Soekris dac2541 technical measurements

Discussion in 'Source Measurements' started by atomicbob, Feb 1, 2021.

  1. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    Soekris dac2541 technical measurements

    Standard Prologue
    If you are unfamiliar with audio measurements please use a search engine with the query:
    "audio measurements" or "audio measurement handbook"
    Look for publications by Richard C. Cabot and also by Bob Metzler, both from Audio Precision. There are other useful publications as well. These will provide basic knowledge.
    Interpretation of the following measurements is beyond the scope of technical measurements posts.


    The data presented were collected as follows:

    1. PrismSound dScope III, picoscope 5243B, Cal Test CT2593-2 balanced probe
    2. Tecnec 75R spdif cable
    3. Balanced XLR cables Belden 1800F with Neutrik 110R AES connectors
    4. Single Ended cables Mogami 2964 1 meter with Amphenol RCA connectors
    5. 100 Kohm load used for measurements
    6. dScope analyzer sample rate 48 KHz unless otherwise noted
    7. DAC 44.1 KHz sample rate, 24 bit depth unless otherwise noted
    9. Audioquest Forest and Schiit Pyst USB cables
    9. Vaunix Lab Brick USB hub
    10. Shielded 14AWG and 16AWG power cables
    11. Lin (red) filter mode unless otherwise noted

    Measurements are made in accordance with AES17:2015

    Soekris dac2541
    Measurements for record commenced after warmup of at least 72 hours.
    Measurements were performed over a period of several months.
    Balanced measurements were performed twice, one month apart for consistency verification.

    Index
    Post 1 - measurement setup description, highlights
    Post 2 - AES input - Bal XLR outputs
    Post 3 - AES input - SE RCA outputs
    Post 4 - USB ASIO input - Bal XLR outputs
    Post 5 - USB ASIO input - SE RCA outputs
    Post 6 - Filter response part 1
    Post 7 - Filter response part 2
    Post 8 - Gain Staging part 1
    Post 9 - Gain Staging part 2 and Jitter Attenuation
    Post 10 - reserved for additional data and corrections

    Notable highlights:
    dac2541 demonstrates excellent performance in a small package and a properly designed SMPS is nothing to fear.
    Very low distortion
    Power supply noise spectrum are at or below -160 dBFS in balanced output
    The power supply spectrum residual noise was so low it was necessary to reconfigure y-axis scale low enough to see the noise floor.
    Balanced output Dynamic Range of 126 dB
    Balanced output Cross-talk is dual mono with > 145 dB isolation
    Balanced output Gain Linearity is nearly perfect to -110 dBFS, less than ±1 dB to -130 dBFS
    Exceptionally low jitter
    Single Ended performance is slightly less than Balanced output, as is typical
    Single Ended performance is still exceptional compared to many competitor DAC SE outputs

    Well done Søren!

    Commentary:
    Power supply is at least 50% of any audio component's design. The higher the level of performance the more the power supply matters. Soekris recognizes this, enhancing dac2541's power supply, keeping noise to 100s of nanovolt levels while supplying 10s of volts throughout the system. Exceptional performance by design.

    Measurement setup example:
    DSC_0597_small.jpg

    Listening picture example:
    DSC_0594_small.jpg

    A time allocation dilemma occurs when components of such high caliber are in the lab for evaluations and measurements. Listening can become so involving as to lose track of time. In the blink of an eye an allocated time slot is consumed. dac2541 kept me immersed listening with various amps. Two favorites appear in the picture above.
     
    • Like / Agree Like / Agree x 35
    • List
    Last edited: Feb 1, 2021
  2. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    AES input Bal output measurements


    dac2541 Dynamic Range AES input Bal output
    01 20210103 dac2541 dynamic range FFT AES Bal watermarked.png
    Excellent performance

    Dynamic range, in simplified terms
    1. find maximum output voltage at 0 dBFS
    2. find residual noise with a -60 dBFS 1 KHz stimulus, removed from analysis by window notch
    3. Dynamic Range is the difference between maximum output and residual noise
    Specifics are in AES17 section 9.3 (measurebators do your homework please)
    Example here:
    https://www.maximintegrated.com/en/design/blog/spec-dynamic-range.html

