Header

Objective Reviews & Commentary - An Engineer's Perspective

August 24, 2011

O2 Summary

o2 v11 boardTHE SHORT VERSION: By popular demand, here’s an “executive summary” and Q&A for the Objective2 (O2) headphone amp. There are links sprinkled throughout leading to revised information in the previous three large articles and elsewhere.

QUESTIONS ANSWERED: Before you post a comment to one of the articles, forum threads, etc., please see if your question has already been answered below.

WHAT’S AN O2? The Objective2 (O2) is my attempt at a “One-Size-Fits-Nearly-All” headphone amp. It can drive most any headphone from the most sensitive IEMs to some really power hungry full size cans from 16 to 600 ohms. It’s small, inexpensive, and runs on batteries or AC power. The design is freely available to everyone subject to the License terms. See: O2 Headphone Amp

WORRIED YOU’RE NOT GETTING THE BEST SOUND? Reading through the forums it’s obvious a huge number of headphone lovers are worried they’re not getting the best sound from their headphones. Their worry may be justified as most of the headphone amps (and headphone DACs) on the market have one or more significant limitations or problems. Many only work well with a limited range of headphones. See: Going Shopping

o2 see thru

LESS WORRY: Very few manufactures provide anything resembling credible specifications for their headphone amps and very few are independently tested on proper equipment. So you rarely know what you’re getting. I’ve discovered significant problems or limitations with everything I’ve tested with a headphone jack on it. The O2 is designed to ease such concerns by working well with nearly all headphones and having fully documented performance. See: O2 Headphone Amp

nwavguy-o2-purple-largeWHY DID YOU CALL IT THE OBJECTIVE2? I’m an electrical engineer, serious audio geek, and took a very objective approach to the design. It’s a minimalist amp focused on the best performance possible for the least amount of money. See: Objective Pitch

POWER TO SPARE: In the testing I’ve done, with more than a dozen different headphones, either my ears, or the headphones, reach their limits long before the O2 does. I suspect with some worst case headphones (like some of the planars from HIFiMan) the O2 will merely get genuinely loud at full output. But for more typical headphones, the O2 will be well under its full output. Compared to most battery powered amps that are often pushed to the edge of, or well into, clipping that’s a very audible difference. The O2 is loafing when most portable amps are straining. And, unlike dual battery Cmoy amps, the O2 not only has lots of voltage, but it also has lots of current for low impedance loads that require it—like AKG K701s, Audeze LCD-2s and HiFiMan planars. And the O2’s two stage design vastly improves noise and distortion over a single stage Cmoy.

HEADPHONE SAFETY: Unlike every two battery Cmoy or similar design I know of, including the $650 Emmeline SR-71 Blackbird, the O2 won’t try to destroy your headphones if you run the batteries low or one battery becomes disconnected. It has a power management circuit to shut it down if there’s anything wrong with the power preventing dangerous levels of DC at the output.

o2 final pcb largeLOW OUTPUT IMPEDANCE = TIGHT BASS: If you’ve been using a headphone amp or source with a significant output impedance you may be in for a treat with the O2. A sufficiently low output impedance provides improved bass damping with many headphones and the most accurate frequency response instead of an exaggerated peak at resonance. With my Beyer DT770-Pro 80 headphones, for example, the difference in the bass performance is easily noticeable (and survives a blind test) when using a high output impedance source versus the O2. With the O2 the bass is tighter and more controlled. Lots of headphone source gear has a high enough high output impedance to cause audible problem. The O2 solves these problems. See: Amplifier Impedance

WHAT DOES IT COST? The O2 can be built in functional “bare board” form for $30 - $40. The total parts cost with everything, including a pre-made front panel, batteries, enclosure, and AC adapter is around $75 plus shipping. The DIYer can have a complete amp for $80 – $100 all inclusive. Even fully assembled complete with batteries and AC power adapter it can still be under $150. See: O2 Price Details.

o2 hd650 dscopeWILL THE O2 DRIVE MY HEADPHONES? Short of a few discontinued models, like AKG K1000 “earspeakers”, the answer is very likely yes. The O2 can drive normally challenging headphones including the AKG K701 and even HiFiMan planars. See: More Power?

DOES THE O2 HAVE ANY HISS WITH SENSITIVE IEMs? The O2 is the first amp I’ve tested that’s dead silent with any headphones I know of at any volume setting, and any applicable gain, including with my ultra sensitive UE SuperFi Pros. The noise is –116 dBv or –133 dB referenced to full output.

WILL YOU BUILD ONE FOR ME? There are already individuals and businesses offering to build the O2 for others either as a completed circuit board or a fully assembled amp. See O2 Resources for various options. I’m staying out of the commercial aspect. I expect as the O2 becomes more established there will be additional purchasing options in the future and I’ll be updating the O2 Resources links.

nwavguy o2 pcb enclosuresHOW CAN I GET ONE? Right now the O2 is mainly a DIY project and there will likely be more options in the near future. For what’s available today, see O2 Resources.

WHERE ARE THE LATEST UPDATES? Check out: O2 Important Information.

WHY DID YOU DESIGN IT? I shopped around for a decent headphone amp that didn’t have significant problems or limitations and couldn’t find any that were reasonably priced. Sadly, I mostly found heavily flawed designs unable to deliver what the recording engineer intended. I also wanted to show what can be done by properly implementing about $30 worth of parts. See: O2 Motivation

DO YOU HAVE A COMMERCIAL INTEREST IN THE O2? No. I don’t get any money, revenue or income of any kind from the O2. It’s open source hardware and I’m not selling anything. I don’t have a commercial interest in anything audio related at the moment. My day job is in another field.

