• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Budget passive preamp with stepped attenuator?

  • Thread starter Deleted member 58865
  • Start date
You've gained 2 dB dynamic range at -20 dB volume position with a perfect analog volume control. Without a lower noise / lower gain amplifier you will not be gaining much from using a higher dynamic range DAC or an analog volume control.
-20dB is just DAC volume. You need to also take application volume into account. That can be additional -12dB ~ -20dB attenuation.
 
Anyway, noise floor is different from dynamic range.

I don't think you understand the issue, they are intimately related as I showed in my calculatiosn. What do you think you actually hear, residual noise or dynamic range? Dynamic range is just a comparison of output signal to residual noise. What do you think you are actually gaining by using an analog volume control?

For example you can have an amplifier (or DAC) with excellent dynamic range because it is capable of very high output. For example let's say your amplifier has a dynamic range of 110 dB at 500 W in to 4 ohm. Residual noise in this case is sqrt(500 x 4) x 10^(-110/20) = 141 uV which is actually not very good, for comparison a Hypex NC252MP has residual noise of less than 50 uV.

Michael
 
What about application volume, sir? I guess application volume is usually between 0dB and -20dB. That adds to DAC volume of -16dB ~ -20dB.
 
-20dB is just DAC volume. You need to also take application volume into account. That can be additional -12dB ~ -20dB attenuation.

That's fine, they will still only be 2 dB different across different volume control positions (whether from application volume or DAC volume). Here is a plot showing a perfect analog volume control (orange trace) vs digital volume control (blue trace) for your system. Note above volume position of -10 dB isn't really credible as you run out of amplifier power.

1680182545662.png


Michael
 
Last edited:
Where did you get this? Reducing application volume certainly doesn't lower noise floor.

I think reducing application volume can cut out quiet details because it meddles with digital bits while reducing DAC volume only lowers analogue signal.
 
I don't really understand how noise increases as volume goes down in the graph. If you reduce analog volume, noise should go down.

I think your calculation is off.
 
Last edited by a moderator:
Reducing DAC volume doesn't eliminate bits from audio.

However, I think reducing application volume can eliminate digital bits from audio. Maybe, I'm wrong.

I also don't understand how gain in dynamic range by stepped attenuator stays at roughly 2dB even if DAC volume is nearly at -60dB
That seems impossible. I don't know where your equations come from.

Can you give me an intuition on how gain in dynamic range stays fixed at all DAC volume levels I care about?

Maybe, noise floor is so negligibly low in the first place that dynamic range is not affected much by it for all DAC volume levels I care about?
 
Last edited by a moderator:
You are simplifying this too much. You need to look at residual noise at the amplifier output as a function of volume control position considering residual noise and gain from the amplifier as well as residual noise inherent in your source.

If you don't have issues with audible hiss currently, adding an analog volume control is not going to help you.

Based on PMA's measurements -> https://www.audiosciencereview.com/...s-power-supply-measurements.19507/post-642669 A07 residul noise is 10^(-75.2/20) = 174 uV residual noise from the A07 with an improved power supply. For reference this is a 5 W in to 4 ohm dynamic range of 20 x log[sqrt(5 x 4) x 10^6 / 174] = 88 dB.

Actual system noise performance will depend on the volume control position of the A07, lower gain = lower amplification of upstream noise but this comes with a channel imbalance penalty. For this analysis let's assume 29 dB gain like the ASR review.

Residual noise from D10S = 2 x 10^(-112/20) = 5 uV
Amplified residual noise from D10S = 5 x 10^(29/20) = 142 uV
Total residual noise from amplified DAC output and amplifier = sqrt (142^2 + 174^2) = 225 uV

225 uV residual noise is pretty good and is likely not audible depending on speaker sensitivity, hearing and listening position.

Now let's say you typically listen at -20 dB, that will reduce your output signal but noise remains constant with digital volume control.

Voltage at DAC output at -20 dB = 2 x 10^(-20/20) = 0.2 V
Voltage at amplifier output at -20 dB = 0.2 x 10^(29/20) = 5.6 V
Dynamic range at amplifier output at -20 dB = 20 x log(5.6 x 10^6 / 225) = 88 dB

Now let's look at the case of perfect analog volume control. In this case both upstream signal and noise are attenuated.

