• 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!

Audio Buffer – unity gain link stage/headphone amp with ultimate output drive capability

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
Audio Buffer – unity gain link stage/headphone amp with ultimate output drive capability

I have been using a similar circuit in measuring instruments for HV tests since 1980's and later decided to use it for audio, for its excellent output drive capabilities. I described the reasons in my web page in the year 2002

http://pmacura.cz/buffer_en.html

“Output circuits of standard signal sources as CD players suffer from limited capability to drive capacitive load and low impedance load. Depending on signal cable length (capacitance) and amplifier input resistance/capacitance there is an interaction between output circuits of CD player and the cable.
This may result in a loss of resolution of high frequencies, harsh (grain) sound and soft undefined bass. The following circuit may solve the problem and can be also used as a high quality headphone amplifier. It has high input resistance and low capacitance and low output resistance. It is able to deliver output current of some 250mA. The frequency range is far beyond audio needs. The distortion is very very low, order of 0.0001%. This circuit should be connected to the CD output by coaxial cable no longer than some 10 – 15cm.”


Output impedance of the CD players was usually between 200 ohm and 1 kohm and output stage was a low cost commercial opamp. This has not changed much till nowadays for DAC converters, so the issue often remains the same.

My solution is a small box with unity gain, input impedance of 100 kohm, output impedance of 50 ohm, which is able to drive load down to 50 ohm and almost unlimited capacitance.

audiobuffer.jpg


Here are some measurements

max_out_swing.png

Maximum output voltage is 9Vrms into 10k load, 4.5Vrms into 50 ohms.

square_10kHz.png

Response to 10kHz square wave

square_10kHz_50ohm.png

Response to 10kHz square wave, 50 ohm load

load_47nF.png

Response to 10kHz square wave, 47nF capacitive load

load_1uF.png

Response to 1kHz square wave, 1uF capacitive load

HF_noise.png

Output HF noise up to 10MHz

thd_1kHz_1kohm_log2.png

THD at 1kHz / 1.8Vrms / 1kohm load (LF noise rise is an issue of my measuring gear)

thd_1kHz_50ohm_log2.png

THD at 1kHz / 0.9Vrms / 50 ohm load (LF noise rise is an issue of my measuring gear)

thd_freq_5dBV_1kohm.png

THD vs. frequency at 1.8Vrms / 1kohm

thd_ampl_50ohm_1kohm.png

THD vs. amplitude at 1kHz and 1kohm and 50 ohm load

Conclusion

This circuit eliminates influence of the link cable to output of CD players or DACs. It can be loaded with as low impedance as 50 ohm and capacitance up to 100nF will not degrade its audio parameters. Capacitances up to 1uF (and more) may be driven, at the expense of high frequency roll off.

Most of the distortion measurements of this unit are already affected by resolution of my measuring gear.
 
Last edited:
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
I had to change the 50 ohm load resistor for the professional oscilloscope 50 ohm terminator in the THD 1kHz spectral plot ...., and repeat the measurement.
It is incredible but probably a sort of bad contact (connector??) resulted in unexpected high harmonics in the spectrum, which decreased to a "normal" level after changing the terminating resistor.
 

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
This is very interesting as an exercise, but have you found and measured equipment that needs a buffer of this kind? Do you have examples? A lot of people may suddenly worry that their standard interconnects are actually degrading performance, or that they haven't accounted for the input characteristics of the load (amp, preamp, etc). Your conclusion would be stronger if it clarified that aspect.
 

Speedskater

Major Contributor
Joined
Mar 5, 2016
Messages
1,610
Likes
1,333
Location
Cleveland, Ohio USA
Of course the line output of a hi-fi component and a 50 Ohm load are not compatible. But some headphone amplifiers might be. On the same topic, I wonder why Stereophile magazine also tests line outputs into a 600 Ohm load? It doesn't represent a real world condition.
 

pozz

Слава Україні
Forum Donor
Editor
Joined
May 21, 2019
Messages
4,036
Likes
6,827
Of course the line output of a hi-fi component and a 50 Ohm load are not compatible. But some headphone amplifiers might be. On the same topic, I wonder why Stereophile magazine also tests line outputs into a 600 Ohm load? It doesn't represent a real world condition.
The 600 ohm load used to be the previous standard during the "impedance matching" era. I think it's done more as a curiousity than anything else. If you read @pma's website you'll see that he says at the end that the same circuit can be used as a headphone amplifier, which, as you said, makes it inappropriate for other components.
 

tw99

Senior Member
Joined
Mar 13, 2019
Messages
467
Likes
1,069
Location
West Berkshire, UK
This is very interesting as an exercise, but have you found and measured equipment that needs a buffer of this kind? Do you have examples? A lot of people may suddenly worry that their standard interconnects are actually degrading performance, or that they haven't accounted for the input characteristics of the load (amp, preamp, etc). Your conclusion would be stronger if it clarified that aspect.