WHY ARE YOU GIVING IT AWAY? I’m mainly trying to prove proper design goes a lot further than using expensive designer parts, following audiophile myths, trying to re-invent the wheel, resurrecting ancient single-ended designs, etc. The idea is to raise the bar in the industry as a whole. There are just too many marginal and/or overpriced headphone amps—both DIY and commercial. See: O2 Motivation

O2 V11 Battery Both Ch 1 Khz 10mV  THD N vs Output Left to Right 15 33 80 150 600 Ohms commentsWHERE ARE THE MEASUREMENTS? The O2’s performance has been thoroughly documented using industry standard measurement practices. For a summary see: O2 Review. For the full measurements see: O2 Measurements 

WHAT ABOUT THE SOUND QUALITY? The O2 compares very favorably in listening tests against the critically acclaimed Benchmark DAC1 Pre’s headphone output. An O2 user has compared it to a $1000+ AMB beta22 (b22) and wasn’t sure he could hear any difference. As more amps get evaluated more and more favorable reviews are coming in. I have encouraged blind listening tests against other gear as well. See: O2 Subjective Pitch

COMPETITIVE CHALLENGES: Is the Camaro faster than the Mustang? Let’s take them both to the track and find out. First across the finish line wins! I’ve challenged others to beat the O2’s objective measured performance with amps up to three times the assembled price. And for the subjectivists, I’m willing to put the O2 up against suitable amps of any price in a proper blind listening test. See An Open Challenge for the details.

o2 ipod dscopeDOES THE O2 HAVE ADJUSTABLE GAIN? The O2 has a gain switch on the front panel. You might use one setting with IEMs and the other with power hungry full size cans. Or one gain for portable use and the other for home use. The standard gains are 2.5X and 6.5X but can be internally changed to anything from 1X to 12X. See: Gain Settings and All About Gain.

HOW LONG IS THE BATTERY LIFE? The regular version runs for about 8 hours and, by swapping out 3 ICs, it increases to about 25 hours. See: O2 Low Power Option

ARE THERE FORUM DISCUSSIONS? Yes, you can find the links in the O2 Resources section.

WHERE’S THE SCHEMATIC AND PARTS LIST? See: O2 Resources

WHERE CAN I GET THE PARTS? See: Obtaining The Components

IS THERE CONSTRUCTION INFORMATION? Anyone building the O2 should read the O2 Cautions and O2 Circuit Board Construction sections. I would also suggest: Using The O2

IS THE O2 A GOOD FIRST DIY PROJECT? Novice DIYers might want to wait until at least several others have built the O2 before deciding if they want to try building one.

o2 v11 dscope cablesHOW DID YOU COME UP WITH THE DESIGN? The high level objectives can be found in the Design Highlights section and the O2 design itself is explained in the Circuit Description section. The entire design process including the Requirements, Circuit Design, and Prototyping/Testing is explained in the O2 Design Process article.

CAN I UPGRADE THE OP AMPS? You could, but it would be a waste of money and might degrade the performance. I tested nearly two dozen op amps and the O2 is optimized for the op amps that proved best. See: Op Amp Myths and Op Amp Measurements. I have offered $500 for charity to anyone who can hear differences between op amps that measure reasonably well. Also check out: Designer Components.

CAN I USE A DIFFERENT ENCLOSURE? Yes. Several have already made some amazing enclosures including an elaborate one out of laminated wood. See: Enclosure Options

WHAT ABOUT CLAIMS ON HEAD-FI THE O2 IS FLAWED? As far as I know, the criticism was started by those with financial interests in more expensive products. And it’s mostly been put to rest. Please consider where the O2’s critics obtain their income (selling parts, amps, ads to sponsors, etc.) and/or their personal investment in far more expensive gear. I’ve requested the critics participate in either of the two official O2 forum threads on ABI and diyAudio but it’s telling they have refused. Instead, they seem to be hiding behind the admins who banned me at Head-Fi. I welcome an open discussion on either of the official O2 threads, valid criticisms, proper blind comparisons, and professional measurements. See Competitive Challenges above. The O2 isn’t perfect, but when you consider its size, price, battery operation, etc. I believe it’s hard to beat and any criticisms should be kept in proper perspective for a portable amp that costs $25 – $150 and can drive most any headphone.

o2 8-5V battery input stage clippingDOES THE INPUT OVERLOAD? The O2 will not overload in any sort of normal use. Here are the facts (all input voltages are Vrms):

  • Understanding Gain – Please see All About Gain.
  • On AC Power - At 2.5X gain the O2 can handle 2.8V and at 6.5X gain it can handle over 1V.
  • On Battery Power – At 2.5X gain the O2 can handle 1.8V and at 6.5X it can handle 0.7V (most LODs are 0.5V).
  • The Competition – The very popular battery powered FiiO E7 overloads at 1.2V vs 1.8V for the O2. The equally popular AC powered FiiO E9 desktop amp overloads at 2.1V vs 2.8V for the O2. The Mini3 has about 6dB less overall headroom than the O2 with both running on battery power. The Mini3 also lacks a gain switch making it less flexible.
  • The Math – The maximum gain on AC is 7 / Vin and on battery it’s 4.5 / Vin. So with a 0.5V iPOD LOD cable it’s 7/0.5 = 14X on AC and 4.5/0.5 = 9X on battery.
  • Home Sources – The Redbook standard for home digital audio gear is 2V. Anything without a volume control above 2.8V would be extremely rare. But if you have such a device, just clip two resistor leads and the O2 will drop to 1X for one of the two gain settings. Or you can optimize the gain. See: All About Gain
  • Portable Sources – Nearly all battery powered players have a maximum output of 0.5 – 1.0V and nearly all USB powered portable DACs are 1.4V or less. All of these are 100% compatible with the O2 without changing anything.
  • Excess Gain – Some have claimed the O2 may not have enough “excess gain”—that is extra volume control range to “boost” improperly recorded/ripped music. Apple designed only a few dB of excess gain into their iPods because more would have been a serious compromise for 95% of the music people listen to. Too much excess gain is a problem because it renders a large portion of the volume control’s range unusable with most music. There are other negative side effects as well. Plus anyone using a PC or laptop as their source can add extra gain on the PC side to compensate for oddball recordings. They can also normalize any “quiet” tracks. With any headphones that work with the Mini3, for example, the O2 can provide at least 6 dB of excess gain even when operating on battery power. 6 dB of excess gain is plenty for more applications. With more sensitive headphones it can provide even more. See: All About Gain
  • More Details – For anyone still concerned there are more details, including scope shots of the O2’s actual performance, under Maximum Input and Gain Stage Limitations. See also: All About Gain

o2 portable desktopWHAT ABOUT A DESKTOP ONLY VERSION? It’s possible to use the O2 board in a variety of ways in a desktop amp. This is discussed in the Enclosure Options and Circuit Board Construction sections. I’m also planning a different amp, based on the O2 design principals, for desktop-only use with the following possible benefits:

