Goldpoint SA2X - Nobsound NS-05P passive attenuator comparison technical measurements

Discussion in 'Portable and Other Gear Measurements' started by atomicbob, Jan 10, 2019.

  1. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    Goldpoint SA2X - Nobsound NS-05P passive attenuator comparison 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 usually beyond the scope of technical measurements posts. This post will be somewhat of an exception.

    The data presented were collected as follows:

    1. PrismSound dScope III, picoscope 5243B
    2. Balanced cables Canare L-4E6S starquad and Belden 1800F with Neutrik XLR connectors
    3. Single Ended cables Audioblast HQ-1 1.5 ft with Rean RCA connectors or RG6 with BNC connectors
    4. 75R RCA to BNC adapters as necessary
    5. Vaunix Lab Brick USB hub for measurement equipment
    6. Audioquest Forest and Schiit Pyst USB cables used for measurement equipment
    7. 14 and 16 AWG sheilded power cables used

    All testing performed at +20 dBu input and 20 dB attenuation unless otherwise noted.


    Post table of contents
    1. Introduction and Background
    2. Select Measurement Comparisons
    3. Square Wave Response and Channel Imbalance
    4. Frequency Response vs Attenuation
    5. Frequency Response vs cable type and length


    Introduction
    The BIG QUESTION: will the NS-05P at $50 provide similar performance to an SA1X which is 10x $$$?
    The BIG ANSWER: it depends.

    Unsatisfying for TL;DR but unfortunately this is the correct answer. With optimal setup conditions both will perform similarly. But when the setup is more challenging, serious shortcomings of the NS-05P are revealed.

    SA2X and SA1X employ the same stepped attenuator. The SA2X has two inputs and an input selector switch. I'm using the SA2X for this comparison as it was within easy reach when I began this evaluation exercise. All performance assessments apply equally to the SA1X.


    Background
    Most amplifiers, both speaker and headphone have the following basic block diagram:
    00 20190110 Typical Amp block diagram.png
    Output level is controlled by a potentiometer attenuating input signals before sending to a fixed gain amplification block. In essence nearly all amplifiers have a passive attenuator.

    When a high output signal source is applied to an amplifier designed for a much lower input signal level, the amp volume control will often need to provide large attenuation with the volume knob at the minimum extremes. Many potentiometers will change levels greatly with little movement on the knob and have increased channel imbalance issues at this extreme operation. One solution involves attenuating the hot signal source with an external passive attenuator allowing the internal potentiometer to operate in a more favorable performance range.


    01 20190110 Typical Amp block diagram + ext passive attenuator.png
    An external passive attenuator (or passive preamp) can provide the desired level match with minimal signal degradation, but it is important to know limitations and constraints. One extremely important consideration is the distributed capacitance of the interconnecting cable between the output of the attenuator and input of the amp. The connection forms a simple RC low pass filter. Such a filter has potential to roll off high frequencies. Such constraints will be explored while comparing measured performance of two passive attenuators.

    Very important design differences between SA2X and NS-05P
    SA2X 10K ohm stepped attenuator, 4 gang switch with precision low noise resistors
    NS-05P 50K ohm potentiometer, ALPS 4 gang


    Picture of the Audio Lab setup for SA2X and NS-05P measurements and listening
    02 DSC_0017_F13_small.jpg
    RME ADI-2 Pro reference output set to +24dBu for listening evaluation
     
    Last edited: Jan 10, 2019
  2. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    Select Measurement Comparisons

    Frequency response with 20 dB attenuation from +20 dBu
    SA2X

    10 SA2X G2_L1T2_1_B.png

    NS-05P
    11 NS-05P G2_L1T2_1_B.png
    SA2X has negligable high frequency roll off while NS-05P drops approximately 0.65 dB at 20 KHz.


    Gain linearity with 20 dB attenuation from +20 dBu
    SA2X

    12 SA2X G2_L1T3_1_A.png

    NS-05P

    13 NS-05P G2_L1T3_1_A.png
    NS-05P gain linearity deteriorates approximately 6 to 7 dB higher than SA2X


    Crosstalk with 20 dB attenuation from +20 dBu
    SA2X

    14 SA2X G2_L1T11_1.png

    NS-05P
    15 NS-05P G2_L1T11_1.png
    Really no contest here. SA2X has far superior (lower) crosstalk isolation

    Residual noise with 20 dB attenuation from +20 dBu
    SA2X

    16 SA2X FFT_2_L1T16_1_A.png

    NS-05P
    17 NS-05P FFT_2_L1T16_1_A.png
    NS-05P is more susceptible to external mains hum and demonstrates higher residual noise
     
    Last edited: Jan 10, 2019
  3. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    Square Wave Response and Channel Imbalance

    NS-05P square wave 20 Hz 2 Vpp 10 mS / div 10K load 100 KHz measurement BW 20 dB attenuation

    20 20190109 SigGen NS-05P 20 Hz square 2000mVpp input 20 dB atten 10mS div 100KHz filter.png
    Same low frequency response as any other potentiometer or stepped attenuator.

