• 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!

Passive I/V stage, why not?

DKT88

Active Member
Joined
Feb 26, 2019
Messages
221
Likes
232
Location
South Korea
I am looking for technical comments on a DAC design that uses resistors and audio transformer (a so called passive no feedback stage) in place of an active output stage. As an example, I was looking at an older Neko D100 DAC (based on the old pcm1794a). The designer makes this claim:

The unique aspect is the passive transformer-based output stage. A passive analog stage because was chosen avoid the possible audio degradation that can happen with an active analog stage. There are many good active designs, but there is also a lot more to analog music reproduction that THD+N measurements because music is much more complicated than sine waves.

Evidence of the passive design's benefits can be seen in the noise floor, which is at the limits of the Audio Precision ATS-2 (this is a measuring instrument that analyzes distortion) at around -130dB. This offers excellent sine-wave linearity even without feedback (as seen in the linearity graph on the web site). I wanted there to be as little as possible influencing the signal fidelity after it comes out of the PCM1794A chip. The output transformer is part of that, as well as providing a nicer solution to removing the PCM1794A output's DC offset than a capacitor which is used in other passive designs.


I can't find any measurements that support his performance claims.
 

SIY

Grand Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
10,386
Likes
24,752
Location
Alfred, NY
There are many good active designs, but there is also a lot more to analog music reproduction that THD+N measurements because music is much more complicated than sine waves.

Evidence of the passive design's benefits can be seen in the noise floor, which is at the limits of the Audio Precision ATS-2 (this is a measuring instrument that analyzes distortion) at around -130dB.

I can't find any measurements that support his performance claims.

"What are three reasons that this company's products should be avoided?"

Ding! Ding!

"OK, I'll take Wire and Cable Scams for $200, Alex."
 
OP
DKT88

DKT88

Active Member
Joined
Feb 26, 2019
Messages
221
Likes
232
Location
South Korea
"What are three reasons that this company's products should be avoided?"

Ding! Ding!

"OK, I'll take Wire and Cable Scams for $200, Alex."
yeah, I spotted the "music is more than sine waves" subjectivist trope. And the mention of noise floor without distortion measurements.
 

SIY

Grand Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
10,386
Likes
24,752
Location
Alfred, NY
yeah, I spotted the "music is more than sine waves" subjectivist trope. And the mention of noise floor without distortion measurements.

Also, "noise floor" is fairly meaningless without a description of test conditions (specifically FFT bin width). Noise floor is not the same as noise. So either they don't know the difference (and hence you should stay away) or they do know the difference and are happy to mislead people (and hence you should stay away).
 
OP
DKT88

DKT88

Active Member
Joined
Feb 26, 2019
Messages
221
Likes
232
Location
South Korea
I am not thinking about buying it, I'm interested in the issues with passive output stages in DACs. I didn't make this clear in the post.
 

SIY

Grand Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
10,386
Likes
24,752
Location
Alfred, NY
Basically, distortion and noise, as you can surmise by the seller's pre-emptive "those things don't matter." An active solution will give a much better virtual ground for the DAC, far lower noise than trying to voltage amplify microvolt and nanovolt signals, and a lower distortion/hysteresis stepup than the transformer.

Note that exactly zero of the highest performance DACs use resistor/transformer combos. That should tell you something.

edit: to clarify my first point, an active solution will use an I-to-V converter, which works off the full current range output of the DAC chip (often 0-2 mA), with typical transimpedances of 1V/mA. A passive solution has to trade off resistor size vs distortion, with smaller resistors giving lower distortion. But they also give lower voltages! So if we drop in, say, a 10R resistor (a common choice here), that 0-2 mA translates into 0-20 mV. So you're faced with the same issue that phono preamp designers have, and getting better than 100dB of S/N or dynamic range is a massive challenge and usually not met.
 
Last edited:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Issues with passive output stages are transformer distortion and bandwidth (which is not really a problem for line level).
One would have to load the output of the transformer correctly but this can be done inside the DAC as any normal load would likely be >10kOhm.
The output voltage will also be load dependent.