  • Easier to construct as a desktop amp with everything self contained on the PC board rather than having to panel mount jacks, the volume control, etc. and use point-to-point wiring.
  • Built-in 1/4" headphone jack (in addition to a 3.5mm jack)
  • RCA input jacks (in addition to a 3.5mm jack)
  • Most of the connections on the back for more tidy desktop use
  • Higher input voltage range to handle even “very hot” sources at higher gains
  • A more flexible gain structure allowing better gain matching
  • A headphone protection relay preventing turn/off transients and offering DC protection
  • Perhaps a few other enhancements as well such as an improved power supply

WILL THE DESKTOP AMP COST MORE? The desktop amp will require two machined panels instead of one adding about $20 and the other upgrades will add another $10 – $20 but you also save $11 as you don’t need the batteries. So if you buy both panels it might be around $30 more.

o2 v1 board and dscope

WILL THE DESKTOP AMP SOUND BETTER? It’s hard to say. The desktop amp will have even more gain flexibility than the O2 and there can be benefits to using a 1/4” headphone jack and RCA input jacks. The power supply may also have lower noise.

WHEN WILL THE DESKTOP BOARD BE AVAILABLE? Hopefully in the first quarter of 2012.

WHICH ONE SHOULD I CHOOSE? There’s a lot to be said for a portable amp. I’m using my O2 on battery power far more than I would have guessed. Do you want a BMW coupe or convertible? Both are great cars with their own benefits and disadvantages. There will be more details on the desktop version in the next several weeks.

O2 LICENSE: See: O2 Creative Commons CC-BY-ND License

WHERE ARE THE OTHER ARTICLES? There are three main O2 articles plus the op amp measurements:

  • O2 Headphone Amp – The first article covering the design premise, review and measurements.
  • O2 Design Process – The second article covering the requirements and design process.
  • O2 Details – The third article with mostly practical information about buying/building an O2 as well as a detailed circuit description.
  • Op Amp Measurements – Many different op amps tested in the O2.
nwavguy headphone manBOTTOM LINE: I’ve tried to raise the price/performance bar for headphone amps with the O2. I’ve provided detailed measurements and the theory behind it. I hope the O2 encourages others to make, and publish, more accurate and complete measurements. I also hope it helps people smile more and worry less when listening to their music!

157 comments:

  1. Great summary post for the O2 amp, It'll be easier and cause less "I didn't read that, its too technical / long" responses when I send people to your blog when they ask why I'm so interested in building the O2 amp instead of buying something that's already out there.

    BTW you have the question "WHAT’S AN O2?" twice.

    ReplyDelete
  2. Thanks Sebastien. I'm getting better at this blogging stuff... slowly... I fixed the duplicate paragraph and appreciate the BTW.

    ReplyDelete
  3. This is a great Summary! Now people can get as much (or as little) technical information as they want by following the links.

    Just so you know, I think one of your tags is broken. The link to "list of how most of them fail" and links under that all go to the "O2 resources" link. Tested with FF6, Win 7.

    -odigg

    ReplyDelete
  4. Thanks Odigg! I have no idea how that happened, but I think they're all fixed. As usual, you all are better proofreaders than I am.

    ReplyDelete
  5. Is MrSlim a good person to buy the O2 from?

    ReplyDelete
  6. I have no personal experience with MrSlim but he's been doing this sort of thing for a while and I would think if he did it poorly he would have had to go into hiding by now ;) I can't "endorse" him at this point but once he builds a few O2's, and all is well, I'll be happy to. I have no reason to believe he's not 100% legit.

    ReplyDelete
  7. This might be a stupid Question to ask... but is this the finale design, or is there a possibility that you will change/improve it in the next few months?
    I'm in no hurry building the O2, so I dont want to "miss" something due to impatience.
    Btw, keep up the good work, you're doing a great job!

    ReplyDelete
  8. That's a good question. Unless something major is discovered I don't plan to change the O2 design once the group buy boards are ordered. I am encouraging everyone not to order parts until Oliver puts the board order in just in case something gets tweaked in the next week. The boards will have the longer lead time so there's no reason to buy parts yet.

    I'll hopefully be kicking the O2 out of the nest and letting it fend mostly for itself. I want to move onto testing DACs, working on the upgraded desktop amp, and other assorted blog priorities.

    ReplyDelete
  9. Thanks for the nice summary page and for your time and effort on this amp! I can't wait to start building it. Just a suggestion - perhaps you could also provide links to the relevant forum discussions on ABI and diyAudio on this page?
    Also, just wondering what changes you were thinking of implementing in the desktop version?

    ReplyDelete
  10. Please see above under "What about a desktop version" for what I have planned so far for the desktop amp.

    ReplyDelete
  11. Suggestion for desktop: 4 mounting holes

    ReplyDelete
  12. Good design. I hope you have kept your identity secret coz admins from head-fi might have hired Ninja assassins to dissect you & O2 by now ;-)

    ReplyDelete
  13. I'm not certain wether this is the best place for such a question, but I can't think of anywhere significantly more appropriate.