    NS-05P 20 Hz square 2000mVpp input 20 dB atten 500nS div 20 MHz filter

    21 20190109 SigGen NS-05P 20 Hz square 2000mVpp input 20 dB atten 500nS div 20 MHz filtr-BW calc.png
    NS-05P square wave 20 Hz 2 Vpp 500 nS / div 10K load 5 MHz measurement BW 20 dB attenuation

    Bandwidth estimation: BW (MHz) = 0.35 / RT (mS)
    Where RT = 10 to 90% Rise Time
    0.35 / 1.986 uS = 0.176 MHz

    SA1X 20 Hz square 2000mVpp input 26 dB atten 200nS div 10MHz filter

    22 20181117 SigGen SA1X 20 Hz square 2000mVpp input 26 dB atten 200nS div 10MHz filter - BW calc.png
    SA1X square wave 20 Hz 2 Vpp 200 nS / div 10K load 5 MHz measurement BW 26 dB attenuation

    Bandwidth estimation: BW (MHz) = 0.35 / RT (mS)
    Where RT = 10 to 90% Rise Time
    0.35 / 0.239 nS = 1.464 MHz

    SA1X / SA2X has much better transient response with 1.464 MHz estimated bandwidth vs. NS-05P at 176 KHz estimated bandwidth


    Channel Imbalance
    SA2X A03 Gain vs Volume Control Setting

    23 20190109 SA2X A03 Gain vs Volume Control Setting.png

    NS-05P A03 Gain vs Volume Control Setting

    24 20190109 NS-05P A03 Gain vs Volume Control Setting.png
    Stepped attenuators may not achieve exact dB attenuation due to limitations of discrete resistor values while a potentiometer is nearly infinite in value setting. But the stepped attenuator can achieve precise, repeatable level settings which is more difficult with a potentiometer. Channel imbalance is also an issue as can be seen above. Left and Right channels overlay exactly on SA2X graph while the imbalance is apparent in the NS-05P graph.

    Left - Right channel tracking
    25 20190109 SA2X Gain vs Volume Knob Setting.png
    26 20190109 NS-05P Gain vs Volume Knob Setting.png
    Another way to look at channel imbalance. Consider the Left channel as the correct attenuation level and compare Right channel difference. In the SA2X graph L-R channel tracking is impeccable. NS-05P begins reasonably well with minimal attenuation but as attenuation is increased channel tracking deteriorates considerably.
     
    Last edited: Jan 11, 2019
  4. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    Frequency Response vs Attenuation

    42 dB Attenuation - Belden 1800F cable 2 ft
    30 20190109 SA2X - NS-05P A02 frequency response 42 dB attenuation Belden 1800F 2 ft.png
    SA2X (blue) NS-05P (red) performance is essentially identical out to 40 KHz


    36 dB Attenuation - Belden 1800F cable 2 ft
    31 20190109 SA2X - NS-05P A02 frequency response 36 dB attenuation Belden 1800F 2 ft.png
    SA2X (blue) NS-05P (red) performance is essentially identical out to 20 KHz


    24 dB Attenuation - Belden 1800F cable 2 ft
    32 20190109 SA2X - NS-05P A02 frequency response 24 dB attenuation Belden 1800F 2 ft.png
    SA2X (blue) remains unchanged while NS-05P (red) drops 0.25 dB at 20 KHz


    12 dB Attenuation - Belden 1800F cable 2 ft
    33 20190109 SA2X - NS-05P A02 frequency response 12 dB attenuation Belden 1800F 2 ft.png
    SA2X (blue) drops approximately 0.1 dB at 20 KHz while NS-05P (red) drops 2 dB at 20 KHz
    The 2 dB drop might be objectionable to many listeners.


    6 dB Attenuation - Belden 1800F cable 2 ft
    34 20190109 SA2X - NS-05P A02 frequency response 6 dB attenuation Belden 1800F 2 ft.png
    SA2X (blue) drops approximately 0.2 dB at 20 KHz while NS-05P (red) is estimated to drop 3 to 4 dB at 20 KHz. Again the high frequency attenuation might be objectionable to many listeners.

    With attenuations 20 dB or greater, the two devices perform similarly. But when attenuating between 1 and 20 dB SA2X offers far less high frequency loss than the NS-05P.
     
