser known drawback is the massively decreased input impedance when using attenuation. It usually plummets from 1MΩ to around 57kΩ composed of the 47kΩ + 10kΩ voltage divider for the required -15dB. The specific amount of attenuation however only applies when the bass itself has near zero output impedance. In other words, for active basses. When using a passive bass with volume control, dialing the output back just a bit can increase the bass’output impedance easily to 50kΩ. This impedance is added to the already present 47kΩ. Total attenuation will no longer be -15dB, but -20.6dB. You can see how using the controls on a passive bass will have a dramatic effect on the actual attenuation.

But, gain control right in front of the first gain stage is the worst option of them all. Because its value solely determines the amplifiers input impedance, it should be of high value. Preferably 1MΩ, which is very common, to get the maximum signal transfer from low output basses. The problem starts when the control is not at its min/max position. Let’s take an attenuation setting of 0.5 (-6dB). The gain potentiometer will be set halfway and the gain stage will now have 500kΩ is series with its grid (we’re talking tubes here). This will add a tremendous amount of white noise (Johnson–Nyquist noise) to the signal. Especially if a carbon element potentiometer is used. Other noise, which prefers to travel the point of least impedance, will have nowhere to go either.

As if that’s not enough to ditch this topology, let’s get a bit more technical. All resistance added up in series with the tubes’ grid will create a low pass filter together with the tube’s miller capacitance. In short, this is a number composed of the various capacitances seen at the input multiplied by the gain of the stage.
Mathematically: Cin = Cgk + (Cgp*(A+1))
Cin = input capacitance
Cgk = grid to cathode capacitance
Cgp = grid to plate capacitance
A  = stage gain

For the commonly used 12ax7 with a typical circuit gain of 60, this makes:
Cin = 1.6pF + (1.7pF * (60+1))
Cin = 105.3pF

In practice, the number will be a bit higher due to various stray capacitances, but this is fine for this example. Now we can calculate the low pass -3dB corner frequency:
f = 1/(2*π*R*Cin)
f = 1/(2*π*500k*105.3p)
f = 3022Hz

This will severely impact the high frequency response, resulting in a dull sounding bass. And this doesn’t even include the bass’ own output impedance, possible grid-stopper or stray capacitances, let alone if more attention is used which will put more resistances in series with the grid. One solution is to decrease the value of the potentiometer, but the lower input impedance is an unwanted compromise. Gain control directly at the input? Not an option.

So now what?
With the typical issues spread out above, we can create a set of design goals.
– It has to be simple
– No need for distortion
– We need more than one fixed attenuation value
– Accommodate for every active or passive bass
– Provide the best possible signal to noise ratio

A variable or multi-settings attenuator. Commonly seen in audiophile equipment under the name stepped-attenuator.

There’s no need for 20 different settings, but cutting the signal in half with each step seems ideal: 0dB, -6dB, -12dB and -18dB. This will still keep it simple while providing a lot more flexibility than the single pad button. Plus, there’s only need for one physical input.

Another benefit, the switching circuitry can use high quality, low noise resistors instead of relying on e.g. a gain pot’s carbon element, improving the overall signal-to-noise ratio.

And the ability to implement the resistor network as I see fit, allows for a per-setting optimum input impedance. This will truly add support for every single bass I’ve ever come across.

Given my ideas and convictions on how an amp should function and sound, this input arrangement is the perfect solution. The initial tests with point-to-point wired circuitry confirmed this. A brand new input prototype PCB is on its way. Build photos, tests and more will of course come soon after.