    Many years ago Jan Meier designed a portable amplifier targeted at DIYers that that employed a (to my knowledge) rather novel approach for attenuating output power.
    Instead of using a potentiometer that adjusts the voltage entering the amplifier, he used a potentiometer to adjust the gain factor in a continuous fashion.

    Schematic and some limited TechSpeak here:
    http://www.meier-audio.homepage.t-online.de/

    NwAvGuy, what do you think of the general premise of adjusting output power in this way?

    ReplyDelete
  14. Aksel, Meier has done some interesting things. My personal opinion is he keeps trying to re-invent the wheel when there's no need to. His "virtual ground" arguments, for example, make no sense to me. See my Virtual ground article for lots more about that topic.

    If you're talking about his "portable amp" with the LM6171 output stage. His use of the volume control in the feedback loop is very noise prone. When the pot becomes noisy (dirt/wear on the track/wiper) that "noise" in the pot will cause the amp to briefly go to full volume. The effect will be to make the pot wiper noise massively loud.

    It's also Audio Design 101 to never run DC through a volume control. To do so just about guarantees you'll get noise when adjusting the volume. But that's what Meier has done.

    He's used really high value resistors in that circuit so the Johnson Noise is probably off the chart. The pot in the feedback loop is likely a band-aid to try and reduce all the Johnson Noise.

    And if you're talking about the Corda 2Stepdance. Near as I can tell it uses a stepped attenuator with CMOS switches to control the volume. I can almost promise that solution is much less linear than a simple volume control. Doug Self has written about all the issues with using CMOS switches for audio. That amp almost certainly has noise issues looking at the PC board. The DC-DC converters and digital circuitry are just too close to the audio circuitry. It also has losses in the DC-DC converter which shortens battery life. I do like the option of two different rail voltages.

    ReplyDelete
  15. Thanks.
    It was the portable amplifier DIY project I was thinking of.
    Right, never DC through a volume control. I think I can remember that.
    I've read pretty much all you've written, but that stepdance amplifier was new to me. What a packed circuit board, no wonder why Head-Fiers think it sounds so good :)

    ReplyDelete
  16. I'm in on the PCB group buy with a friend (who's fortunately a whiz with a soldering iron...), and we just gave Mouser the parts order since they are running low here and there (a couple items are out of stock until next week or so). However, if you change anything on the BOM in the next few days, I will forgive you :-)

    For batteries, eBay seller "Tywong888" has 600 mAh 9V batteries. These are about twice the price and rating of the 300 mAh variety, which I assume(?) is the basis of the 8-hour spec you have for the O2. Do you have any sense of how linear battery life is (i.e., twice the mAh rating = twice the life)?

    ReplyDelete
  17. I would stick with the 250 mAh Tenergy batteries in the BOM. I strongly suspect 600 mAh is pure eBay fantasy.

    I've ordered a variety of batteries from Asian eBay vendors that are nowhere close to their rated capacity. I have a battery analyzer that plots discharge curves and precisely determines capacity. The most laughable were some "BTY 3000 mAh" (printed right on them) double AA Ni-Mh cells that all tested around 500 - 600 mAh or 1/6 their rated capacity! For < $1 each they know you're not going to try and return them to Asia.

    By comparison, the Tenergy batteries actually measure pretty close to their rated capacity. I'll bet you'll get longer or similar run time with those than the no-name 600 mAh batteries. Plus they're cheaper :)

    The only higher capacity battery that I've heard good things about is the Accu-Power 270/280/300 mAh. But they're a lot more expensive and I haven't tested any.

    And All-Battery is the main USA importer for Tenergy. They have a big website, a real e-store, and a reputation to protect. So you're much less likely to get fake cloned knock-off products from them. Tywong888, on the other hand, I'm much less sure about.

    To answer your other question at the power drains involved here run time scales fairly linearly with the real capacity.

    ReplyDelete
    Replies
    1. My ambitious friend tried to fit in series of 12 AA size "2000 mAh eneloop" into a (insanely) bulky O2 enclosure, (6 x 1.2volt) for each O2 9V battery plug. Though I doubt the original designed VAC adapter has enough DC offset to feed all the batteries overnight. Is there a way to make it work? As said in O2 Detail article, 8-24 hours to fully charge the suggested 9V battery type. Otherwise the eneloop may take ..weeks?

      Delete
    2. R1 and R2 set the max charge current. 2000 mAh NiMh batteries can handle up to 200 mA continuous charge current but 100 mA is better. At full charge 6 NiMh cells are around 8.4 volts. So with the O2's 12 volt rails, you get 12 - 8.4 = 3.6 volts. Ohms law gives 3.6/0.1 amps = 36 ohms. But, more important, if the batteries are completely dead, the maximum available current from the O2 regulators is around 150 mA. So ohms law yields 12/.15 = 120 ohms.

      So your ambitious friend should reduce R1 and R2 to 120 ohms. The charging time will depend on how deeply discharged the batteries are, but roughly the average charge current will be around 35 mA. So for 2000 mAH completely discharged batteries that's 2000/35 = 57 hours of worst case charging time.

      Delete
  18. Thanks - we have the Tenergy 250 mAh batteries on order, which are on sale for $4.99 today.

    ReplyDelete
  19. Any chance, the desktop amp will have a DAC?

    ReplyDelete
  20. I'm trying to find a commercial DAC that will pair well with either amp. The only cost effective USB interface available for DIY DACs are the ancient TI PCM2xxx parts which are limited to only 16 bits. Commercial large volume manufactures have better options available which gives them a huge advantage over most any DIY DAC. DAC's also require surface mount parts making them much more challenging for most DIYers.

    ReplyDelete
  21. I would say an Music Streamer DAC would fit. Not that expensive, extremly simple and match the size of the O2.

    ReplyDelete
  22. Great stuff here, but... there is one major issue stopping me. As an average person I don't have a bunch of testing gear, after the last solder cools how can I know I've done it 'right'? Just plugging it in at the end isn't objective, if you see what I mean. I would really want it to sound the way it was designed, not the way I happened to build it and have my subjective ears fill in the gaps.