    Why we have a -60 dBFS stimulus
    Some clever codec designers include an output gate that shuts off when there is no signal present. This has the dual benefit of lowering output idle noise while also gaming the measurements. AES became wise requiring a -60 dBFS stimulus so any such gates are open during measurement and true Dynamic Range performance measured.


    dac2541 -120dBFS sine AES input Bal output
    02 FFT_0_L1T18_1.png

    dac2541 A04 THD+N THD nth-HD FFT AES input Bal output
    03 20210103 dac2541 A04 THD+N THD nth-HD 4+HD+N 60Hz FFT Bal 0dBFS - AES.png

    dac2541 50 + 7000 Hz AES input Bal output - Left Channel
    04 FFT_2_L1T13_1_A.png

    dac2541 Gain Linearity AES input Bal output - Left Channel
    05 G2_L1T3_1_A.png

    dac2541 THD+N vs Frequency AES input Bal output - Left Channel
    06 G2_L1T5_1_A.png

    dac2541 Residual Noise Bal AES input Bal output
    07 20210103 dac2541 residual noise FFT AES Bal - 160 dB range watermarked.png

    dac2541 Residual Noise Bal AES input Bal output - 180 dBFS range
    08 20210103 dac2541 residual noise FFT AES Bal - 180 dB range watermarked.png
    It was necessary to extend the low end of the x-axis down another 20 dBFS to visually assess FFT of residual noise.
    I think Søren may safely claim a properly designed SMPS can be exceptionally noise free as demonstrated by his dac2541 design.

    dac2541 12 multitone AES input Bal output - Left Channel
    09 20210106 dac2541 12 multitone AES Bal Lin.png

    Complete AES input Bal output analysis report pdf attached
     

    Attached Files:

    • Like / Agree Like / Agree x 15
    • List
    Last edited: Feb 1, 2021
  3. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    AES input SE output measurements

    dac2541 Dynamic Range AES input SE output
    01 20210124 dac2541 dynamic range FFT AES SE.png

    dac2541 A04 THD+N THD nth-HD FFT AES input SE output
    02 20210124 dac2541 A04 THD+N THD nth-HD 4+HD+N 60Hz FFT SE 0dBFS - AES - Lin.png

    dac2541 50 + 7000 Hz AES input SE output - Left Channel
    03 FFT_2_L1T13_1_A.png

    dac2541 Gain Linearity AES input SE output - Left Channel
    04 G2_L1T3_1_A.png

    dac2541 THD+N vs Frequency AES input SE output - Left Channel
    05 G2_L1T5_1_A.png

    dac2541 Residual Noise AES input SE output
    06 20210124 dac2541 residual noise FFT AES SE - 160 dB range.png

    dac2541 12 multitone AES input SE output - Left Channel
    07 20210124 dac2541 12 multitone AES SE Lin.png

    Complete AES input SE output analysis report pdf attached
     

    Attached Files:

    • Like / Agree Like / Agree x 11
    • List
    Last edited: Feb 1, 2021
  4. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    USB ASIO input Bal output measurements

    dac2541 Dynamic Range USB input Bal output
    01 20210125 dac2541 dynamic range FFT ASIO Bal.png

    dac2541 -120dBFS sine USB input Bal output
    02 FFT_0_L1T18_1.png

    dac2541 A04 THD+N THD nth-HD FFT USB input Bal output
    03 20210125 dac2541 A04 THD+N THD nth-HD 4+HD+N 60Hz FFT Bal 0dBFS - ASIO.png

    dac2541 50 + 7000 Hz USB input Bal output - Left Channel
    04 FFT_2_L1T13_1_A.png

    dac2541 Gain Linearity USB input Bal output - Left Channel
    05 G2_L1T3_1_A.png

    dac2541 THD+N vs Frequency USB input Bal output - Left Channel
    06 G2_L1T5_1_A.png

    dac2541 Residual Noise Bal USB input Bal output
    07 20210125 dac2541 residual noise Bal FFT ASIO Bal - 160 dB range.png

    dac2541 12 multitone USB input Bal output - Left Channel
    08 20210125 dac2541 12 multitone ASIO Bal - Lin.png

    Complete USB input Bal output analysis report pdf attached
     

    Attached Files:

