O2 Measurements & Review

Discussion in 'Headphone Amplifier Measurements' started by tomchr, Jan 26, 2017.

  1. Poleepkwa

    Poleepkwa Friend

    Friend BWC
    Joined:
    Oct 1, 2015
    Likes Received:
    602
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Finland
    I find the Butte much more engaging than the O2, although I am sure the measurements will favour the O2.
    The DIY instructions are excellent and Pete has been very friendly and helpful.
     
    Last edited: Dec 2, 2018
    Cryptowolf, rtaylor76 and ultrabike like this.
  2. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    7,247
    Dislikes Received:
    15
    Trophy Points:
    113
    Location:
    Irvine CA
    OK guys, here are some THD sweeps with low gain (2.5X) using different loads (I skipped the 6.5x gain because I don't really use it, and it can indeed clip. Use that one only if your DAC is weak sauce):

    Note: All of these measurements were done with a 14VAC/500mA wall-wart. I don't use this amp from batteries. It really works better as a small desktop amp.

    300 ohms

    THD vs Vin
    Sweep_THDvsVin_lowgain_300ohm.JPG

    Vout vs Vin
    Sweep_VoutvsVin_lowgain_300ohm.JPG

    With a resistive 300 ohm load we easily get 0.0005% THD or so over most of the operating range. The output voltage can reach about 7.5 Vrms w/o issues and 2.5x amplification is achieved. Maximum input voltage range seems to be around 3 Vrms or so. Not 8 Vrms, but should handle most DACs I know IMO. I'm calling this 7.5^2/300 = 187 mW max output power.

    100 ohms

    THD vs Vin
    Sweep_THDvsVin_lowgain_100ohm.JPG
    Vout vs Vin
    Sweep_VoutvsVin_lowgain_100ohm.JPG

    Things are starting to strain a bit. Now things are more close to 0.001% and above 1 Vrms input (2.5 Vrms output) things start to degrade harder. That said 0.001% is not bad at all. I'm going to call this 6.8^2/100 = 462 mW max output power.

    33 ohms

    THD vs Vin
    Sweep_THDvsVin_lowgain_33ohm.JPG

    Vout vs Vin
    Sweep_VoutvsVin_lowgain_33ohm.JPG

    A more difficult load. The system is now proly closer to 0.05% THD or so and we no longer touch 6.6 Vrms output. A 5 Vrms output is as far as we go before shit hits the fan. Therefore I'm calling this 5^2/33 = 757 mW max output power.

    HD600 load

    THD vs Vin
    Sweep_THDvsVin_lowgain_HD600.JPG

    Vout vs Vin
    Sweep_VoutvsVin_lowgain_HD600.JPG

    Now we are talking business. These sweeps are with the HD600 acting as load. Not as clean as 300 ohms, but not bad IMO. This is a bit of a less linear load perhaps. But still things look decent. Max output power seems to be about 187 mW still.

    Making measurements w the HD600 as load is annoying as hell though. The HD6x0 cans are very sensitive, and 187 mW is a lot of power into them. We can hear the 1 kHz tone loud and clear through the house.

    This sweeps are logarithmic and distortion is shown in % format so it can be compared to @tomchr own measurements.
     
    Last edited: Dec 3, 2018
    Lyer25, Cryptowolf, purr1n and 3 others like this.
  3. tomchr

    tomchr MOT - Neurochrome

    Joined:
    Oct 28, 2016
    Likes Received:
    427
    Dislikes Received:
    19
    Trophy Points:
    63
    Location:
    Calgary, Canada
    Home Page:
    Looks similar to my measurements. Maybe slightly higher distortion. Do you know what the measurement bandwidth of the SRS is? I use 20 kHz (AES 17) for 1 kHz measurements.

    Heh... Yeah... When I developed my DG300B, I performed many sine sweeps both vs output power and vs frequency. The vs frequency measurements were fun. The 300B is a quite microphonic tube and the output transformers will give off some hum at the signal frequency as well, so when testing at 10 W output power, the amp is very audible - even with resistive load. Every time the sine sweep hit 8-10 kHz my neighbour's dog would go bonkers. It was pretty funny. The dog was a very annoying and aggressive pit bull so I didn't really mind getting it in trouble. :)

    Tom
     
    Hrodulf and ultrabike like this.
  4. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    7,247
    Dislikes Received:
    15
    Trophy Points:
    113
    Location:
    Irvine CA
    Yup. I think I used a BW of 32 kHz instead of AES17. I'm still working on a better presentation for the THD sweeps and might redo with the AES17 filter as default. If not using it for some reason I'll try to remember to mention it.
     