    ReplyDelete
  23. Thanks Anon. See: O2 DIY Testing You'll find various tests that greatly improve the odds you assembled your O2 correctly. I might be adding even more that section in the future.

    The nice thing about a proven design using a known PC board is most errors will tend to be obvious. The odds of getting something wrong that creates a subtle problem are fairly remote.

    That's very different than all those people and companies designing things from scratch without the proper equipment to measure the results of their efforts. There's an almost infinite number of things they can get wrong--many of which might escape their biased sighted listening tests.

    ReplyDelete
  24. NwAvGux - you mentioned this company in another blog post as a company that publishes detailed measurements:
    http://www.violectric.de/Pages/de/shop/index.php
    (sadly I can't afford them - yet anyway)

    Maybe there is a way you could design the desktop amp to be able to accept their DAC in the back?

    ReplyDelete
  25. DetlevCM, that is an interesting idea! I wonder if Violectric would provide the technical data for the 10 pin connector? I would guess the DAC board uses the power supply in the headphone amp. It is also switching between the analog inputs and the USB input (hence the relay).

    The 16/48 board is not so interesting as I think that is the TI PCM2xxx in the Behringer UCA202. You might as well buy a UCA202 and save money. :)

    But the 24/96 DAC uses a Tenor USB chip and a better DAC. So that is more interesting. There are not many low cost DACs (yet anyway) that can properly do 24/96 over USB. It would be interesting to test the Violectric board. But I might also need a V100. And that, as you said, is expensive.

    ReplyDelete
  26. It might be an idea to just contact them - ask and see what they reply. (That can't do any harm)

    If you had too much time it might also be possible to reverse engineer it, but if you need to go that far another alternative is possibly the better choice.

    ReplyDelete
  27. Perhaps someone in Germany can contact them in German? They might be more friendly that way.

    ReplyDelete
  28. I'd really expect them to speak English - and I'd actually think the other way round: International interest is better than only national interest.

    Alternatively, I could try to translate an English email to German for you - at least right now I'm not busy at all. My only worry is, what do I do when it becomes technical?

    ReplyDelete
  29. Either Violectric is willing to provide the specifications for the 10 pin connector, or they're not. It all comes down to if they make a decent profit margin on the USB 24/96 itself or if they make most of their money on the much more expensive headphone amps. If they want to sell more USB 24/96 boards they should be OK to share the technical information. But if they want to protect sales of their expensive headphone amps, don't expect a favorable response.

    If they're paying any attention to their website stats, I'm probably already on their radar as I suspect many have visited Violectric from my blog.

    ReplyDelete
  30. So - will you be emailing them then?
    Or would you like me to translate something for you?

    If you would like me to translate something, I guess you have my contact details somewhere - either from my old email or my Google account.

    It will be interesting to see what their response will be. To be honest I don't think they'll be making a loss on those boards as they also sell a very expensive DAC with USB built in - so offering a USB board for the "cheaper amps" only makes sense if it is at least a zero loss product, especially at these prices.

    ReplyDelete
  31. Perhaps the "Lake People" are really nice guys and will just want to help out? :) I'll dig up your email. Thanks.

    ReplyDelete
  32. instead of looking for a commercial dac to pair it to the desktop version, why not a Objective dac? something minimal similar to the 4397 ebay dac ( http://theartofsound.net/forum/showthread.php?t=315 ) but done your way, as the O2.

    thank you for your effort and time

    ReplyDelete
  33. I've had lots of requests to do my own DAC and perhaps this might be a future article topic. But the short answer is I believe USB is the way to go for most users. And I also think 24 bit support over USB is useful so you can use software volume controls without loss of resolution. But there are no good USB interfaces available to DIYers that support 24 bit audio. Nearly every DIY DAC uses the ancient TI PCM2xxx USB interface and it's limited to 16 bits, prone to jitter, etc. The large OEMs making commercial DACs, however, have access to high resolution USB solutions.

    ReplyDelete
  34. Hmm... it looks like I'm able to complete an order for one PCM1796DB at $5.34 (+shipping) on components.arrow.com Haven't actually entered CC info at the last step, though. But shipping will probably be more expensive than the unit itself. They don't show shipping price. Only method.
    But with a group buy this might be an option.

    ReplyDelete
  35. Hello, I have found your blog very useful and there is much to be commended here.

    Just out of interest how much of a difference to the sound does a DAC make? From what I've read it seems that headphones have the largest affect on sound followed by an amp and then DAC. So if I run 02 amp from my ipod touch 2nd gen (using line out) to my hd 650s will I be really hindering the sound?

    If a DAC really makes a difference and the DAC on my ipod is inadequate then instead of the 02 I have been considering the audio-gd NFB-12, which is an amp+DAC, which retails for $200, so not too expensive but at the maximum end of my budget that I would pay for an amp + DAC. It has had very good reviews and I have heard that is has an excellent DAC and amp, better than those of the similarly priced fiio e7 + e9 combo.

    I was wondering if you could give me a quick opinion of it, and possibly review it some time in the future.

    I am guessing that the 02 has a better amp than nfb-12 but IF the DAC matters than the nfb-12 seems like a more attractive option since it has a good amp+ DAC for $200. Unless of course you can find a good DAC match for the 02 which brings the total cost of the 02 amp + "x" DAC under or equal to $200, bearing in mind that I will probably not build the 02 amp myself so it will cost roughly $130.

    Of course if you believe that DACs don't make much of a difference and my ipod's DAC is good enough than I will more than happily just get the 02 amp.

    ReplyDelete
  36. The PCM1796 doesn't support USB so you're still stuck using something like the PCM27xx or PCM29xx for the USB interface and they only support 16 bits.

    ReplyDelete
  37. Hey really awesome stuff here, helped me understand the importance of headphone impedance and saved me from buying overpriced amps. Now I just need to wait for someone to sell these pre-made since my DIY skill are non existent.

