CHANGSTAR: The Headphone Mutual Adulation Society (of a different flavor)

Please login or register.

Login with username, password and session length
Advanced search  

News:

Pages: [1] 2

Author Topic: EXCLUSIVE: OBJ2 vs Magni full spectrum distortion at various power levels.  (Read 2324 times)

marvey

  • Don of the Beats Mafia
  • Master
  • Pirate
  • *****
  • Karma: +285/-1
  • Online Online
  • Posts: 3863
  • aka Purrin

OK. I don't know if I did this test right, but here's what I did.
  • x2 100 ohm high current resistors paralleled to get 50 ohms.
  • CLIO box probes connected at the ends. The input Z of these probes is 64kohms
  • Instead of a 1kHz tone, which is kind of industry standard BS for measuring power / distortion, I used a two sine tones at 120Hz and 1kHz (still kind of BS, but more realistic). We all know reproducing bass is tough and requires a ton of power.
  • I also hooked up my Fluke DMM at either ends to get voltage readings at 0.040, 0.060, 0.080, and 0.120 volts
These results are very preliminary and someone correct me or do the math if they see something wrong.


Notes:
  • The full spectrum graphs were NOT normalized for 0db, so just relative readings. So noise floor is not optimal.

« Last Edit: January 08, 2013, 01:46:12 PM by purrin »
Logged

marvey

  • Don of the Beats Mafia
  • Master
  • Pirate
  • *****
  • Karma: +285/-1
  • Online Online
  • Posts: 3863
  • aka Purrin

Magni measurements. Higher distortion compared to Magni at lower power levels, but this is expected for a discrete design. Before anyone goes stupid over them, people should understand that audio transducers have a magnitude or two greater distortion; and measurements at the headphone will likely be dominated by such distortion. In other words, human beings don't listen to the electrical output from the amp. They listen to what comes out of the headphones, which typically measures 10-100 times worse than any amp.
« Last Edit: January 08, 2013, 01:32:42 PM by purrin »
Logged

marvey

  • Don of the Beats Mafia
  • Master
  • Pirate
  • *****
  • Karma: +285/-1
  • Online Online
  • Posts: 3863
  • aka Purrin

Objective 2 results

The super low distortion for the first two graphs is not unexpected. That's what opamps do. However, I was surprised at the rapid escalation in distortion when I turned up the volume knob. These are not high levels of power (I briefly plugged in the HE500 - it should noted Z is 33ohms - and the sound was kind of loud, but not deafening.) It's obvious the OBJ2 is clipping on the last two graphs.

Anyways, I did something retarded and left the OBJ2 at the 0.120v reading for a few minutes (forgot to turn off the tone generator), and then I smelled smoke...

So I'm scratching my head because I thought the OBJ2 would have done better based on existing measurements / data. Some thoughts:
  • The 120Hz tone threw a wrench in the works. Op-amps were never designed to drive low Z loads.
  • One of the JRC4556s (output op-amp) was defective.
  • There's much more voltage (hence power) that's being put out than I realize (some error in wiring, measuring, etc.)
« Last Edit: January 08, 2013, 02:27:12 PM by purrin »
Logged

marvey

  • Don of the Beats Mafia
  • Master
  • Pirate
  • *****
  • Karma: +285/-1
  • Online Online
  • Posts: 3863
  • aka Purrin

Exploding op-amps smell like sh1t. And that was 30 minutes ago. Fortunately, I've got a few extra JRC4556s handy.
« Last Edit: January 08, 2013, 01:33:59 PM by purrin »
Logged

Willakan

  • Pirate
  • **
  • Karma: +16/-16
  • Offline Offline
  • Posts: 109
  • I'm quite reasonable really.

At the risk of stating the blindingly obvious, there's definitely something wrong here :D. Opamps do not suddenly panic when faced with a 120hz and 1khz tone. Seriously, the 19khz and 20khz IMD test is absolutely brutal in comparison, in terms of the challenge it poses for the amplifier - if anything, it's an unrealistically worst-case scenario, as opposed to going to easy on the amplifier.

I'm also presuming you weren't making the O2 clip deliberately, which wouldn't make any sense, so that would imply...you get the point. Broken stuff is broken.
« Last Edit: January 08, 2013, 02:03:04 PM by Willakan »
Logged
Indecent lover of cheap opamps...

schiit

  • Schiit Audio
  • Pirate-at-Heart
  • Powder Monkey
  • *
  • Karma: +48/-0
  • Offline Offline
  • Posts: 73
  • Terrible gear for the least discerning listeners.
    • Schiit Audio

Interesting. Kinda like a IMD test, but with nonstandard tones. Not a bad idea.

But...was it normalized for 0db? The scale is weird.

I can post up what typically what we get out of a Magni at 600 and 32 ohms, 1VRMS out, 1K tone, if I can figure out how to use the image attachment deal--or I can send them to you separately.

Edit--sent to Purrin. I suck at the internets.

Of note, the 60Hz peak is a ground loop that comes and goes on one of the Stanfords--still trying to chase down, or get it back to them for service. At least it's the SR1, not the 1+ with all the crazy low-jitter, high-res options.
« Last Edit: January 08, 2013, 01:57:19 PM by schiit »
Logged
Lots o Schiit around here.

marvey

  • Don of the Beats Mafia
  • Master
  • Pirate
  • *****
  • Karma: +285/-1
  • Online Online
  • Posts: 3863
  • aka Purrin

The full spectrum graphs were NOT normalized for 0db, so just relative readings. So noise floor is not optimal.
Logged

marvey

  • Don of the Beats Mafia
  • Master
  • Pirate
  • *****
  • Karma: +285/-1
  • Online Online
  • Posts: 3863
  • aka Purrin

Two Magni measurements provided by Jason attached.

I'll repair and re-run the OBJ2 tests.
Logged

Willakan

  • Pirate
  • **
  • Karma: +16/-16
  • Offline Offline
  • Posts: 109
  • I'm quite reasonable really.

This presumably isn't the same O2 used for the listening tests? Also, any chance of a HF IMD test (CCIF)? Some very interesting-looking sidebands, and as mentioned earlier as far as IMD is concerned this isn't at all a punishing test.
« Last Edit: January 08, 2013, 02:26:50 PM by Willakan »
Logged
Indecent lover of cheap opamps...

marvey

  • Don of the Beats Mafia
  • Master
  • Pirate
  • *****
  • Karma: +285/-1
  • Online Online
  • Posts: 3863
  • aka Purrin

Different one. Anax brought over his O2 before he took off for CES. The one I used in the listening test was part of my test rig.
Logged

Willakan

  • Pirate
  • **
  • Karma: +16/-16
  • Offline Offline
  • Posts: 109
  • I'm quite reasonable really.

Good good. Not sure how the timing on my edits worked out (EDIT: badly), but how about the IMD tests? Please :D?

« Last Edit: January 08, 2013, 02:41:11 PM by Willakan »
Logged
Indecent lover of cheap opamps...

marvey

  • Don of the Beats Mafia
  • Master
  • Pirate
  • *****
  • Karma: +285/-1
  • Online Online
  • Posts: 3863
  • aka Purrin

LOL! Patience! I need to examine the amp carefully. Don't want something exploding again!
Logged
Pages: [1] 2