The PCM1794 has 7.8mA pp = 2.7mArms will need a 360 Ohm load for 1V.
This will partly be from a load resistor before the transformer and partly be from a load resistor behind the transformer (to get good FR linearity)
The mkII has 2V so the trafo will probably have a different ratio making it more dependent on the load the equipment after the DAC provides.

I suspect that the performance of a good IV stage (using opamps) will be measurably better than that of a transformer.
The plots shown by the manufacturer and the specifications are well... not very revealing and not reaching the performance levels of the DAC chip.
(there are plots below the specifications tab)
This indicates that the 'special' output stage performs less than those pesky opamp stages.
 
Last edited:
OP
DKT88

DKT88

Active Member
Joined
Feb 26, 2019
Messages
221
Likes
232
Location
South Korea
Issues with passive output stages are transformer distortion and bandwidth (which is not really a problem for line level.
One would have to load the output of the transformer correctly but this can be done inside the DAC as any normal load would likely be >10kOhm.
The output voltage will also be load dependent when that .

The PCM1794 has 7.8mA pp = 2.7mArms will need a 360 Ohm load for 1V.
This will partly be from a load resistor before the transformer and partly be from a load resistor behind the transformer (to get good FR linearity)
The mkII has 2V so the trafo will probably have a different ratio making it more dependent on the load the equipment after the DAC provides.

I suspect that the performance of a good IV stage (using opamps) will be measurably better than that of a transformer.
The plots shown by the manufacturer and the specifications are well... not very revealing and not reaching the performance levels of the DAC chip.
(there are plots below the specifications tab)
This indicates that the 'special' output stage performs less than those pesky opamp stages.
ah, I didn't see the plots on his web page.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
I can only make something of the FR plot which is impressive for a transformer (4Hz -3dB)
It looks like it uses a sharp filter so that is nice.
I see a multitone plot that either has all the multitones removed and shows the residual only or the mutlitone is so low in level it says nothing.
Distortion levels are higher than that of a 'normal' PCM1794 DAC.

Anyway the given plots and data make me believe distortions will remain below audibility levels so the DAC will sound just fine.
No output voltage nor output resistance is given. The latter will be unusually high so is probably why it is left out.
The output voltage value is something I saw mentioned in a review (for what's that worth as no measurements were shown)

Unfortunately the given data does not show the 'magic' level in sound quality.
The word of the subjective reviewers means nothing to me.
Guys that swallow the nonsense will be cheaper and better of with this one instead of the TotalCrap DAC.
For the same amount as the Neko one can get RME ADI2 DAC with some nice audio transformers behind it.
Something tells me that is a better deal.
 
Last edited:

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,498
Location
Monument, CO
Looks like the DAC's datasheet specs are measured with active I/V converters FWIW. TI must like them...

A transformer-coupled output probably does not add significant noise, and can provide galvanic isolation (to prevent potential ground loops), but there are a number of other reasons TI and others choose to specify their DACs with an active I/V stage. Here are a few reasons off-the-cuff:
  1. A DAC's output devices are not ideal current sources (or sinks) thus their output varies nonlinearly with load. That means higher distortion when the output is swinging voltage instead of driving a 0-ohm (low-impedance) load like an active I/V converter.
  2. Without an active buffer the output is load-sensitive; the output performance (voltage swing, bandwidth, noise, etc.) depends upon the load impedance. That is true with an active buffer as well, but the low-impedance input and output of a typical I/V stage means perhaps orders of magnitude reduced sensitivity to the load. Note a transformer does not remove load sensitivity though does transform the impedance so may help (or hurt).
  3. An active I/V converter typically incorporates filtering as well to reduce very high-frequency edges and switching spikes from getting to the following components. Without that filter right at the DAC's output, very high frequency switching spikes will hit the transformer, usually leading to ringing and nonlinearity at the output.
  4. On the low-frequency side, transformers do not work down to DC, but rather have some LF cutoff and tend to saturate (distort) with large LF signals. A good transformer will minimize issues in the audio band, natch, but cannot eliminate them entirely.
  5. As with an active circuit, the transformer's design and layout (placement) must be carefully considered to prevent noise/hum pickup. A transformer is a coil of wire around a core (which may be air) and is great at picking up local EM fields and adding them to the output. Good ones have shielding around them to minimize that, natch.
A good transformer design can approach the performance of an active stage and perhaps exceed it for noise if the impedances and output voltage works synergistically with the following stages/components. In general I would expect a transformer-coupled output to perform worse in terms of distortion and bandwidth than an active stage and exhibit much greater sensitivity to the load (cables, input impedance of the next component in the chain, etc.)