    • Like / Agree Like / Agree x 9
    • List
    Last edited: Feb 1, 2021
  5. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    USB ASIO input SE output measurements

    dac2541 Dynamic Range USB input SE output
    01 20210125 dac2541 dynamic range FFT USB SE.png

    dac2541 A04 THD+N THD nth-HD FFT USB input SE output
    02 20210125 dac2541 A04 THD+N THD nth-HD 4+HD+N 60Hz FFT SE 0dBFS - USB.png

    dac2541 50 + 7000 Hz USB input SE output - Left Channel
    03 FFT_2_L1T13_1_A.png

    dac2541 Gain Linearity USB input SE output - Left Channel
    04 G2_L1T3_1_A.png

    dac2541 THD+N vs Frequency USB input SE output - Left Channel
    05 G2_L1T5_1_A.png

    dac2541 Residual Noise USB input SE output
    06 20210125 dac2541 residual noise FFT USB SE - 160 dB range.png

    dac2541 12 multitone USB input SE output - Left Channel
    07 20210201 dac2541 12 multitone ASIO SE - Lin.png

    Complete USB input SE output analysis report pdf attached
     

    Attached Files:

    • Like / Agree Like / Agree x 9
    • List
    Last edited: Feb 1, 2021
  6. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    Filter Response part 1

    Measurement conditions
    1. dac2541 gain set to -4 dBFS
    2. 44.1 KHz sample rate except where noted
    3. USB ASIO input
    4. SE output to oscilloscope or dScope

    Filter step responses are independent of inputs and outputs.

    dac2541 10mS/div Lin (red)
    01 20210125 dac2541 20 Hz square 3682mVpp 100uS div 5MHz filter AES SE Lin BW calc.png
    dac2541 100uS/div Lin BW calculation
    Bandwidth estimation: BW (MHz) = 0.35 / RT (mS)
    Where RT = 10 to 90% Rise Time
    0.35 / 17.72 uS = 19.75 KHz

    dac2541 10mS/div Mix (org)
    02 20210125 dac2541 20 Hz square 3682mVpp 100uS div 5MHz filter AES SE Mix BW calc.png
    dac2541 100uS/div Mix BW calculation
    Bandwidth estimation: BW (MHz) = 0.35 / RT (mS)
    Where RT = 10 to 90% Rise Time
    0.35 / 20.96 uS = 16.70 KHz

    dac2541 10mS/div OS Min (grn)
    03 20210125 dac2541 20 Hz square 3682mVpp 100uS div 5MHz filter AES SE Min BW calc.png
    dac2541 100uS/div Min BW calculation
    Bandwidth estimation: BW (MHz) = 0.35 / RT (mS)
    Where RT = 10 to 90% Rise Time
    0.35 / 44.21 uS = 7.92 KHz

    dac2541 10mS/div Soft (blk)
    04 20210125 dac2541 20 Hz square 3682mVpp 100uS div 5MHz filter AES SE Slow BW calc.png
    dac2541 100uS/div Soft BW calculation
    Bandwidth estimation: BW (MHz) = 0.35 / RT (mS)
    Where RT = 10 to 90% Rise Time
    0.35 / 12.25 uS = 28.57 KHz

    From the BW estimates above Slow filter has the fastest transient response (highest bandwidth) while Min filter has the slowest by a very large margin, 28.57 KHz BW vs 7.92 KHz BW.

    dac2541 Imaging all four filters AES input
    11 20210103 dac2541 imaging FFT AES Bal - 4 filters.png

    dac2541 Imaging all four filters ASIO input
    12 20210125 dac2541 imaging FFT ASIO Bal - 4 filters.png

    Slow filter mode roll off behavior is markedly different between AES and ASIO inputs. The other filter modes are substantially similar between the two inputs. ASIO Imaging is necessarily an asynchronous measurement and thus is noisier than the synchronous AES measurement.

    dac2541 Imaging all four filters AES input zoom
    13 20210103 dac2541 imaging FFT AES Bal - 4 filters zoom 1.png

    dac2541 Imaging all four filters ASIO input zoom
    14 20210125 dac2541 imaging FFT ASIO Bal - 4 filters zoom.png

    dac2541 Imaging Slow filter mode ASIO input SE output at 44K, 48K, 88K and 96K sample rates
    30 20210128 dac2541 Imaging 44 48 88 96 ASIO SE Slow.png

    As sample rate is increased, the Slow filter mode slope remains constant. However note how reduction lowers at 44KHz from approximately 26 dB for 44KHz to approximately 56 dB at 88 KHz, an improvement of 30 dB sample rate supression. This will be important in the next two sections.
     
