Attenuator for Passive pre amp, best sounding solution

[UMarcus]

Hi,
i want to build a passiv pre amp and looking for the best attenutator solution.

On my list there are following :

1. Potentiometer ALPS (motorized)
2. Relay based Attenuator with resistor array
3. IC based volume control (MUSE 72320)
4. Auto Transformer
5. Lightspeed Attenuator


It should be possible to remote the volume, so i exclude the switch based solution which is near to relay based solution (?)

Did anyone could advise or had compare the different attenutator solutions ?
Is there a noticeable sound difference herable (maybe in a high end system) ?

Thanks, regards
Marcus

 

[Zen Mod]

AVC & Input Selector for Papa’s Koan M2 ; temporary case ? | Zen Mod Blog

best when buffered on input

for my ears , superior to , say, Penny-Giles attenuator , TKD and similar

 

[UMarcus]

Hi Zen,
wow, looks great. Thank you for your response and pointing me to that. This sounds interessting, i have to update my mind.

Which buffer design you prefer for balanced signals ?
Is the signal dc decoupled by using an auto transformer ?

Thanks,
Marcus

 

[thommy]

Would you add an LDR attenuator to that list?

I have just built one of the Stereo Coffee attenuators and am very pleased with it. Have it between DAC and power amps, takes nothing away from the sound other than gain.

It works for balanced signals and cost me $160. It's capable of switching between 3 balanced inputs which makes it great value for money. You could attach an output buffer if needed but I use it without and get fantastic results.

There's no remote with the kit, but any remote controlled pot of the correct impedance will control the LDRs.

 

[UMarcus]

Would you add an LDR attenuator to that list?

I have just built one of the Stereo Coffee attenuators and am very pleased with it. Have it between DAC and power amps, takes nothing away from the sound other than gain.

It works for balanced signals and cost me $160. It's capable of switching between 3 balanced inputs which makes it great value for money. You could attach an output buffer if needed but I use it without and get fantastic results.

There's no remote with the kit, but any remote controlled pot of the correct impedance will control the LDRs.

Hi Thommy,
many thanks for your response, the LDR are not on my scope, i add it to the list.

Do you compare some solutions with the LDR ?

 

[Neilsan]

I've had a Goldpoint 25kohm attenuator for a while after trying other "passive pots" over the years, and love it.But now I've gone to one of the Stereo Coffee LDR's and it's definitely a step up sound-wise.
Have a look at this thread.
http://www.diyaudio.com/forums/vend...ht-dependant-resistor-3-input-preamp-kit.html

 

[abraxalito]

A transformer is the best sounding attenuator. Resistive attenuators will add more loading to the source, reducing its dynamics as well as increasing the impedance seen by the destination component.

 

[UMarcus]

Many Thanks for your input Guys.

Somebody know a remote control transformer solution ?

Would it be possible to combine the transormer with a µC controlled relais array ? Drawbacks ?

 

[abraxalito]

I don't know of a commercially available one but sure this would be an excellent DIY project, combining hi-tech uC with low-tech trafo I think relays probably wouldn't be needed, MOSFET switches might well do the job well enough.

 

[Calvin]

Hi,

I´d say a transformer is certainly not the best solution, as the ´best´ always depends on the circumstances, as price, resolution/steps, precision, etc. etc.
- Due to its inductive nature a Autoformer is a nonlinear device. Typically the resolution restricts to a few large steps and size is large. Beeing specialty devices cost is high. Most offers use mechanical switches offering only low comfort, though relais may be used for more handling comfort. Voltage range and power can be high. Due to low comfort, it doesn´t lend itself well to multichannel use.
- Integrated chips like the Muses excel in compactness, number of steps and probabely also in handling convenience, especialy in multi-channel applications. It may be argued about the (sound) quality of silicon resistors and numerous MOS-switches which add THD. Voltage range and power capability is typically small but sufficient for Audio.
- Potis are still the workhorses and come in countless styles. They are compact and may offer some comfort if motorized. Interchannel balance and dB-linearity may be from mediocre to good. For a passive Preamp their varying and rather highish output impedance may be troublesome depending on the load. Typically only short cable runs and higher impedance loads are suitable. Cost and quality may vary over a huge range. Voltage range may be high, Power is typically low, due to the filigrane wiper contacts.
- LDRs offer a compact, low parts number count solution. Number of suitable parts is low, tolerances are high and need screening. Technically an elegant solution allowing for easy handling and comfort and multichannel use.
- Resistor array offer a high dedree of precision. Depending on the chosen number of Steps the number of parts and cost may be high. Build size is not compact.
If intended for a passive Preamp a network should be chosen that offers a lowish and constant output impedance. If a bit-wise weighted network is chosen to reduce the number of switches and resistors, the resistors need to be very tightly tolerated. Only then is a high level precision in dB linearity achieved. Each step itself offers the highest linearity of all solutions as a high quality resistor is without doubt the most linear device one could use. Interchannel balance is excellent and handling and comfort may be high. Adopts well to multi-channel useage. Voltage range and Power can be high.

