"Bicycle Shifting" Stepped Attenuators?

[Syzygies]

I've been reading with great interest online various places about the qualities of stepped attenuators, how hard they are to wire, how people wish there was a good $1 resistor in between the 20 cent and $2 choices, how even a poor, multiple resistor attenuator design sounds better than a good pot, how most people only use a limited range of levels when they're actually listening.

Meanwhile, it's amusing to look at the circuit for an R/2R ladder digital-to-analog converter. Similar problem, yet they only use a couple of resistor values; one could grade and sort a handful of resistors by actual value to lay out a balanced stereo attenuator, no more hunting for odd resistor values to fill out the log curve, then hoping the two you ordered match.

Stepped attenuators also don't show up in the smallest portable headphone amps; they're too large. All this brings me to...

Would people put up with a pair of stereo volume controls, big-step and little-step like the two shifters on a bike? I think I would. I've never heard a biker in my life wish they had a single 27-speed shifter, they all suck it in and ride. I'm thinking two shorting 2x6 rotary switches, tied into an array of resistors giving 36 attenuator levels. There are many, many ways one could lay out the resistors; I love this sort of problem.

Any serious objections anyone can raise? I'm hoping to improve on the most basic stepped attenuator in a very small form factor, shunt quality, not ladder quality.

 

[rickcr42]

heh,been done man .Called "Volume" and "Trim"

As for the DAC/analog ladder comparison,also been done by adding logic chips to the feedback loop of an opamp which selects parallel/series groups of resistors hence changing the overall signal level.

but thinking on a subject is cool and you should pursue your ideas to a conclusion and let us know what your findings are.

clickety rick

 

[Syzygies]

Thanks. Do you have a link to the best resistor layout for this?

 

[rickcr42]

have far too many windows open at the moment (working on an AD Converter design) and my mind is scattered in like twenty directions but yes,i can provide links later.If i forget just bump up this thread by responding and my email will nudge the old memory banks

Rick

 

[AndrewFischer]

Quote:

Originally Posted by Syzygies Stepped attenuators also don't show up in the smallest portable headphone amps; they're too large. All this brings me to... Would people put up with a pair of stereo volume controls, big-step and little-step like the two shifters on a bike?

Certainly been done before. I've got a pair of these:

Have to impedance match the input and output to 600ohm for best operation. I has been years since I opened them up. As I recall they used .1% 5W resistors. Not at all small.

 

[Syzygies]

How about using toggle switches to write the dB volume reduction in binary? Have people seen this circuit before?

This circuit selects one of 16 stepped attenuator values, using four toggle switches. Using DPDT toggles, one could control both channels at once. The R values are all different, and need to be computed; this is one channel. The circuit could be made small enough to bring stepped attenuator quality to rather small headphone amps. I thought of this circuit by analogy to algorthms that exponentiate by writing the exponent in binary then multiplying.

The trouble with R/2R ladders is they're additive; this circuit is multiplicative, and the resistor values can be chosen so it presents a constant impedance to the source. It also doesn't have "make before break" type issues with the switches. Nevertheless, having the signal pass through four switch contacts is certainly a weak point. Are there switches up to the task?

This circuit is most easily understood by working recursively from the right. Believe that everything to the right of one's eyes actually works as claimed with the toggles either way, and at the moment they all happen to be down. Prove the same for the current stage, and move your eyes left.

 

[ChickenScrtchBoy]

Quote:

Originally Posted by Syzygies How about using toggle switches to write the dB volume reduction in binary?

Why not DIP switches?

They are sorta made for binary input... compact too.

Quote:

Originally Posted by Syzygies Using DPDT toggles, one could control both channels at once.

That would be a problem for the DIPs, but it could be advantagous if you want more volume on on side or the other...

 

[kevin gilmore]

These are remarkably good for this sort of thing, a little big though.

http://gilmore.chem.northwestern.edu/gertsch.jpg

The RF version is all surface mount resistors...
picture of those tomorrow.

 

[Syzygies]

Quote:

Originally Posted by ChickenScrtchBoy Why not DIP switches?

I've only seen single throw DIP switches; one would lose the top row of resistors in the circuit. They are there only to provide fixed impedance to the source. Giving this up is an accepted flaw in other attenuator designs, and then we'd fit in an Altoids tin. But would the contacts be up for the task? They're the weak link in the design, as most of the current is going through most of the contacts.

Oops! Solving for the resistor values, I noticed that the circuit itself depends on constant impedance down the line, for the switches to have a consistent effect. So double throw switches are necessary. However, small slide switches would work nicely.

Gold or silver? I seem to recall this depends on how often the switches are thrown?

 

[ChickenScrtchBoy]

Ha, never thought of that.

Good for a portable solution, though.

But then why not use a pot?
I wonder if the sound quality increase would be signifigant over the panasonic pot used standard in a mint...

 

[rickcr42]

hey kevin.what is the in/out impedance of those babies ?