Agreed. Given the huge number of snake oil tweaks being sold by the industry that address problems that are either irrelevant, or simply do not exist, it'd be surprising that there's a clear need for a box like this, or we'd be snowed under with them.
 

raindance

Major Contributor
Joined
Sep 25, 2019
Messages
1,037
Likes
968
I don't understand how parallel capacitance, which rolls off high frequencies, can cause harsh, grainy sound or how it can cause rolled off bass. These symptoms are the opposite of the stated issue. You might mention series capacitance also.... but I don't think I've even seen, even, a "high end" cable that introduces series capacitance.

Most line level devices are already designed with low output impedance to drive loads where the input of the other device is many times more of a load than the cable that gets the signal there, but is also at least 10x the output Z of the driving device. In addition, we're in the high impedance world at line level in the consumer business, so unless you have a passive "preamp" or tube preamp whose output impedance is close to that of the load being driven, you should find almost no contribution from the cable unless it is massively long or completely unshielded (like Kimber). Rolled off bass is almost always due to a capacitor coupled output with a too small output capacitor for the load.

Ragged, grainy, nasty sounding treble could be due to instability (oscillation), clipping - you'd be amazed how many people love "golden age" solid state gear that has a 150mV input sensitivity and literally clips when fed CD level audio. Cables don't cause this.

In my opinion, cables are a problem that has already been solved. Let's put energy into convincing the industry to (1) be honest with their specs and (2) to add friggin' tone controls back so people can dial in the "color" they want without spending a fortune on imagined charges from poorly designed expensive cables.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
This is very interesting as an exercise, but have you found and measured equipment that needs a buffer of this kind? Do you have examples? A lot of people may suddenly worry that their standard interconnects are actually degrading performance, or that they haven't accounted for the input characteristics of the load (amp, preamp, etc). Your conclusion would be stronger if it clarified that aspect.

Yes, of course. The following units that I own have output impedance of about 500 ohm

https://www.audiosciencereview.com/...-of-technics-sl-pg100-vintage-cd-player.9964/
https://www.audiosciencereview.com/...-onkyo-dv-sp503e-universal-player-2005.10016/

In fact my dissatisfaction with the Technics about 20 years ago was the impulse that started the development of the Audio Buffer.

Many existing soundcards have output impedance about 1kohm. As an example, my Roland Duo-Capture Ex. 1kohm is too much. When I send the signal from one room to another one through a 10m of cable, the Audio Buffer makes a great improvement.


I don't understand how parallel capacitance, which rolls off high frequencies, can cause harsh, grainy sound or how it can cause rolled off bass. These symptoms are the opposite of the stated issue. You might mention series capacitance also.... but I don't think I've even seen, even, a "high end" cable that introduces series capacitance.

Most line level devices are already designed with low output impedance to drive loads where the input of the other device is many times more of a load than the cable that gets the signal there, but is also at least 10x the output Z of the driving device. In addition, we're in the high impedance world at line level in the consumer business, so unless you have a passive "preamp" or tube preamp whose output impedance is close to that of the load being driven, you should find almost no contribution from the cable unless it is massively long or completely unshielded (like Kimber). Rolled off bass is almost always due to a capacitor coupled output with a too small output capacitor for the load.

Ragged, grainy, nasty sounding treble could be due to instability (oscillation), clipping - you'd be amazed how many people love "golden age" solid state gear that has a 150mV input sensitivity and literally clips when fed CD level audio. Cables don't cause this.

In my opinion, cables are a problem that has already been solved. Let's put energy into convincing the industry to (1) be honest with their specs and (2) to add friggin' tone controls back so people can dial in the "color" they want without spending a fortune on imagined charges from poorly designed expensive cables.

It is not only the ability to drive higher cable capacitance, but also acting as a LP filter for the frequencies in MHz and tens of MHz present at output of digital sources, as I have already shown several times in the example of DacMagic Plus. The unit like Audio Buffer which contains very fast components and LP 300kHz filter effectively filters these spikes and prevents the amplifier from possible input slew rate limitation.


Of course the line output of a hi-fi component and a 50 Ohm load are not compatible. But some headphone amplifiers might be. On the same topic, I wonder why Stereophile magazine also tests line outputs into a 600 Ohm load? It doesn't represent a real world condition.

Because 600 ohm is a PA link standard.