    • Like / Agree Like / Agree x 12
    • List
    Last edited: Feb 1, 2021
  7. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    Filter Response part 2

    Measurement conditions
    1. dac2541 gain set to 0 dBFS
    2. dScope signal generator set to -20 dBFS
    3. 44.1 KHz sample rate except where noted
    4. USB ASIO input except where noted
    5. SE output to oscilloscope or dScope

    At -20 dBFS dac2541 SE output will be approximately 570 mVpp. A headphone amplifier running 0 dB gain would mirror this output level which is also approximately -12 dBu. HD600 and HD650 will produce 88 dBSPL at this level.

    dac2541 imaging -20 dBFS FFT AES 44K SE Slow
    15 20210129 dac2541 imaging -20 dBFS FFT AES 44K SE Slow.png
    The cursor at 44.1 KHz indicates -30.66 dB reduction of sample rate components for AES input.

    dac2541 imaging -20 dBFS FFT ASIO 44K SE Slow
    16 20210129 dac2541 imaging -20 dBFS FFT ASIO 44K SE Slow.png
    The cursor at 44.1 KHz indicates -25.77 dB reduction of sample rate components for ASIO input.

    dac2541 40 Hz sine FFT -20dBFS 44K USB SE Slow
    21 20210128 dac2541 40 Hz sine FFT -20dBFS 44K USB SE Slow.png
    Using a large number of averages, FFT indicates primary 44 KHz component is 32 dB lower than the 40 Hz stimulus. When all the other spikes are incorporated into the signal the result will be 4 to 6 dB higher, around 26 to 28 dB below the 40 Hz stimulus.

    dac2541 40 Hz sqr 100us div -20dBFS 44K USB SE Slow
    22 20210128 dac2541 40 Hz sqr 100us div  -20dBFS 44K USB SE Slow.png
    From the 40 Hz square wave response a very short damped ring is observed before settling into the 44 KHz oscillation. From the downward slope of the Overview window we may infer a DC blocking filter is in the signal path.

    dac2541 20 Hz sine 570mVpp 20uS div 5MHz filter USB SE Slow
    23 20210127 dac2541 20 Hz sine 570mVpp 20uS div 5MHz filter USB SE Slow BW calc.png
    Switching to a 40 Hz sine and zooming to a relatively level portion we observe the 44 KHz sine has an amplitude of 29mVpp which is 26 dB lower than the 570 mVpp average sine level. This is consistent with both the dScope Imaging measurement and the oscilloscope FFT measurement.

    dac2541 40 Hz sine FFT -20dBFS 88K USB SE Slow
    24 20210128 dac2541 40 Hz sine FFT -20dBFS 88K USB SE Slow.png
    Changing sample rate to 88 KHz the averaged FFT indicates the primary 88 KHz component is 57 dB below the 40 Hz stimulus, an additional reduction of 31 dB as compared to the filter characteristic at 44 KHz.

    dac2541 40 Hz sqr 100us div -20dBFS 88K USB SE Slow
    25 20210128 dac2541 40 Hz sqr 100us div  -20dBFS 88K USB SE Slow - no markers.png
    88 KHz ripple is only very slightly apparent in this time domain view, much cleaner than at 44 KHz. The filter ring is much shorter due to the higher sample rate, an additional bonus.
     