    Last edited: Jan 10, 2019
  5. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    Frequency Response vs cable type and length

    SA2X blue - NS-05P red in graphs below

    42 dB attenuation Belden 1800F 2 ft
    40 20190109 SA2X - NS-05P A02 frequency response 42 dB attenuation Belden 1800F 2 ft.png

    42 dB attenuation Canare L-4E6S 1_5 ft
    41 20190109 SA2X - NS-05P A02 frequency response 42 dB attenuation Canare L-4E6S 1_5 ft.png

    42 dB attenuation Canare L-4E6S 10 ft
    42 20190109 SA2X - NS-05P A02 frequency response 42 dB attenuation Canare L-4E6S 10 ft.png
    At 42 dB attenuation again there is little performance difference between the two passive attenuators


    6 dB attenuation Belden 1800F 2 ft - estimated cable capacitance 24 pF
    43 20190109 SA2X - NS-05P A02 frequency response 6 dB attenuation Belden 1800F 2 ft.png

    6 dB attenuation Canare L-4E6S 1_5 ft - estimated cable capacitance 69 pF
    44 20190109 SA2X - NS-05P A02 frequency response 6 dB attenuation Canare L-4E6S 1_5 ft.png

    6 dB attenuation Canare L-4E6S 10 ft - estimated cable capacitance 460 pF
    45 20190109 SA2X - NS-05P A02 frequency response 6 dB attenuation Canare L-4E6S 10 ft.png
    At 6 dB attenuation very large performance differences are observed. Cable capacitance has a dominating influence on high frequency roll off.

    Lesson: keep cables short between Passive Attenuator Output and Amplifier Input. Lower capacitance cables provide better performance.


    Conclusion
    When used with at least 20 dB attenuation and very short, low capacitance cables the Nobsound NS-05P will come close to the performance of Goldpoint SA1X. With attenuations between 1 and 20 dB high frequencies will be diminished as demonstrated above and deteriorate further with higher capacitance cables and longer cable runs.
     
    Last edited: Jan 10, 2019
  6. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    reserved 6/6.
     
  7. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,990
    Trophy Points:
    113
    Location:
    Irvine CA
    Last edited: Jan 10, 2019
  8. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    Each of the 4 gangs measured approximately 50K in the NS-05P.

    IMG_0343_small.jpg IMG_0346_small.jpg
     
  9. Taverius

    Taverius Smells like sausages

    Pyrate
    Joined:
    Dec 27, 2017
    Likes Received:
    3,032
    Trophy Points:
    113
    Location:
    Rapallo, Italy
    @atomicbob is there a difference in capacitance between coaxial cables and braided, generally?

    I'm going to use my goldpoint at the workshop later this year and I have both on hand, but not the equipment to measure them.

    Trying them both now I don't really hear a difference (they're both 25cm so not surprising) but I'm kinda curious.
     
  10. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,990
    Trophy Points:
    113
    Location:
    Irvine CA
    In a nutshell, it seems that @ 20 dB attenuation from 0 dBu:
    1. NS-05P drops 0.65 dB @ 20 kHz
    2. NS-05P linearity goes down by 3 dB @ -85 dB attenuation.
    3. NS-05P xtalk @ 1 kHz is about -75 dB
    4. NS-05P 60 Hz hum is -110 dBu
    5. NS-05P BW is 176 kHz
    6. NS-05P L and R tracking falls below 0.5 dB @ -50 dB attenuation.

    1. SA2X drops 0.01 dB @ 20 kHz
    2. SA2X linearity goes down by 3 de @ -95 dB attenuation.
    3. SA2X xtalk @ 1 kHz is below -120 dB
    4. SA2X 60 Hz hum is -120 dBu
    5. SA2X BW is 1.4 MHz
    6. SA2X L and R tracking is flawless @ -60 dB attenuation.

    So, SA2X is superior. It seems the extra cash does give you extra performance. But the above don't really scare me away from the NS-05P. Except your results with 6 dB attenuation.

    Why is the BW so dependent on attenuation level?

    Is it parasitics?
     
    Last edited: Jan 10, 2019
  11. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,990
    Trophy Points:
    113
    Location:
    Irvine CA
    BTW, this makes (IMO) a case of active pre-amps (Hot Pots/Current Drivers/No-Gain "Chu-Moys"). Because the pot is located before the active buffer (a short little trace ideally) which should be able to drive a longer cable for a lot less $.
     
  12. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    Good question. Output impedance from a passive attenuator varies with the parallel of the two resistances creating the divider.

    Typical balanced passive attenuator schematic:
    20190110 Typical Bal passive attenuator.png
    Highest output impedance will occur at attenuation of 6 dB presenting highest R to an RC lowpass filter.

    Simulating a 10K pot in a balanced configuration (20K total) set for 6 dB attenuation,
    100K load and 10 ft of Canare L-4E6S:
    20190111 SA2X 10K bal set 6dB atten to 100K load - Canare L-4E6S 10 ft.png
    Prediction is -1.15 dB loss at 20KHz, actual measured in posts above was -0.75 dB.