The final signal quality of the listed solutions depends on the quality of the parts and their implementation. And each of the different techniques has proven good results as well as inferior results. Factors like cost, build size and especially comfort and easy of handling should not be understimated in the evaluation process.

I´ve made excellent experiences with bit-wise weighted switched resistor networks. Sonically its as neutral as one could wish for. At typical listening levels no more than 2-3 resistors are within the signal path. Imho one couldn´t come much closer to the proverbial piece of wire. If it sounds limited in dynamics, its implementation certainly is flawed. Done right the dynamics are outstanding (... as the resistors are the most linear and lowest limiting devices).

jauu
Calvin

 

[thommy]

Interesting advice abraxalito. "The thing you need doesn't exist and the things everyone uses instead are inherently unsuitable".

I think the OP wants a solution that is actually available for sale.

It was the stereo coffee attenuator that I bought and can't recommend it enough. The comment about resistive attenuators destroying dynamics is just rubbish. Potentiometers and stepped attenuators use resistors to control gain, and with good effect.

 

[abraxalito]

Where was the comment that resistive attenuators destroy dynamics? They certainly limit dynamics for the reasons already stated - introducing greater loading on the source. The implementation isn't at issue, that's assumed to be competent (i.e. blamelessly linear).

@Marcus - a quick search turned up this review that you might find interesting - http://www.enjoythemusic.com/diy/0611/slagle_autoformer_volume_control_modules.htm

 

[UMarcus]

I don't know of a commercially available one but sure this would be an excellent DIY project, combining hi-tech uC with low-tech trafo I think relays probably wouldn't be needed, MOSFET switches might well do the job well enough.

But with MOSFET in signal path it is no passive preamp anymore

 

[UMarcus]

Hi,

I´d say a transformer is certainly not the best solution, as the ´best´ always depends on the circumstances, as price, resolution/steps, precision, etc. etc.
- Due to its inductive nature a Autoformer is a nonlinear device. Typically the resolution restricts to a few large steps and size is large. Beeing specialty devices cost is high. Most offers use mechanical switches offering only low comfort, though relais may be used for more handling comfort. Voltage range and power can be high. Due to low comfort, it doesn´t lend itself well to multichannel use.
- Integrated chips like the Muses excel in compactness, number of steps and probabely also in handling convenience, especialy in multi-channel applications. It may be argued about the (sound) quality of silicon resistors and numerous MOS-switches which add THD. Voltage range and power capability is typically small but sufficient for Audio.
- Potis are still the workhorses and come in countless styles. They are compact and may offer some comfort if motorized. Interchannel balance and dB-linearity may be from mediocre to good. For a passive Preamp their varying and rather highish output impedance may be troublesome depending on the load. Typically only short cable runs and higher impedance loads are suitable. Cost and quality may vary over a huge range. Voltage range may be high, Power is typically low, due to the filigrane wiper contacts.
- LDRs offer a compact, low parts number count solution. Number of suitable parts is low, tolerances are high and need screening. Technically an elegant solution allowing for easy handling and comfort and multichannel use.
- Resistor array offer a high dedree of precision. Depending on the chosen number of Steps the number of parts and cost may be high. Build size is not compact.
If intended for a passive Preamp a network should be chosen that offers a lowish and constant output impedance. If a bit-wise weighted network is chosen to reduce the number of switches and resistors, the resistors need to be very tightly tolerated. Only then is a high level precision in dB linearity achieved. Each step itself offers the highest linearity of all solutions as a high quality resistor is without doubt the most linear device one could use. Interchannel balance is excellent and handling and comfort may be high. Adopts well to multi-channel useage. Voltage range and Power can be high.