    Ben

    ReplyDelete
  38. Anon, I generally agree it's headphones first by a mile, headphone amps (even if built-in) come next of there are audible problems like insufficient power, higher output impedance, etc. And last would be the DAC. The other issue is audible noise (especially "digital" noises with laptops, etc.).

    The DAC in the newer/better iPods is quite good. The one in my Touch 3G measures very well (some Touch 3G results are shown in the Clip+ review). I don't agree with outboard DACs for iPods.

    ReplyDelete
  39. >And I also think 24 bit support over USB is useful so you can use software volume controls without loss of resolution.

    Vista and 7 convert the bit-depth to 32-bit floating point during the mixing and DSP stage in shared mode. There should be no dynamic range problem with changing the volume in those OSes.

    ReplyDelete
  40. Ferongr, you're correct for some on-board audio systems where the volume control is part of the on-board audio. But for a USB 16 bit DAC (and nearly all USB DACs can only do 16 bits over USB with native OS drivers) you do get less than 16 bit resolution at anything less than full volume. A 6 dB reduction in volume reduces it to 15 bits. A 12 dB reduction is 14 bits. Etc.

    Some argue this still isn't as big of problem as it might seem because the ambient noise floor from the room remains constant but you do lose bit resolution. With the volume at -12 dB you're listening to no more than 14 bits of digital audio. Someone else posted this link, and it's a rather self serving explanation, but it does get the point across: http://www.wadia.com/technology/technicalpapers/Digital_Volume_Control_2.pdf

    ReplyDelete
  41. That's fine and dandy, but you missed what I wrote in my previous comment.

    The Vista/7 UAA mixer operates at 32 bit floating point (and the volume controls as a consequence), so there should be no problems with attenuating the signal.

    >Vista will adjust the volume using internal floating point values rather than integer. So it does not reduce resolution as XP did.

    From http://www.avsforum.com/avs-vb/showthread.php?p=8256922#post8256922

    General consensus at HydrogenAudio too seems to be that UAA's mixer is transparent.

    ReplyDelete
  42. I agree the mixer in Vista/Win7 is transparent. The issue is the external DAC. Some external DACs have a digitally controlled analog volume control in them. The Windows mixer, with the right USB endpoints, can control such a volume control. But few reasonably priced DACs offer that feature.

    Most external USB DACs are at a full 16 bits at 0 dBFS. If you reduce the volume at the PC there's no way around sending less than 16 bits to these DACs. With the PC's (or software player's) volume control at -12 dB, a 0 dBFS signal on the recording becomes a -12 dBFS signal before it leaves the PC. And that's only 14 bits. The top two bits will never be used at that volume setting.

    ReplyDelete
  43. I get it now, it was a misunderstanding. You weren't criticizing the digital volume control but describing a fundamental limitation of DACs.

    This concern indeed has merit, though I personally avoid all PC side volume controls because anyway. Using the amp's volume control is the safest from a usability standpoint (no possibility of jerking a cable and sending a tweeter-frying click, switching inputs into a 2V source and blasting the cones to the other side of the room) and most easy way to keep the combined noise-floor of the chain as low as possible.

    ReplyDelete
  44. My wife is not happy I've spent the last 4 hours doing nothing but reading your blog but its been so worth it! The O2 is AMAZING and I want one. But I have no DIY ability so I will look into ordering one ready made or hope a good O2 shows up on eBay someday soon.

    I also want to add my name to the list of someone looking for a good USB DAC that either can handle lots of headphones on its own or would mate up with the O2. I'm one of those guys wants "no regrets" good sound at the lowest price possible. The O2 is by far the best thing I've come across yet towards that goal. Thanks for everything and please keep up the great work!

    And, PS, I don't blame for you giving up on headfi. I tried to find solid info there but had little luck. Too much empty hype and not enough facts. There's also lots of stuff there that's really suspicious and I doubt is genuine. Plus the constant flashing, scrolling, whatever ads are more than obnoxious. Kudos for having your blog be ad free!

    ReplyDelete
  45. so stoked for the desktop version! I love my cmoy's as much as anyone but an affordable and versatile reference class amp (like the 02) might be the last amp i ever need.

    ReplyDelete
  46. And the boards have shipped from the Chinese manufacturing plant! Won't be long now...

    ReplyDelete
  47. I'm selling 8 assembled boards with batteries and power supplies for $120 each. I can ship by Oct 24th most likely.
    If you are interested check out my thread at diyaudio.

    ReplyDelete
  48. Although I've ordered 2 PCBs of your portable version, I also can't wait for a "custom" desktop version of your PCB. Any idea on when or if you'll be producing one? Ad if yes, will it fit in the B3 enclosure?

    ReplyDelete
  49. I'm still hoping to have at least the first article on the desktop amp published before winter (mid December). Not sure about the B3 enclosure yet. I'm hoping to have room for an optional 24/96 DAC board in the same enclosure which might not be possible with the B3-080.

    ReplyDelete
  50. A question to you NwAvGuy - might sound daft to you, but I'm curious:

    If one looks at other products, the line in impendance is significantly higher. Project Audio uses 49kOhm, Beyerdynamic 50kOhm - but then Violectric uses 10kOhm as you do in the O2 Amp.

    Which is better? Unless I am misunderstanding your articles, you want an impendance as high large as possible to negate any possible effects from an excessive output impendance, or not? And the Line Out impendance tends to be large, or not?

    (I hope it fits into this section - best I can think of - and I'm curious as to what the reasoning for 50kOhm vs. 10kOhm is)

    ReplyDelete
  51. @DetlevCM, that's a fair question and it's addressed in the O2 Circuit Design section of the Design article and in the Circuit Description section of the Details article. The input impedance is a trade off. A lower input impedance reduces noise--especially when the source component is powered off. But if it's too low, some sources may have low frequency roll off due to their output capacitors and some tube gear may have difficulty (such as higher distortion and/or reduced output).