Attenuated residual noise from DAC output at -20 dB (analog) = 5 x 10^(-20/20) = 0.5 uV
Amplified residual noise from DAC output at -20 dB (analog) = 0.5 x 10^(29/20) = 14 uV
Total residual noise at -20 dB (analog) = sqrt (14^2 + 174^2) = 175 uV

Now noise is dominated by your amplifier where as before it was roughly equal parts DAC and amplifier.

Dynamic range at amplifier output at -20 dB (analog) = 20 x log(5.6 x 10^6 / 175) = 90 dB

You've gained 2 dB dynamic range at -20 dB volume position with a perfect analog volume control. Without a lower noise / lower gain amplifier you will not be gaining much from using a higher dynamic range DAC or an analog volume control.

Michael

Michael, thank you so much for going in to so much detail here, I too am on the hunt for a good passive and may wind up making one (or several) since my chain is similar, dac -> (pot/stepped att) -> nc252mp

A minor note, from PMA's measurements of the D10S at https://www.audiosciencereview.com/...g-d10s-owner-measurements-quick-review.28462/

Noise is below 2uVrms measured at 22Hz-22kHz unweighted, a little bit better than by manufacturer specs! S/N is better than 120dB then. Please note that I use @22kHz bandwidth unweighted, which is more strict than A-weight used by Topping. So the noise parameters are better than by manufactirer specs! With A-weight, I measured 1.44uVrms noise.
Note: my noise measurement limit is 0.7uVrms (22Hz-22kHz unweighted)

I'll save the noise, but honestly finding a simple passive with a rca/xlr in, xlr out that takes in to account bruno's notes and has a good pot / stepped attenuator is very hard, I keep coming back to a couple of nob sounds devices in my hunts (ns-05p, or their 256 Stepped Attenuator under different brands that's hard to get quickly).

I'd give my left arm for a simple device, and 1.5 for one which cancelled dc too. I may as well ask, what do you use / know of / recommend?

edit:
It's a spreadsheet I built after investigating audibility of residual noise here -> https://www.audiosciencereview.com/...a-with-topping-performance.24768/post-1008785.
This is gold Michael, one of the most valuable posts I've read so far, thanks so much for sharing.
 
Last edited:
@sonder According to him, you don't need a passive volume control for at least up to -60dB of DAC volume.
 
I don't really understand how noise increases as volume goes down in the graph. If you reduce analog volume, noise should go down.

I think your calculation is off.

I think I've reached a newbie stage where I can perhaps explain, or do it wrong and be corrected.

signal -> attenuator -> amplification

consider amplification constant, signal constant, and the volume/attenuation in the middle a bottleneck that turns the signal up and down, any noise/uV introduced at any step is amplified, and the attenuator add's a bit too.

I'm quickly learning that this one little bit is the funnel through which all else goes, and could POTentially be one of the most critical bits in my setup.
 
Paul McGowan of PS audio said his company started producing amplifiers without attenuators. The volume knobs on their amplifiers control amplification gain instead.

Since their amplifiers do not have attenuators, their amps produce cleaner sounds supposedly.
 
Without a lower noise / lower gain amplifier you will not be gaining much from using a higher dynamic range DAC or an analog volume control.
Does the volume knob on AIYIMA A07 not control amplifier gain?
 
@mdsimon2 I guess I will just upgrade AIYIMA A07's power supply to Mean Well RS-100-24 instead of adding a stepped attenuator between Topping D10s and AIYIMA A07.

24V is better because it allows me to keep the amplifier volume knob at a higher level and have less volume imbalance.

Do you suggest I connnect the ground of RS-100-24 to an electrical outlet's ground? Or, should I just ground AIYIMA A07 through RCA cables?
 
I don't really understand how noise increases as volume goes down in the graph. If you reduce analog volume, noise should go down.

In the case of analog volume control noise at the DAC output certainly goes down, again at -20 dB you are going from 5 uV to 0.5 uV. However, noise at the amplifier output of 174 uV is a constant and now dominates your system noise.

Reducing DAC volume doesn't eliminate bits from audio.

However, I think reducing application volume can eliminate digital bits from audio. Maybe, I'm wrong.

I've never seen a difference between DAC volume control and application volume control. All digital volume control "eliminates bits" by reducing the signal level.

I also don't understand how gain in dynamic range by stepped attenuator stays at roughly 2dB even if DAC volume is nearly at -60dB
That seems impossible. I don't know where your equations come from.