Ids the transformer usable at audio frequencies and are they crazy expensive ?
plus,are the various adjustments clutch connected or are they all free spinning ?

Just wondering man but looks like an interesting solution for an audio attenuator project if the specs are workable in the audio range and the impedances right.


but not if it costs like thousands man,I know how you can get with these "toys" and i can't keep up with your private candy collection

 

[rickcr42]

Quote:

I've only seen single throw DIP switches; one would lose the top row of resistors in the circuit. They are there only to provide fixed impedance to the source. Giving this up is an accepted flaw in other attenuator designs, and then we'd fit in an Altoids tin. But would the contacts be up for the task? They're the weak link in the design, as most of the current is going through most of the contacts.

you would need an instruction sheet just to change the volume with all the combination possibilities.Better is to use a matrixing switch and then translate it to the logic which will do the actual resistor combination switching.
Another very old idea and something i read about first in Popular electronics magazine around 1975 or so in the "experimenters corner" section.Not EXACTLY as described above but damn close.

another possible and good for portable would be the use of voltage controlled analog switches.The data sheets are availably at Analog Devices and have many good ideas.Supply three volts at low current through a switch matrix and off you go

 

[Syzygies]

Could this be what you saw?

This is the circuit my first drawing wanted to be, which became clear under analysis. As shown, it has a constant input impedance of 10K ohms, easily scaled to other values. The selected attenuation shown is 10 dB + 2.5 dB = 12.5 dB. In binary, the toggles are set at 0101 = 5 * 2.5 dB = 12.5 dB. At least for me, it takes a little thought to see that this circuit actually adds the dB of attenuation together; it isn't obvious. Get the upper resistors wrong, the circuit stops working.

I'm a newbie who hasn't seen much, so I'm still at the stage where it's faster for me to reinvent the wheel, then ask experienced folks what's already known about the circuit.

This circuit seems far simpler than the stepped attenuators in common use, with a small resistor count making it possible to afford better components, etc. I'm wondering why others don't use it, why I shouldn't use it.

Quote:

Originally Posted by rickcr42 you would need an instruction sheet just to change the volume with all the combination possibilities.

Actually, I have the impression as a teacher that many people coming of age in this computer era find base 2 far more natural than base 10. Base 10 is unbelievably stupid after all, we just all learned to suck it in and not complain in grade school, and no one revisits the question as an adult.

Moreover, there is a definite placebo effect in cooking, audio, etc., where people need to demonstrate their commitment to the cause. Having to enter attenuation in binary in order to achieve sonic purity in a portable headphone amp? I think I've met the type. In any case, the amp is for me.

I need to digest your other suggestions. They appeal, but I wasn't sure that route wouldn't distort more, KISS and all that.

 

[dsavitsk]

Quote:

Originally Posted by Syzygies This circuit seems far simpler than the stepped attenuators in common use, with a small resistor count making it possible to afford better components, etc. I'm wondering why others don't use it, why I shouldn't use it. Actually, I have the impression as a teacher that many people coming of age in this computer era find base 2 far more natural than base 10. Base 10 is unbelievably stupid after all, we just all learned to suck it in and not complain in grade school, and no one revisits the question as an adult.

I, for one, am very interested in your volume switch design. I tried to do the same thing a few months ago (I even ordered a bunch of switches) but was unable to compute the values of resistors to use.

I think that binary adjustments are pretty straight forward to think about -- even more, I am very tactile and love playing with switches.

The real benefit of the design, however, is having one switch that goes from very loud to very soft w/o fully turning off the volume, and thus allowing returning to the original level. Levinson preamps do this with their mute button, b/t/w/.

Last, I think the natural evolution of this is to use transistors as the switches allowing for a very very inexpensive digitally controlled rotary stepped attenuator.

-d

 

[rickcr42]

Quote:

Actually, I have the impression as a teacher that many people coming of age in this computer era find base 2 far more natural than base 10.

probably where i get into trouble then.i "came of age" in the early days of color TV and way before the CD !

Quote:

The real benefit of the design, however, is having one switch that goes from very loud to very soft w/o fully turning off the volume, and thus allowing returning to the original level. Levinson preamps do this with their mute button, b/t/w/.

but you will need to use caution unless you have a redundant system.I can see the potential for the accidental full volume out scenario

Quote:

Last, I think the natural evolution of this is to use transistors as the switches allowing for a very very inexpensive digitally controlled rotary stepped attenuator.

and why not cmos analog switches ? Just a prepackaged FET switch that uses little power and you can even get "clickless" versions plus they are a plug and play solution.with a transistor you will need to set the operating points for each one externally.
Which brings up another thing : Switch Bounce.All mechanical switches have this and you will need to add "debounce" circuits if you also use logic circuits or the simple act of activating a switch may cause multiple triggers .Not hard to do (simple cmos inverter ) but essential .

good luck guys