BTW, the best immunity to RF is when I use 50 ohm coax cable, driven from Audio Buffer (which has 50 ohm output impedance) and the cable is terminated by a 50 ohm resistor. As you could see, the Audio Buffer happily drives 50 ohm without rise of distortion.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
One more set of measurements to demonstrate loopback distortion of my measuring system and what is "added" by the Audio Buffer.

1st, loopback THD at 1kHz
Spectrum_loop.png

Base frequency amplitude is -3.33dB, H2 is -115.38dB and H3 is -105.30dB

2nd, same system with Audio Buffer inserted in the loop
Spectrum_1kHz.png

Base frequency amplitude is -3.38dB, H2 is -113.6dB and H3 is -105.16dB

So, in fact the Audio Buffer copies my system's distortion, with very very small addition of H2. H2 of the system loop -112.05dBr (0.00025%), H2 with Audio Buffer is -110.22dBr (0.00031%).
Noise is unaffected.

Noise level measured is less than -110dBV(A), this makes S/N = 130dB(A) to full output.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
The next 2 plots show how the Audio Buffer reduces HF noise and spikes present at the output of the DacMagic Plus D/A converter:

DM+_RCAout.png


DM+_RCAout_buffer.png


This clearly shows that though the Audio Buffer has wide bandwidth (300kHz), it greatly reduces HF stress and reduces possibility of SR input limitation in the amplifier that follows behind the DAC (or the CD player).
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
If the dac magic is producing that level of random noise, 30mV pk to pk, with a normal load then it is a broken design. Throw it in the trash.

Also, an amp fed with the dac is also unlikely to have an unfiltered response to 300kHz so it will get filtered anyway.
 
Last edited:

Speedskater

Major Contributor
Joined
Mar 5, 2016
Messages
1,610
Likes
1,333
Location
Cleveland, Ohio USA
Because 600 ohm is a PA link standard.
That's a legacy standard from well over a half century ago. The 600 Ohm output stages were small power amplifiers, to drive the 600 Ohm inputs.
BTW, the best immunity to RF is when I use 50 ohm coax cable, driven from Audio Buffer (which has 50 ohm output impedance) and the cable is terminated by a 50 ohm resistor. As you could see, the Audio Buffer happily drives 50 ohm without rise of distortion.
Because the Radio Frequency Characteristic Impedance of that cables is 50 Ohms only above about 1 MHz, this is a common RF technique, but it has little to do with audio.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
Because the Radio Frequency Characteristic Impedance of that cables is 50 Ohms only above about 1 MHz, this is a common RF technique, but it has little to do with audio.

Yes and above 1MHz is the issue of digital audio output. I have measured numerous "audio" cables and have results in my archive, the best are industrial RG-58 and RG-59 terminated at both sides by its characteristic impedance. This assures lowest reflections and lowest HF interferences and noise. If you have measurements as well then we may discuss, otherwise I will not reply to word speculations. I guess you have no measurements of HF noise in audio cables. Audio engineers are often narrow minded and concentrate their analysis to 50kHz max. Then they wonder why digital audio products sound different and try to improve audio band distortion parameters to -130dBr or so, which is absolutely pointless. Very few have deeper insight.
 

LTig

Master Contributor
Forum Donor
Joined
Feb 27, 2019
Messages
5,760
Likes
9,442
Location
Europe
Yes and above 1MHz is the issue of digital audio output. I have measured numerous "audio" cables and have results in my archive, the best are industrial RG-58 and RG-59 terminated at both sides by its characteristic impedance. This assures lowest reflections and lowest HF interferences and noise. If you have measurements as well then we may discuss, otherwise I will not reply to word speculations. I guess you have no measurements of HF noise in audio cables. Audio engineers are often narrow minded and concentrate their analysis to 50kHz max. Then they wonder why digital audio products sound different and try to improve audio band distortion parameters to -130dBr or so, which is absolutely pointless. Very few have deeper insight.
I'm puzzled about what to make of your statement. Do you talk about analog signals coming out of a DAC, or about digital signals feeding a DAC?
WRT the former: an output stage with a high source impedance creates a natural low pass filter with the connected cable. Adding a low impedance highspeed output would prevent this so if you really want to get rid of RF noise then such a stage would be counterproductive.
WRT the latter: adding an output stage meant for analog signals to an existing SPDIF Coax (75 Ohm) or AES output (110 Ohm) makes no sense, except the output of the feeding source is broken. Of course you have to use cables with the matching wave impedance.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,580
Likes
38,283
Location
Gold Coast, Queensland, Australia
My solution is a small box with unity gain, input impedance of 100 kohm, output impedance of 50 ohm, which is able to drive load down to 50 ohm and almost unlimited capacitance.