FWIWFM - Don
 
OP
DKT88

DKT88

Active Member
Joined
Feb 26, 2019
Messages
221
Likes
232
Location
South Korea
Looks like the DAC's datasheet specs are measured with active I/V converters FWIW. TI must like them...

A transformer-coupled output probably does not add significant noise, and can provide galvanic isolation (to prevent potential ground loops), but there are a number of other reasons TI and others choose to specify their DACs with an active I/V stage. Here are a few reasons off-the-cuff:
  1. A DAC's output devices are not ideal current sources (or sinks) thus their output varies nonlinearly with load. That means higher distortion when the output is swinging voltage instead of driving a 0-ohm (low-impedance) load like an active I/V converter.
  2. Without an active buffer the output is load-sensitive; the output performance (voltage swing, bandwidth, noise, etc.) depends upon the load impedance. That is true with an active buffer as well, but the low-impedance input and output of a typical I/V stage means perhaps orders of magnitude reduced sensitivity to the load. Note a transformer does not remove load sensitivity though does transform the impedance so may help (or hurt).
  3. An active I/V converter typically incorporates filtering as well to reduce very high-frequency edges and switching spikes from getting to the following components. Without that filter right at the DAC's output, very high frequency switching spikes will hit the transformer, usually leading to ringing and nonlinearity at the output.
  4. On the low-frequency side, transformers do not work down to DC, but rather have some LF cutoff and tend to saturate (distort) with large LF signals. A good transformer will minimize issues in the audio band, natch, but cannot eliminate them entirely.
  5. As with an active circuit, the transformer's design and layout (placement) must be carefully considered to prevent noise/hum pickup. A transformer is a coil of wire around a core (which may be air) and is great at picking up local EM fields and adding them to the output. Good ones have shielding around them to minimize that, natch.
A good transformer design can approach the performance of an active stage and perhaps exceed it for noise if the impedances and output voltage works synergistically with the following stages/components. In general I would expect a transformer-coupled output to perform worse in terms of distortion and bandwidth than an active stage and exhibit much greater sensitivity to the load (cables, input impedance of the next component in the chain, etc.)

FWIWFM - Don

The passive stage seems like it is intended to appeal to audiophiles who believe that active stages using opamps with high negative feedback are somehow non-musical sounding. (I don't subscribe to this belief).
 

watchnerd

Grand Contributor
Joined
Dec 8, 2016
Messages
12,449
Likes
10,409
Location
Seattle Area, USA
I am looking for technical comments on a DAC design that uses resistors and audio transformer (a so called passive no feedback stage) in place of an active output stage. As an example, I was looking at an older Neko D100 DAC (based on the old pcm1794a). The designer makes this claim:

The unique aspect is the passive transformer-based output stage. A passive analog stage because was chosen avoid the possible audio degradation that can happen with an active analog stage. There are many good active designs, but there is also a lot more to analog music reproduction that THD+N measurements because music is much more complicated than sine waves.

Evidence of the passive design's benefits can be seen in the noise floor, which is at the limits of the Audio Precision ATS-2 (this is a measuring instrument that analyzes distortion) at around -130dB. This offers excellent sine-wave linearity even without feedback (as seen in the linearity graph on the web site). I wanted there to be as little as possible influencing the signal fidelity after it comes out of the PCM1794A chip. The output transformer is part of that, as well as providing a nicer solution to removing the PCM1794A output's DC offset than a capacitor which is used in other passive designs.

I can't find any measurements that support his performance claims.

Slap some big Hashimoto SUTs on to a NOS R2R DAC

11840652694_f3cb29907e_o.jpg



You know it would get the reviewers all slobbery and sell like hot cakes....
 
Top Bottom