    Simulating a 50K pot in a balanced configuration (100K total) set for 6 dB attenuation,
    100K load and 10 ft of Canare L-4E6S:
    20190111 NS-05P 50K bal set 6dB atten to 100K load - Canare L-4E6S 10 ft.png
    Prediction is 8 dB loss at 20KHz, actual measured in posts above was approximately the same.

    Conclusion: 10K per gang potentiometers are a much better choice than 50K per gang for this application. Substituting a 4 gang 10K ALPS in the Nobsound NS-05P, which would probably make it an NS-01P, would reduce the high frequency losses to nearly that of Goldpoint SA1X. Channel imbalance issues would remain.

    I guess there is a Nobsound NS-01P and it's a tube preamp. So much for product numbering extrapolations on my part.
     
    Last edited: Jan 11, 2019
  13. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,990
    Trophy Points:
    113
    Location:
    Irvine CA
    Indeed a 10K ALPS makes more sense all around.

    Nobsound is all over the place with their specs on the NS-05P, and their nomenclature. As discussed before, some sites say the pot is 10K, some others 20K, and even the part family is all over the place.

    The nomenclature is confusing. I would have guessed that NS stands for Nobsound, 05 for model, and the P for passive. But this is broken with the NS-01P.

    If I go to Amazon here:
    https://www.amazon.com/Nobsound-Precision-Passive-Controller-Preamplifier/dp/B07GRLV7XG

    They offer SE models that change a little the presentation and maybe the ALPS family. But all are 50K which based on our discussion here is a bit high. Hard to identify by model.

    BTW guys. This may sound like I'm shilling. But it is really a point to all inclined to use a decent and affordable passive pre-amp:

    @schiit SYS is a 10K passive pot. So the high frequency roll over due to long cable RC low pass filter effect will not be a big deal as shown by @atomicbob. You also don't get a box of chocolates.

    What I got from this is that, if in search for other alternatives using the affordable passive pot approach, check the pot impedance and pot model. For very long lines, I might consider using an active balanced pre-amp. The output buffer lowers the output impedance to the sub-ohm range and if not unstable, might be able to drive long lines like no passive solution depending on my application.

    For lines below 10 ft, a passive 10K pot might do.

    Most pots do indeed have issues of channel imbalance as the attenuation goes lower than -50 to -60 dB. But it doesn't bother me because I don't usually attenuate signal that much.

    If in some applications the imbalance at -60 dB attenuation is a problem for some reason, maybe consider a stepper indeed (like the SA2X).

    Anyhow, thanks @atomicbob!
     
    Last edited: Jan 11, 2019
  14. frenchbat

    frenchbat Almost "Made"

    Joined:
    Sep 26, 2015
    Likes Received:
    1,370
    Trophy Points:
    93
    You do hint at an important point @ultrabike and it's been said by a few guys before but I think it's worth repeating : gain structure is very important, because anything too high or too low is gonna be a problem at some point. Ideally one would want the minimum attenuation possible.
     
  15. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    Thank-you @frenchbat !

    REPEATED - This should be at the top of the list before any twiddle / fiddle / diddle with tweaks such as USB cables, esoteric interconnects, power cables etc. Gain structure has first order effects while tweaks are in the tertiary or lower potential at best.
     
  16. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    There will be a more complete answer to this question in the form of a separate post on unbalanced, and maybe balanced cables when I complete that project. Here is a preview for some unbalanced interconnects as measured on a BK Precision 891 LCR meter:
    20181127_interconnect_Cp.png

    The answer is yes, cable capacitance does vary as would be expected given variance in construction details, insulation dialectrics, etc.
     
  17. frenchbat

    frenchbat Almost "Made"

    Joined:
    Sep 26, 2015
    Likes Received:
    1,370
    Trophy Points:
    93
    Looking forward to this. I found that blogpost on the subject a while ago : https://www.belden.com/blog/broadcast/understanding-skin-effect-and-frequency
     
  18. Armaegis

    Armaegis Friend

    Pyrate BWC
    Joined:
    Sep 27, 2015
    Likes Received:
    8,183
    Trophy Points:
    113
    Location:
    Winnipeg
    Interesting that the starquad drops so much... and the pyst increases?
     
  19. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet
    Keep in mind those sweeps go to 300 KHz, so most of the changes occur above 20 KHz. Canare L-4E6S is star quad with four conductors and a shield. In an unbalanced interconnect configuration, two conductors and the shield are connected to RCA shield while two conductors connect to center pin. This may not be the best use case for star quad cables, which are designed for balanced use. This makes for a great example of how blindly or randomly picking components, cables etc. based on excellent performance in one use model doesn't guarantee system synergy in another use model.
     
  20. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    19,646
    Trophy Points:
    113
    Location:
    On planet

Share This Page