My solution is this slightly bigger box. It's 28 years old. It has a 1Hz-500KHz bandwidth, distortion below anything I can measure, can swing 20V and has an output impedance of 1 ohm (RCA). (50ohm 25/25 on the balanced xlr) and @110dB S/N on the RCAs, I'm close to the limit of any CD player's residual.

(brochure shot)
1576447216627.png


It's called a preamplifier Pavel. You should try one sometime ;)

1576447092207.png
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,580
Likes
38,283
Location
Gold Coast, Queensland, Australia
Noise level out of my Motu 8A

Your scale is 10mV per division and the reading is hovering around 1-3 graduations which is 1-3mV. (1000-3000uV) That's an awful lot of noise, even if you have a x10 mag switch on.

A typical D/A converter is in the very low single digit uV. Something is not right (noise/bandwidth etc).
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Your scale is 10mV per division and the reading is hovering around 1-3 graduations which is 1-3mV. (1000-3000uV) That's an awful lot of noise, even if you have a x10 mag switch on.

A typical D/A converter is in the very low single digit uV. Something is not right (noise/bandwidth etc).
I was just keeping same conditions as PMA.

An 8 bit scope is useless for measuring accurate low level noise levels. This was just a comparison to PMAs post and data to show I dont see massive spikes of noise as he is.
 
Last edited:

LTig

Master Contributor
Forum Donor
Joined
Feb 27, 2019
Messages
5,760
Likes
9,442
Location
Europe
Noise output of my RME ADI-2 PRO fs, measured with Siglent SDS 1202X-E DSO

EDIT: all measurements shown below are wrong and do not show the behaviour of the RME ADI-2 PRO fs. @MC_RME gave some hints in posting #21 and when I repeated the measurements with the RME on battery (see posting #24) I noticed that the DSO shows noise and spikes even if the RME is powered off and disconnected from power.:facepalm:

The DAC is connected via USB to the PC I'm using right now. Volume is off (one step below-96.3 dB), Samplerate 44.1 kHz, no audio signal playing.

Unbalanced Out
Yellow=left channel, pink = right channel, 100us/div, 10 mV/div:

ADI2proFS_noise_unbal-RL_dso_100us_10mV.png


Oops - spikes and RF pulses. Lets increase sensitivity to 2mV/div:

ADI2proFS_noise_unbal-RL_dso_100us_2mV.png


Yep. Lets zoom in to the RF pulse in the middle (2 us/div):

ADI2proFS_noise_unbal-RL_dso_2us_2mV.png


Let's zoom in more and measure its frequency (20ns/div):

ADI2proFS_noise_unbal-RL_dso_20ns_2mV.png


With the cursors covering 10 wavelengths the frequency is 10 times as displayed = 68.3 MHz, level left channel ~5mVpp..

Let's look at one of those short spikes:

ADI2proFS_spike_unbal-RL_dso_100ns_2mV.png


That's a ringing pulse. Zoom in to measure its frequency:

ADI2proFS_spike_unbal-RL_dso_10ns_2mV.png

Frequency is around 180 MHz.

I'm not sure what to make of this. Is it a measuring artifact? Do i measure something wrong?

XLR out (right channel)

I measure both signals against ground and use the math function of the DSO to show the difference. The noise was much higher :oops: so I started with 2ms/div and 20 mV/div:

ADI2proFS_noise_XLR-R_dso_2ms_20mV.png


The level of the difference is 40 mVpp and quite lower than the singular signals. When we zoom in (5 us/div) we find only spikes at a distance of 14.4 us (repetition rate 69 kHz):

ADI2proFS_spike_XLR-R_dso_5us_20mV.png


Zoom in to the spike in the middle (50 ns/div) :

ADI2proFS_spike_XLR-R_dso_50ns_20mV.png


It's an RF pulse, 87 MHz at 40 mVpp.

Side Notes
All results are independent of volume and output reference level. Pulling the USB connection has no influence either.

Since I bought it not only for music but also for measurements I'm a bit puzzled what to think about this. Maybe I do something wrong. I'd like @MC_RME to have a look and correct any failures I've possible done.

I'm sure nothing of this is audible at all; the frequencies are much too high and will be swamped in the first stage of any audio amplifier if it reaches them in the first place. But for measuring high gain amplifiers like phono or microphone preamplifiers you are well advised to set the DAC to a high output voltage and use a voltage divider to create the desired input level; just using a volume of -50 dB will give you a nice low level signal but it will be swamped by those high frequency noise and spikes and I don't think this to be a valid input signal for measureing those preamps.
 
Last edited:
Top Bottom