    With headphone listening noise is a priority so it's useful to keep the input impedance as low as practical. 10K isn't a problem for 99.9% of anything anyone would want to connect to the O2. And the 0.1% is probably so flawed in other ways the input loading is unlikely to be the dominant problem.

    The guideline for audio gear is 10K - 50K. So the O2 is still in the normal accepted range.

    ReplyDelete
  52. OK :)

    Thanks for the concise reply - because it did get me wondering, even more after checking and finding Violectric also use 10kOhm.

    ReplyDelete
  53. PCB arrived today. My O2 is operational but not quite finished yet. I had not soldered anything in years and was using a $8 Walmart soldering iron, so since it works, I attribute that to the build being pretty easy overall.

    http://tinypic.com/r/2qv8fvo/5

    Thanks for all the hard work on the design and hunting down parts!

    ReplyDelete
  54. @mikeaj, you're welcome. I'm glad it works! You're living dangerously with all those un-clipped leads that could touch each other or something else conductive. I hope it puts a big smile on face listening to it.

    ReplyDelete
  55. I have just finished building my Objective 2 and I have to say it is awesome, top marks for your design and creativity. I agree with most of your comments about the audio industry and the 2 best amps I have now are your Objective 2 and my cMoyBB 2.03.

    My question is how difficult would it be to have a DIY Objective 2 as a surface mount amp. My reasoning is that whilst I love both the aforementioned amps I find the size not really what I want as I use my amp 80% of the time whilst out and about. Something like the Fiio E11 size would be better from my perspective and I guess from a lot of others, it is a little less "chunky". In addition the 9V batteries do not help the depth of the amp and also they are limited to NIMAH and on Li-Ion.

    So is this possible and if so how difficult?

    Thanks for your time and excellent work.

    ReplyDelete
  56. @Phil, Glad you like your O2. SMT is something most don't have the skills and tools to pull off. I can tell there's been some commercial interest in an assembled full SMT version. I think that's a better solution than me spending a bunch of time designing, testing, publishing, and then supporting, a DIY SMT design. I don't even want to think about all that could go wrong with SMT solder bridges etc.

    If you want a solution ASAP, you could do worse than the FiiO E11.

    ReplyDelete
  57. I am watching this with interest. I hope you can find a way to make available a desktop unit that has is dac + amp. I would buy that in a heartbeat.

    ReplyDelete
  58. I saw that and thanks. I am very excited, good luck with the project. I might be selling my Audio-gd Sparrow in the future :)

    ReplyDelete
  59. Just completed constructing my O2, and the sound is outstanding! Many thanks for releasing this design open source. All your hard work is very much appreciated.

    I have one problem. The amp sounds great when I have the high gain button pushed in; sound is nice and balanced. When I have it on low-gain the sound is unbalanced - the left channel isn't getting as much volume as the right...

    ReplyDelete
  60. @zzzzrrr glad you like the O2! The volume balance problem is very likely from the front corner of the gain switch touching the via (solder) pad by R21. If you heat up the front pins of the switch and use a small tool under the center front of the switch to pry it up slightly the gain problem should go away. There are notes about it in the Circuit Board Construction section and also in the Trouble Shooting section.

    ReplyDelete
  61. Fixed.... I checked the solder leads for the low gain resistors and noticed that I needed to do some touching up. Fixed!!!! Awesome!

    Now I have to make sure I don't give myself hearing damage by listening to this awesome amp at high volume :-)

    Thanks!

    ReplyDelete
  62. Hi NwAvguy,

    I really appreciate all your good work. I have one question for you though.

    Is the default gain setting for the O2 still 2.5x and 6.5x? I believe that Epiphany Acoustics ehp-O2 have a default gain setting of 1x and 2.5x, while jdslabs have the original settings.

    I will just briefly list my headphone rig:

    Headphones: Allesandro ms1 (grado), Hifiman He-300
    Dac: HRT music streamer II
    Amp: soon to be O2

    What will be most beneficial in the long run, I want to keep the O2 for a long while and may decide to buy more demanding headphones in the future. So my ultimate question, which gain settings would be best?

    Thanks!

    ReplyDelete
    Replies
    1. The HRT is a slightly "hot" source. If you plan to use it on battery power you should use no higher than 2X gain. On AC power you can use 2.5X gain. If the HE-300's are typical power hungry HiFiMan planars, you'll want to use one or the other.

      The MS1s are probably much more sensitive and might be happy at 1X gain but I don't know for sure. You need to look up reviews/specs on your headphones and read my All About Gain article.

      Delete
  63. @Anon, it's not that hard to change gain settings later. And it's even easier to clip the resistors to drop either the 6.5X or 2.5X setting to 1X. Without spending time researching all your hardware I don't really know what the ideal settings would be but I've provided all the info you need figure it out yourself. Start with:
    All About Gain

    ReplyDelete
  64. Any chance you'll ever design a ODA with Balanced connections? I know you mentioned a negative on this based on previous Blog updates, but was just curious. I appreciate you mentioned Balanced and RCA basically sounds the same, and I would agree except on my current DAC, the Rein Audio X-DAC, there definitely is an audible difference. Given a specific segment of a certain track blind test A/B'd, I can actually tell them apart. I think it's less to do with the cable itself and more to do with how the X-DAC processes it.

    Secondly, are the O2's designed to match or perform better with a specific headphone or Ohm rating? I'm new to this stuff but I know I've read things before on the importance of impedance matching. My main can I use with the O2 is actually a T1, which can at times sound ever so slightly bass shy, thin and articulate on the O2. Wondered if this was to do with the headphones themselves or the match up.

    Lastly, any possibility of providing an ETA on the ODA? I'm eagerly awaiting one. Currently in the process or reviewing and comparing both the V200 and the O2!