  5. tomchr

    tomchr MOT - Neurochrome

    Joined:
    Oct 28, 2016
    Likes Received:
    427
    Dislikes Received:
    19
    Trophy Points:
    63
    Location:
    Calgary, Canada
    Home Page:
    I've made it a habit to list the test conditions in the chart title. That way there's no ambiguity or assumed defaults.

    Tom
     
  6. Roen

    Roen Rando

    Joined:
    Mar 3, 2017
    Likes Received:
    7
    Dislikes Received:
    4
    Trophy Points:
    8
    Location:
    NYC / TO
    With a 16V peak, 11.313708 is Vrms. It's just coincidence that 0.7A is short of close to 1/SQRT(2).

    Pretty sure the 11.2VA figure is another way of saying 11.2W.

    Anyway, those with 2.0Vrms DACs should be getting the unity / 3.5x gain versions, and those with 2.1Vrms DACs should be getting the unity / 3.3x gain versions.
     
  7. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    39,703
    Dislikes Received:
    62
    Trophy Points:
    113
    Location:
    Antarctica
    Just get JDS cMoyBB (lower power / efficient high impedance) or Magni 3 (slightly more power / will work with orthos too). Both sound better than O2.

    Swapping opamps on O2 gets people into trouble. Stock voltage gain opamp makes the most of the power rails, but sounds like ass. Other opamps don't swing nearly as close to the rails, so clipping happens earlier than expected. The 4556 output opamps were chosen for current capability, but they are colored, e.g. the Grado headamp used a single one of them (effectively CMOY), and they matched well with Grado headphones.
     
  8. Roen

    Roen Rando

    Joined:
    Mar 3, 2017
    Likes Received:
    7
    Dislikes Received:
    4
    Trophy Points:
    8
    Location:
    NYC / TO
    I went with the O2 because I drive both IEMs and cans, and I like to be able to use it portably when necessary. (i.e. on a train, bus)

    Would love to hear about other desktop amps that can be used in portable mode, but get a nice power boost with an AC adapter.

    For full desktop use, I use my 789. I used to use the GO2Pro, but the D50 + O2 seems to be doing a better job.
     
  9. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    39,703
    Dislikes Received:
    62
    Trophy Points:
    113
    Location:
    Antarctica
    I call bullshit on this unless you post a photo of yourself on a NYC or Toronto subway with headphones and O2 and source. Yes, I expect to see a photo of a source (phone, portable player), an O2, and a headphone in your bag.
     
  10. Roen

    Roen Rando

    Joined:
    Mar 3, 2017
    Likes Received:
    7
    Dislikes Received:
    4
    Trophy Points:
    8
    Location:
    NYC / TO
    Lol, sure.

    I'll take a photo of my iPad, feeding my D50 sitting on top of a USB battery pack, which is on top of the O2 feeding my cans.

    "Portable"

    But my use case is more generally commuter rail (LIRR, GO Transit), where I can take up an extra seat with my quad stack of devices. I may consider switching out the D50 + Battery Pack for a Tone Board if the enclosure is robust enough for portable transport.
     
    Last edited: Dec 18, 2018
  11. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    7,247
    Dislikes Received:
    15
    Trophy Points:
    113
    Location:
    Irvine CA
    It was confusing to me.

    Consider this US wall-wart:

    wall-wart.jpg

    16VAC 500mA = 8.0W.

    So VAC seems to mean Vrms (Volts AC). Take the C out (VA) and it seems we are talking W (W = VI). Oh well. That leaves about 16 VAC 700mA = 11.2 W or VA. Clipping issues can happen at 6.5x. But unity gain and 2.5x not so much. The build can be defective though.

    Also, the standard O2 version has no unity gain. It's 2.5x Low and 6.5x High. I use 2.5x pretty much all the time.

    I use my iPhone + no-amp.

    My no-amp is very transparent and super-portable.
     
    Last edited: Dec 18, 2018
    Erikdayo likes this.
  12. Roen

    Roen Rando

    Joined:
    Mar 3, 2017
    Likes Received:
    7
    Dislikes Received:
    4
    Trophy Points:
    8
    Location:
    NYC / TO
    I'm sure it is, we all have different standards of what we consider portable. But yeah, certain stacks are portable for me as long as they're battery powered.