    • Like / Agree Like / Agree x 12
    • List
    Last edited: Feb 1, 2021
  8. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    Gain Staging part 1

    Measurement conditions
    1. dac2541 gain set to 0 dBFS
    2. 44.1 KHz sample rate except where noted
    3. AES input
    4. SE output to oscilloscope or dScope

    dac2541 40 Hz sine Lin (red)
    20210131-01 dac2541 40 Hz sine 5847mVpp 5mS div 5MHz filter AES SE Lin.png

    dac2541 40 Hz sine Mix (org)
    20210131-02 dac2541 40 Hz sine 5847mVpp 5mS div 5MHz filter AES SE Mix.png

    dac2541 40 Hz sine Min (grn)
    20210131-03 dac2541 40 Hz sine 5847mVpp 5mS div 5MHz filter AES SE Min.png

    At 0 dBFS SE output is 5.847 Vpp and the dac2541 clip indicator LED remains dark for the above three filter modes.

    dac2541 40 Hz sine Slow (blk)
    20210131-04 dac2541 40 Hz sine 5847mVpp 5mS div 5MHz filter AES SE Slow.png

    At 0 dBFS SE output + 44 KHz oscillation has a higher output and the dac2541 clip indicator illuminates.

    dac2541 40 Hz sine Slow zoom
    20210131-05 dac2541 40 Hz sine 5847mVpp 5mS div 5MHz filter AES SE Slow Zoom.png

    dac2541 40 Hz square Lin
    20210131-09 dac2541 40 Hz sqr 5847mVpp 5mS div 5MHz filter AES SE Lin 0 dBFS.png

    dac2541 40 Hz square Mix
    20210131-10 dac2541 40 Hz sqr 5847mVpp 5mS div 5MHz filter AES SE Mix 0 dBFS.png

    dac2541 40 Hz square Min
    20210131-11 dac2541 40 Hz sqr 5847mVpp 5mS div 5MHz filter AES SE Min 0 dBFS.png

    dac2541 40 Hz square Slow
    20210131-08 dac2541 40 Hz sqr 5847mVpp 5mS div 5MHz filter AES SE Slow 0 dBFS t3.png

    At 0 dBFS SE output, observe varying amounts of clipping for 40 Hz square wave response. The clipping amount is independent of inputs and outputs.

    The above examples indicate gain stage design choices of where to set analog output with respect to 0 dBFS signal input. Many DAC designers make similar tradeoffs for their respective designs. In the next section some useful guidelines to avoid clipping if listening to heavily loudness enhanced source material is presented.
     
    • Like / Agree Like / Agree x 14
    • List
    Last edited: Feb 2, 2021
  9. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    Gain Staging part 2

    Measurement conditions
    1. dScope signal generator set to 0 dBFS
    2. dac2541 volume control set as indicated for each filter mode
    3. 44.1 KHz sample rate except where noted
    4. AES input
    5. SE output to oscilloscope or dScope

    dac2541 40 Hz square Lin (red) - 3 dBFS volume setting
    20210131-21 dac2541 40 Hz sqr 5847mVpp 5mS div 5MHz filter AES SE Lin -3 dBFS.png

    dac2541 40 Hz square Mix (org) - 3 dBFS volume setting
    20210131-22 dac2541 40 Hz sqr 5847mVpp 5mS div 5MHz filter AES SE Mix -3 dBFS.png

    dac2541 40 Hz square Min (grn) - 5 dBFS volume setting
    20210131-23 dac2541 40 Hz sqr 5847mVpp 5mS div 5MHz filter AES SE Min -5 dBFS.png

    dac2541 40 Hz square Slow (blk) - 4 dBFS volume setting
    20210131-24 dac2541 40 Hz sqr 5847mVpp 5mS div 5MHz filter AES SE Slow -4 dBFS.png

    Each dac2541 volume control setting above was the highest achievable before square wave filter ring clipping occurred and the clip indicator illuminated. Keep in mind this assumes highly loudness compressed source material with fast transients. Non "loudness war enhanced" material is likely to be safe with volume control set to -1 dBFS. Soekris digital attenuation math is very good, one shouldn't fear using it.

    Søren has created a flexible, precision dac allowing the user to customize tuning for specific system optimization. The above information should help in making choices for personal use cases.


    Jitter Attenuation

    Measurement conditions
    1. dac2541 gain set to -1 dBFS
    2. 44.1 KHz sample rate
    3. AES input
    4. Bal output to dScope

    dac2541 Jitter Attenuation AES in Bal out
    20210104 dac2541 jitter attenuation AES Bal watermarked.png

    Another example of design excellence is the inclusion of a very capable jitter attenuation system as observed in the measurement above.

    Jitter attenuation vs. frequency is indicated by the yellow line.
     