At the amplifier output you have residual noise from your amplifier plus whatever noise you have at your DAC / preamp output multiplied by your amplifier gain. The noise sums as root squared.

Can you give me an intuition on how gain in dynamic range stays fixed at all DAC volume levels I care about?

Again, your amplifier noise is dominating system noise.

Does the volume knob on AIYIMA A07 not control amplifier gain?

It does, at the expense of channel imbalance. If you lower amplifier gain the benefit of analog volume control becomes even less because you are multiplying the noise from the DAC even less.

Review the thread I linked about audibility of residual noise, it will help you understand the calculations. Until you understand the fundamentals of how noise propagates through the system you won't understand how to increase performance compared to what you have currently. And most importantly if you do not hear noise currently reducing system noise will not help.

Michael
 
It does, at the expense of channel imbalance.
At 9'o clock of the volume knob on AIYIMA A07, I do not notice noticeable channel imbalance. So, I keep it at 9'o clock and fix DAC volume between -16dB and -20dB.

Then, I control audio volume per application. I usually keep application volume between 25% and 70%.

Below 9'o clock, I may start hearing channel imbalance.
 
Michael, thank you so much for going in to so much detail here, I too am on the hunt for a good passive and may wind up making one (or several) since my chain is similar, dac -> (pot/stepped att) -> nc252mp

A minor note, from PMA's measurements of the D10S at https://www.audiosciencereview.com/...g-d10s-owner-measurements-quick-review.28462/



I'll save the noise, but honestly finding a simple passive with a rca/xlr in, xlr out that takes in to account bruno's notes and has a good pot / stepped attenuator is very hard, I keep coming back to a couple of nob sounds devices in my hunts (ns-05p, or their 256 Stepped Attenuator under different brands that's hard to get quickly).

I'd give my left arm for a simple device, and 1.5 for one which cancelled dc too. I may as well ask, what do you use / know of / recommend?

edit:

This is gold Michael, one of the most valuable posts I've read so far, thanks so much for sharing.

Thank you for the kind words, I'm glad it was helpful. It was a fun learning exercise and taught me a lot about how to interpret measurements.

I've personally found that a decently low noise DAC (110 dB DR at 2 V) paired with a NC252MP results in inaudible noise but I don't run very sensitive speakers. As a result I don't have much motivation to pursue analog volume control, especially because all of my systems are DIY active which would require multichannel analog volume control.

I think I've reached a newbie stage where I can perhaps explain, or do it wrong and be corrected.

signal -> attenuator -> amplification

consider amplification constant, signal constant, and the volume/attenuation in the middle a bottleneck that turns the signal up and down, any noise/uV introduced at any step is amplified, and the attenuator add's a bit too.

I'm quickly learning that this one little bit is the funnel through which all else goes, and could POTentially be one of the most critical bits in my setup.

This describes the contribution from the DAC after being multiplied by the amplifier. If your attenuator is digital then only the signal is reduced, noise stays the same. If your attenuator is analog then signal is reduced and noise is reduced (and as you mention the attenuator may add a bit of noise). The key that is missing is you also need to consider the residual noise from the amplifier itself.

Michael
 
Last edited:
A legitimate reason to add a stepped attenuator between DAC and amp is safety.

If, for some reason, DAC or my operating system glitches, DAC volume can be set to 0dB. Then, it may destroy speakers or my ears instantly.

A stepped attenuator cannot glitch. It acts as a safety device.
 
Last edited by a moderator:
I've personally found that a decently low noise DAC (110 dB DR at 2 V) paired with a NC252MP results in inaudible noise but I don't run very sensitive speakers.

I currently have E30II (~117dB), an NC252MP being made by Stefan at KJF Audio, and 90db sensitivity speakers (Triangle Borea BR03, 100w/8ohm).

My specific issues are:
a) lack of XLR on the E30II
b) if anything goes wrong w/ volume on dac, a kid messes with it, or it resets in a power cut or something I'll blow my speakers

Thus I feel I want something in the middle that's as neutral as possible to firewall the NC252MPs input, basically in 3 ways: input v, xlr grounding, dc filter - finding such a thing is oddly difficult.

A legitimate reason to add a stepped attenuator between DAC and amp is safety.

Exactly this.
 
Back
Top Bottom