DIYHiFi.org

What values of resistors will give equal decibel step sizes with only six switches? Or even usable step sizes?

Charles Hansen
Zinsula might be thinking about putting the resistors in networks so that the combination of relays (ie: 2^6 permutations) provide at least 32 different resistance/attenuation steps but he should probably clarify the setup.

zinsula
I guess I was trying to assess the performance of the switch part near its voltage rating. Do we avoid what CH has said about the tweakiness of using optocouplers if we attenuate like you propose above?

You may check these spreadsheets, although the L Pad Style and also the Pass Aleph P1.7 Style Attenuator would need DPDT switches, ie you have to double the number of MOS switches:

viewtopic.php?f=18&t=506

and

viewtopic.php?f=18&t=321

The spreadsheets are for eight switches, but they should also work for six. Just put very high resp. very low values for the non existent resistors if you use less than eight.

I have also double checked and cannot find a serious flaw. It seems important that there are R10 and R11 to reference the output to ground. But these come in handy to define max gain (together with R1)

zinsula
I probably didn't understand your circuit well enough to ask but I was referring to using a quad output switch through one resistor network. The one I looked at had a 24V rating. The ladder attenuators I've seen for <$50 have two or more resistor sets so tracking may still be an issue. In your circuit, are the switches replacing the relays?

Not sure I understand you. The switches are always meant to replace relays or mechanical switch. Anything else doesn't make sense to me.
What do you mean by "my circuit"? The balanced one, with only one pot between the phases?

zinsula
The chip switches in this discussion, do they have the same coloration issues that Charles Hansen mentioned? I would like to understand what exactly is being isolated. Will the attenuation occur through one resistor or a resistor network?

Replace the relay or the mechanical switch with a FET switch. It is really like a relay, study the datasheets and literature.
Then you choose the resistors.
Coloration...Charles speaks about coloration from machanical relays. I do not have gear with FET switches, so I cannot comment.

I think you now have some homework to do, and figure out by yourself how such a circuit could look like, I do not have a cookbook or even a schematic...

Hi,



Allow me to make a little annotation.

For good quality signal relays colorations are low and can be even lower if they switch essentially no signal current (load impedance approaches infinity), where they approach "transparent" to a degree similar to mechanical switches. However not all signal relays are very good switching what is called "dry", plus it can help to run the relays "inverted" (meaning turn them off for the operational position, turn them on to disengage).

For Fet's the colorations are in my experience greater than for good signal relay, however they quickly recede as the load impedance
rises high enough. They are practically unsuitable for switching appreciable signal currents.

I get more differences between resistors than between different signal relays of appropriate construction or signal-relay vs. good cmos switch into a near infinite load (e.g. J-Fet gate or Valve Grid).

Ciao T

_________________
Theory is when you know everything but nothing works.
Practice is when everything works but you don't know why.
Guruhood is when you know everything and everything works accordingly.

Thank you Thorsten for those comments.
While to me it seems easy to guarantee high impedance loads for relays or FET switches when using them for input selection, things may get complicated when we need to build an attenuator.....

On a sidenote, as far as I know, Nelson Pass uses bipolar transistors to attenuate the signal.

For what it's worth, Nelson seems to specialize in power amps. My understanding is that his associate Wayne Colburn designs everything else -- preamps, phono stages, and the now-discontinued DAC.

I believe that the bipolar switches work similarly to the FET switches. I was always a bit leery of them because one of the most elementary mods to a cheap CD player is to remove the muting transistors that short the output to ground. This normally driven with a reasonably low impedance, yet by all accounts they harm the sound. But I have never tried them so I cannot say from personal experience.