    The day I carry a SMSL D1, THX 789, 12V DC to AC Inverter and a car battery on the NYC subway will be the day.
     
    Last edited: Dec 18, 2018
    ultrabike likes this.
  13. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    39,703
    Dislikes Received:
    62
    Trophy Points:
    113
    Location:
    Antarctica
    Most of SBAF has moved away from carrying stacks of shit rubber banded together. That was so 2012. Either self-contained portable players, i.e. Sony, Questyle, Onyko, etc. or phone by itself + IEMs for truly portable use that will work on crowded NYC subways, streets, airports. Yes, Manhattan is like a second home to me now.

    Fulla 2 stashed in my bag for headphones + computer when I arrive at the hotel.
     
    Dr. Higgs and Lyer25 like this.
  14. Roen

    Roen Rando

    Joined:
    Mar 3, 2017
    Likes Received:
    7
    Dislikes Received:
    4
    Trophy Points:
    8
    Location:
    NYC / TO
    If I'm standing, I'm probably going to use a dongle, but my use case generally has me in a static position for an extended period of time, so I don't mind the stack.

    Would be funny running around with a stack though, but I'm sure at that point, all the external noise is going to affect your music anyway.
     
  15. Roen

    Roen Rando

    Joined:
    Mar 3, 2017
    Likes Received:
    7
    Dislikes Received:
    4
    Trophy Points:
    8
    Location:
    NYC / TO
    What is the negatve of matching opamp to DAC Vrms output? e.g. my DAC is 2 Vrms so I went with unity / 3.5x
     
  16. tomchr

    tomchr MOT - Neurochrome

    Joined:
    Oct 28, 2016
    Likes Received:
    427
    Dislikes Received:
    19
    Trophy Points:
    63
    Location:
    Calgary, Canada
    Home Page:
    If the voltage and current are in phase, then yes. With a reactive load, the current and voltage are out of phase. In that case V*A != W.

    Tom
     
  17. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    7,247
    Dislikes Received:
    15
    Trophy Points:
    113
    Location:
    Irvine CA
    LOL!

    That said, the voltage/current inputs (for AC power suply) are for the most part a sinusoidal (ignoring noise and non-idealities). Therefore, in steady state, one might say that the W = V*A in an RMS magnitude sense.

    In general however this indeed does not hold. If the inputs are not sinusoidal (music), then one may need to resort to solve the passive network with differential equations or better yet, a Laplace Transform to make things sort of algebraic like. One will need initial conditions. If one cannot easily represent the input signal with a closed form function of the nice type, forget it. It's simulation time.

    At least that's what me thinks right now. And I just came back from ELP. It was raining and we saw 3 nasty accidents on the 91 as we neared Irvine. It sucked. But we are here. Tired. Good night guys.
     
  18. tomchr

    tomchr MOT - Neurochrome

    Joined:
    Oct 28, 2016
    Likes Received:
    427
    Dislikes Received:
    19
    Trophy Points:
    63
    Location:
    Calgary, Canada
    Home Page:
    The mains voltage is a (distorted) sine wave. The current into the mains transformer is anything but sinusoidal for an audio amp. The rectifier diodes will conduct when the secondary voltage rises above the voltage on the reservoir caps. Hence, the current draw will be a pulse train. The power factor isn't remotely close to 1.0, thus, V*A != W.

    I wholeheartedly agree that working out the exact power factor for an audio amp amplifying a music signal will severely stress your circuit simulator. :)

    Tom
     
  19. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    7,247
    Dislikes Received:
    15
    Trophy Points:
    113
    Location:
    Irvine CA
    If I put a 32 ohm resistor across a 16VAC brick wall adapter output, would I get a distorted sinusoidal current or a pulse train across the resistor load?

    I do get your point though. The amplifier itself will get a sinusoidal voltage, but the amplifier will not draw a sinusoidal current.

    But I think folks use a resistive load to spec their stuff, which IMO is reasonable. Any other random device will not necessarily be representative of any other random device anyway. Furthermore, I think for the most part, as long as the amp load results in a power draw below the specified limits, things should be fine (defects aside).
     
    Last edited: Jan 12, 2019
  20. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    7,247
    Dislikes Received:
    15
    Trophy Points:
    113
    Location:
    Irvine CA
    BTW, There is no need to get lost in amplifier current draw details.

    On more important stuff, I'm a bit behind in making further measurements with the SRS. I will try to get back at it as soon as I can.
     

Share This Page