    • Like / Agree Like / Agree x 20
    • List
    Last edited: Feb 2, 2021
  10. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    reserved for additional data and corrections
     
    • Like / Agree Like / Agree x 2
    • List
  11. Baten

    Baten Friend

    Friend
    Joined:
    Mar 18, 2018
    Likes Received:
    882
    Trophy Points:
    93
    Location:
    EU
  12. soekris

    soekris MOT - Soekris Engineering

    Friend
    Joined:
    Mar 25, 2016
    Likes Received:
    902
    Trophy Points:
    93
    Location:
    Denmark
    Thanks, and thanks to atomicbob for the great work, it must have taken some time....
     
    • Like / Agree Like / Agree x 7
    • List
  13. brams

    brams Rando

    Joined:
    Oct 24, 2015
    Likes Received:
    10
    Trophy Points:
    3
    Great set of measurements! Amazing to see the level of performance currently availabe at reasonable price points

    Has anyone compared the dac performance of the Soekris to that of the Okto dac8? Thinking it makes more sense to pick up one of these than upgrade my Ayre qb 9.
     
  14. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    Loaner tour dac2541 SN13 appears to be operating normally.

    There are some variances between #13 and my lab unit but nothing I would consider out of the ordinary unit to unit variance.

    A couple of AES input SE output examples (#13 vs my unit):
    Dynamic Range: 121.5 vs 125 dB
    *edited for transcription errors - never post when really tired*
    IMD: -95 vs -107 dB
    IMD: -93 vs -103 dBFS
    2nd harmonic distortion: -93 vs -102 dB
    D2 -91 vs -97 dBFS

    Something to consider. Single Ended output RCA center pin is connected directly to XLR pin 2 for each channel. I measured the resistance at 0.04 ohms using a Keysight 34465A with Kelvin 4-wire probes.

    dScope setup:
    DSC_0601_small.jpg


    DSC_0600_small.jpg

    Measurement reports attached in the following order (all 44.1K Linear filter):
    1) AES input Bal output
    2) AES input SE output
    3) ASIO input Bal output
    4) ASIO input SE output[/S]
     

    Attached Files:

    • Like / Agree Like / Agree x 8
    • List
    Last edited: Feb 22, 2021
  15. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Friend MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    11,783
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    While I'm not necessarily debating the audibility of these differences, at first glance these discrepancies would give me pause take a closer look at the two units, my measurement rig and methods, etc.

    Admittedly, I haven't yet been able to dig into the attached reports. Only going off what I quoted.
     
    • Like / Agree Like / Agree x 2
    • List
  16. Vtory

    Vtory Illogical Spock

    Friend MZR
    Joined:
    Sep 12, 2016
    Likes Received:
    7,561
    Trophy Points:
    113
    Location:
    Greater Philadelphia Area, PA
    ^
    This.

    Let's first read what are available... That's internet 101 I was schooled.

    For lazy people, I put together into a single table. There are more numbers but I hugely felt doing more unnecessary...

    upload_2021-2-22_16-26-22.png

    As a day to day stochastic guy, I can't say any of difference meaningful enough to challenge.

    Typo in the last two rows. They should be..

    FFT residual noise A -108.844 -109.740 -108.031 -108.113
    FFT residual noise B -110.830 -105.144 -107.798 -108.120

    PS. I believe we all easily distinguish dbu, db, percentages.
     
    • Like / Agree Like / Agree x 3
    • List
    Last edited: Feb 22, 2021
  17. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    I knew it was a mistake to post quickly the greatest variances I found for the left channel only. @Vtory beat me in tabulating a larger number of measures demonstrating greater similarity than differences.

    Here are the same measures I hastily posted above for both left and right channels converted from % to dB:
    20210222 loaner vs lab unit AES to SE.png

    One might ask why my lab unit Left channel IMD is so much better than the rest. Keep in mind -93 dBFS for dac2541 is 41 uV while -103 dBFS is 13 uV, where 0 dBFS in this case is 7.49 dBu or 1.8357 Vrms. The dac2541 has digital attenuation set for -1 dB.
     
    • Like / Agree Like / Agree x 6
    • List
  18. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    dac2541 #13 (loaner) change after seven days of power on time.
    44.1 KHz 24 bit, Filter set for linear (red), digital attenuator set to -01

    AES input SE output Distortion, first measurements on left, seven days later on right.
    dac2541-13_SE_Dist_Compare_20210220_20210227.png
    Generally all distortions were lower, some significantly, with exception of right channel 3rd harmonic which measured slightly higher.

