Stepped Attenuator - Passive Volume Control [SOLVED]

[jim_griff]

Hi everyone. Long time since I've posted here but just got back into electronics!

I have a question if anyone can help me. It's an Ohm's Law question to do with impedance (with respect to resistance in an AC circuit).

I'm currently building a stepped attenuator for a passive volume control as my studio volume control is playing up. Always wanted to have a passive solution with less circuitry to muddy up the signal.

Criteria:
1. I have two balanced signals (Left +/- and Right +/-) that I need to transfer from my audio interface to my monitors.
2. There needs to be switched 20dB attenuation and Mute as well as volume control from the stepped attenuator.
3. There needs to be 3 switchable inputs and 3 switchable outputs for different monitors.

I've designed and built parts of the circuit but need a little help...


My questions are:

1. Is there a way of designing the circuit to only have attenuation on switch position 2 on SW2? As it is at the moment, there is attenuation in both the fully on (SW2 position 3) and attenuated (SW2 position 2) switch positions.

2. Going off the impedances (below), I can't quite figure out how to work out the input/output impedances required to make the attenuator stage as close to -20dB as possible. I have tried H-bridge configuration, U-bridge, even T-bridge (I know it's for non balanced signals). The trouble is at switch position 3 (fully on) it's still drawing current down to ground through the attenuator pad stage.

Audio interface 150ohm out --> Attenuator (pad) ~2kohm in and out --> Stepped attenuator 2kohm in and out --> Studio monitors 20kohm input impedance.

Circuit is linked in below. You can probably trace through one of the inputs. The rest are the same.

Thanks!

[jim_griff]

In switch position 1 (SW2):
The input signal goes through 2.2k --> 560 ohm --> Ground. So no signal gets through.

In switch position 2 (SW2):
The input signal goes through 2.2k --> and 560ohm to ground and also 2kohm to ground through the stepped attenuator.

In switch position 3 (SW2):
The input signal goes through 2.2k --> 560 ohm --> Ground and also goes through to the stepped attenuator (2kohm to ground).

In these scenarios, I can't figure out how to:
1. Build the circuit to not have any impedance on switch position 3, or
2. Build it as it is, but choose the correct resistor values such that switch position 2 is 20dB lower than switch position 3.

Ohm's Law is all very confusing!

[jim_griff]

This (attached) is an attempt with the switch turned around. Unfortunately, in switch position 3 (fully on) there is still the 560 ohm down to ground, so the signal still won't be full volume.

EDIT: Even in fully on, it still has 446.3ohms down to ground through a series 2.2k resistor, so amplification factor of ~0.163. Simply not good enough. Gah, circuits! EDIT 2: No wait, I looked at that it wrong.

EDIT 3: Okay, think I've figured it out. The way it is now:

SW2 Position 1 -- Input shorted to ground. MUTE.
SW2 Position 2 -- Input goes through series 2.2k and shorted with two resistances (parallel) which is ~446.3 ohms to ground, so 0.163 amp factor.
SW2 Position 3 -- Input goes directly to the stepped attenuator with 560 ohms to ground and 2k through the stepped attenuator. This gives ~0.748 amp factor. Almost there! Just need to find a combination of resistors that makes switch position 2 0.1x that of position 3. Maths was never my strongest subject so trial and error will have to suffice.

[jim_griff]

SOLVED!

Trial and error trumps mathematics (in my case) since I can't math very well.

Will leave the post up in case anyone else is interested in looking at my poor circuits and trial + error diagnosis.

Cheers, Jim.

EDIT: Although, now I'm going to have to check each switch position in the stepped attenuator to make sure the parallel resistance doesn't cause too much skewing of the volume at each step. Doh!

[jim_griff]

Just posting the stepped attenuator excel calculation with the 820 ohm in parallel with it at each switch position. It now has lower total input impedance so I'll have to work out the 20dB pad values again and shuffle it around until it works.

EDIT: Final values:

Switch position 1 -- MUTED (-inf dBref)
Switch position 2 -- PAD (-20.92 dBref  OR  x0.07 amplification factor)
Switch position 3 -- FULL (0.0 dBref  OR  x0.778 amplification factor)

Sorted and happy!

[jim_griff]

Here's the final circuit if anyone's interested in building a passive volume control using 2k potentiometer or a 2k stepped attenuator. The previous post ^ shows the values of the resistors to be soldered between each pin on the stepped attenuator. 2.2 ohm, 3.3 ohm, 4.7 ohm, 4.7 ohm, 6.8 ohm, etc...) They're in the column at the end. The Yellow highlighted part shows the actual resistance values at each switch position when combined with the 820 ohm resistors to ground, which in parallel reduces the overall resistance.

NOTE: This stepped attenuator circuit is the type that has constant input impedance, unlike others where the input impedance changes with each volume step.

The main switches are 4P3T (4-pole 3-toggle) so it can support two balanced audio channels. I'm using Rotary switches from Maplin.

It has a "Mono" switch (DPDT) which electrically sums both the L and R non-inverted (hot) and inverted (cold) signals together.

I may refine the circuit in future so it is almost exactly 600 ohm input impedance, but I'm happy with this in its current form. Most audio interfaces output at <200 ohms anyway, so it should be fine for most setups.