The final signal quality of the listed solutions depends on the quality of the parts and their implementation. And each of the different techniques has proven good results as well as inferior results. Factors like cost, build size and especially comfort and easy of handling should not be understimated in the evaluation process.

I´ve made excellent experiences with bit-wise weighted switched resistor networks. Sonically its as neutral as one could wish for. At typical listening levels no more than 2-3 resistors are within the signal path. Imho one couldn´t come much closer to the proverbial piece of wire. If it sounds limited in dynamics, its implementation certainly is flawed. Done right the dynamics are outstanding (... as the resistors are the most linear and lowest limiting devices).

jauu
Calvin

Hi Calvin,
many thanks for your explanation, this very interessting.

Did you mean with flawed implemetation that the resistor values (impedance) is wrong in the system ? How i could handle different cases, for e.g. i connect source devices with different impedance to the attenuator ?

Thanks,
regeards
Marcus

 

[Zen Mod]

I made my own LDR , also have some pretty exquisite pots , tried few sorts of switched resistive attenuators , tried them all buffered or not (from both sides , if needed ) and - for my ears - xformer based solution is most fun

that's all I wanted to say in previous post

 

[abraxalito]

But with MOSFET in signal path it is no passive preamp anymore

Certainly that's one way of looking at it. To me it depends what function the FET's doing - as a switch that's quite different from as an amplifier.

Incidentally 'signal path' is one of those audiohile buzz terms that doesn't reflect reality - that current travels in loops

 

[Calvin]

Hi,

regarding the impedances and loading.
Working in the voltage domain the load impedance seen by the source should be at least 5 times larger as the source impedance.
With typical source devices we see a output impedance of <<1k.
Load impedances >5k will therefore suffice.
A poti network, be it either a classical poti or a switched resistor network offers a constant load impedance (input) of the value R of the poti and a from 0R to R/4 varying source impedance (output) with the lowest values at each end position and the maximum at midpoint position (linear type).
The bitwise resistor network I mentioned behaves different.
Its input impedance varys while the output impedance can be held constant.
If You design such a network for a output impedance of ~1k, the input may have a minimum value of 4-5k and a maximum that rises with the number of Bits.
It may reach a low Meg range.
Anyway, the minimum input impedance is a certainly high enough load for any decent source device.
There's no technical explanation for a dynamics limiting mechanism and its in contrary to my listening experience.
A output impedance of ~1k is on the other hand low enough to drive almost all devices You connect it to with sufficient bandwidth.
Only for loads <<10k and/or cable runs >10m do I switch in a dedicated Buffer stage.

jauu
Calvin

 

[buzzforb]

Option along with Slagle.
SACThailand

 

[abraxalito]

There's no technical explanation for a dynamics limiting mechanism and its in contrary to my listening experience.

Perhaps you mean 'I'm aware of no technical explanation' ? Its not contrary to mine and other people's listening experiences (that trafos sound better than resistive attenuators) and the technical explanation I've given satisfies me - lower impedance loading makes for more supply noise within the output stage of the driving circuitry. At the same time, signal level is decreased by the attenuation. Therefore more noise, less signal. With a trafo the SNR is preserved, provided its shunt inductance is sufficient.

There's another issue with resistive attenuators when put in-line. The ground used for the shunt arm of the attenuator is often a contaminated one - common-mode currents flow along the screens of unbalanced interconnecting cables. Inserting a trafo attenuates the CM currents by virtue of providing galvanic isolation. Auto-trafos though don't convey this advantage.

 

[Calvin]

Hi,

that trafos sound better than resistive attenuators

I hold this to be self evident, that every human is entitled to err every now and then in its life
No, just kiddin', no pun intended, but You set yourself up for that one.
I'd guess that as everybody has a different listening experience You'll find as many supporters of one principle as You'll find for the other.

Any volume control will have a certain insertion loss.
My network's loss remains <0.5dB up to a source impedance of 1.8k.
If one thinks that this amount of loss is of any noticeable order, well than we hace an issue.
Anyway, I still happily enjoy the good feeling of turning at the wheel of a extremely precise, THD+N transparent, comfortable and sonically neutral volume control ... and btw. did I mention that its IR controllable too?

jauu
Calvin