    AES input SE output IMD, first measurements on left, seven days later on right.
    dac2541-13_IMD_Compare_20210220_20210227.png
    Also notable improvement after seven days.

    Full reports for AES input, Bal and SE output attached.
     

    Attached Files:

    • Like / Agree Like / Agree x 3
    • Epic Epic x 3
    • List
  19. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    dac2541 Power On warm-up series #2 comparing loaner #13 and my lab unit

    Setup Conditions

    The data presented were collected as follows:

    1. PrismSound dScope III, powered on at least 90 minutes prior to measurement suite
    2. Balanced XLR cables Belden 1800F with Neutrik 110R AES connectors
    3. 100 Kohm load used for measurements
    4. dScope analyzer sample rate 48 KHz unless otherwise noted
    5. DAC 44.1 KHz sample rate, 24 bit depth unless otherwise noted
    6. Shielded 14AWG and 16AWG power cables
    7. Lin (red) filter mode unless otherwise noted
    8. AES input to Balanced output

    Measurements are made in accordance with AES17:2015

    A suite of 21 measurement scripts was performed at time intervals of
    T=4, 168 and 528 hrs for loaner #13, and T=4 and 528 hrs for the lab unit during the first warm-up data collection. Some distortions observed for #13 were unusually low at T=168 and 528 hrs. In particular THD+N, THD, D2 and IMD. This motivated a second warm-up series with more frequent data collections at T=0, 1, 2, 4, 8, 24 hrs and then roughly every 24 hours thereafter to 168 or more hours total.

    T=0 is time of dac2541 under test power on. The duration for each test suite is approximately 11 minutes. Therefore unstable, unusual and changing measurements are expected from beginning to end of the T=0 data collection. All data collections at 1 hour or more are stable.

    Data for the first warm-up series has already been presented in posts above. The second warm-up series data follows. Whatever conditions resulting in unusually low distortion observations in the first run was not repeatable after power had been cycled for the second warm-up data collection.

    Lab unit select distortion numbers in percent
    01 dac2541 lab unit warm-up percent.png

    Loaner #13 select distortion numbers in percent
    02 dac2541-13 warm-up percent.png

    Reporting distortion in percentages has been tradition for decades. However percentages may exaggerate apparent differences. For example .05% and .025% which appear to differ by 2x, actually vary by 6 dB but are at -66 and -72 dB respectively. Reporting in dB as shown below assist visualizing differences with better perspective.

    Lab unit select distortion numbers in dB
    03 dac2541 lab unit warm-up dB.png

    Loaner #13 select distortion numbers in dB
    04 dac2541-13 warm-up dB.png
     
    • Like / Agree Like / Agree x 1
    • List
    Last edited: Mar 31, 2021
  20. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    15,524
    Trophy Points:
    113
    Location:
    On planet
    The following plots highlight the unusually low but unrepeatable distortions measured for Loaner #13 between first (Red) and second (Blue) warm-up series with Lab unit second warm-up series for reference depicted in Green.

    There were only 3 data points for Loaner #13 first warm-up series as depicted by the Red circles in the following graphs.

    dac2541 warm-up series THD+N Left channel
    01 dac2541 warm-up THD+N L dB.png

    dac2541 warm-up series THD+N Right channel
    02 dac2541 warm-up THD+N R dB.png

    dac2541 warm-up series THD Left channel
    03 dac2541 warm-up THD L dB.png

    dac2541 warm-up series THD Right channel
    04 dac2541 warm-up THD R dB.png

    dac2541 warm-up series D2 Left channel
    05 dac2541 warm-up D2 L dB.png

    dac2541 warm-up series D2 Right channel
    06 dac2541 warm-up D2 R dB.png

    dac2541 warm-up series 19+20KHz IMD Left channel
    07 dac2541 warm-up 19+20KHz IMD L dB.png

    dac2541 warm-up series 19+20KHz IMD Right channel
    08 dac2541 warm-up 19+20KHz IMD R dB.png
     
    • Like / Agree Like / Agree x 1
    • List

Share This Page