Mido-centered Active Attenuators

Signal attenuation was a function missing from my collection of DIY things based on designs by Peter Blasser. All of these devices use a single +9 volt supply (Peter wanted his designs to be capable of being powered with a nine volt battery). The signals in these designs are biased around the midpoint between 9V and 0V, which Peter calls mido (a sort of Japanese pronunciation of middle). The signal are generally moving in the voltage range between 3V and 6V, though some have a greater range. When patching such a mido-centered control signal the CV inputs on these designs are expecting voltages in this range. Some CV inputs (colored purple) on PB designs have bipolar attenuators. But the blue Verso and Inverso CV inputs do not. This can present a problem when the parameter being controlled gets swept across a range that’s too big. For example, when modulating the Dubernator filter, the sweeps are quite large and exaggerated. (See video demo, below.)

The Circuit Design

Hand drawn Mido Centered Attenuator schematic

It’s a simple design, using the standard 9-volt 7809 type voltage regulator, expecting a +12V wall supply. The 1N4001 diode provides protection in case of reverse polarity on the input.

The mido voltage (4.5V) is done is the standard way of buffering a voltage divider. I used an old LM1458 op amp here. The other half of the 1458 is used for a manual bias feature, so that a 3V to 6V DC offset can be patched or mixed. On the panel this is the red pot and red jack.

The attenuators are simple op amp voltage followers, but referenced to mido instead of zero volts. With nothing patched to the input, or the attenuator all the way down, the output sits at 4.5V. WHY?

Why can’t a simple passive attenuator do the same job? Because it’s ground-referenced. Sending a low impedance zero voltage into a Verso or Inverso CV input is like sending it a -4.5V control, i.e. 4.5 volts below being unpatched. Try it! This is even a technique that Peter recommends to get a very low clock time on the Cocoquantus delays. But we want to attenuate a CV that’s biased around 4.5V.

Construction

I’d been thinking about such a whatchacallit (I mean what should these devices be called?), and then I came across two sets of plexiglass pieces plus a prototyping board exactly the right size! I ordered some new pots and got to it.

Top panel before wiring.

Note the DC inlet jack, super glued to the plexiglass between the pot and red jack. Black banana jack is the ground.

Top of the board, built, on nylon standoffs
Bottom of the built circuit, before wiring up
Full assembly before wiring to check clearances
Pretty close clearances!
Wiring between board and panel
Finished thingy

Demo Video

One Response to Mido-centered Active Attenuators

  1. Joshua Rodriquez says:

    Fantastic~!

Leave a Reply

Please use your real name instead of your company name or keyword spam.