    ReplyDelete
    Replies
    1. Keep an eye on the latest (front page) blog articles here for updates on the ODA/ODAC. If you go back to the first two O2 articles I explain the O2 was designed to drive an unusually wide range of headphones. Unlike many headphone amps, including some very expensive ones, the O2 is almost entirely insensitive to headphone impedance.

      A Violectric V200 offers very similar performance to the O2. The main differences being the V200 has a bit more maximum output, cannot run from battery power, and is much larger, heavier, and more expensive.

      No, I don't plan on messing around with balanced connections as they very rarely offer any real (vs imagined) benefit for home applications. If you honestly found a way to level match both outputs of your X-DAC (balanced outputs are usually higher level than unbalanced outputs) and conduct a true blind test I suspect you would find they sound the same. If they don't, the X-DAC is likely poorly designed. If you're using the X-DAC with unusually long cable runs, or in some environment prone to audible noise, then you may notice less hum and/or interference noise with the balanced connection. But if there is no audible noise, and the X-DAC designer(s) did their job right, they should sound the same.

      Delete
    2. AFAIK balanced connections between gear for typical home cable runs provide no benefit, and are outside the scope of the project.

      Any perceived improvement should most likely be attributed to either bias on the listener (it might not have been the case for you, unless you didn't volume-match correctly) or badly implemented (or otherwise, significantly differently performing) unbalanced outputs on the interface.

      There should be no reason why someone would need a balanced connection for up to 3-5 meters of cable.

      Delete
  65. Just a very slight criticism.

    You listed Lelon electrolytic capacitors on your BOM.

    Lelon are a known bad on various sites such as badcaps and similar motherboard capacitor failure sites.

    Whilst in this application I think it's highly unlikely they would fail, for one it costs next to nothing extra to go with a decent brand, which should increase reliability.

    Too much cheap equipment fails prematurely from cutting corners and screwing the price down, saving money on electrolytic capacitors seems to be pretty detrimental though when they fail and the whole device fails over one component.

    ReplyDelete
    Replies
    1. The main caps in the Bill Of Materials are Nichicon not Lelon. Some O2s have been built with Lelon caps and there have been zero problems. One of the goals of the O2 was to show how good performance can be even using very inexpensive parts. That applies to all the parts. There are a limited number of reasonable priced caps at Mouser that meet the physical and electrical requirements.

      Motherboards severely stress the electrolytic caps associated with the CPU power supply. Modern CPU's require maximum currents in the range of 40 - 100 amps. This creates very high ripple currents in the switching regulator that provides the low voltage CPU core power. The caps are typically run at or even over their design ripple current ratings when the CPU is under full load. In addition they have to be physically close to the CPU which can run at 60C or even hotter. The combination of high ripple current and high ambient temp is what causes all the well documented electrolytic cap failures on motherboards. In that application I agree it's false economy to use cheap and/or undersized caps.

      In contrast, the O2 power supply only has to deliver peak currents of 0.4 amps and the ambient temps are much lower. It's also 60 hz ripple vs 30,000+ hz ripple. So, put simply, it's apples and oranges. It's like comparing tires for a race car with tires for a Toyota Prius. The race car is a much more demanding application and tires wear out much more often.

      Delete
    2. I was referring to C2,3,4 and 5 in your BOM. Unless I'm mistaken.

      I agree, however I have seen capacitor failures in things like cheap domestic hifi equipment with fairly pedestrian power supplies, nowhere near the stresses of a motherboard or SMPS application.

      I agree with your design philosophy completely, I just feel that it is one area that could do with spending that tiny bit extra.

      For me at least decent caps are actually the same price or cheaper than Lelon locally and even cheaper if sourced internationally, so it's a no brainer for me anyway.

      Why not put race tires on my Prius. ;)

      Delete
  66. Greetings once again, NwAvGuy!

    Please allow me to shout the following from any and all convenient rooftops:

    Using my JDS obtained O2, 4G iPod Touch and Grado SR-60s, I am convinced I heard music indistinguishable from live music on more than a single occasion in the recent past. I can't say that about every music file I have on my iPod and I acknowledge that I am not a music 'expert' of any sort. All I can lay claim to regarding any notion of personal 'credentials' is tied to 40+ years of heartfelt listening.

    I can say that such sonic experiences never happened to me prior to my acquisition of the 02. Even those who claim, perhaps with some validity, that the Grados are playing the lead role, must then also acknowledge the inevitable probability that it is the 02's acoustic invisibility that is allowing the Grados to reach their zenith of performance.

    My thanks again to you and to Mr. Seaber. It's been a revelatory (revelationary?) blast.

    Jonathan

    ReplyDelete
  67. Hi,

    recently purchased Audeze LCD-2 revision 2 the specs are;
    • Impedance: 60 Ohms, nominal
    • Maximum diaphragm excursion: 2.5mm p-p
    • Efficiency: 91 dB/1mW
    • Maximum output: 133dB, 15W

    How would the o2 cope with that? What is its power output in Watts to this headphone? how loud will it drive the LCD-2 in decibels, loud enough? I ask as I have purchased your amp and plan to use it with the LCD-2.

    I have a friend electronics engineer and he was saying from a rough quick estimate in his head, that to drive the LCD-2 to full power you would need roughly 400wrms (into 8 ohms) to drive it to full power! Mind you max output of the LCD-2 is a crazy 133db. That's a large full size speaker amp heh.

    So will the o2 suffice?

    thank you,

    Matt

    ReplyDelete
    Replies
    1. There are many happy LCD-2 users using the O2. It has power to spare to drive the LCD-2 to hearing damaging levels and beyond. Your engineer friend needs either read my More Power article or start using the correct LOG button on his calculator. He's very wrong.

      The LCD-2 hits 110 dB with about 1 volt of input. Peaks of 110 dB are considered plenty loud by most people (hearing damage starts past 85 dB). 110 dB is standing on a sidewalk next to a jack hammer tearing up concrete.

      Even if you plan for 115 dB peaks, which is extremely loud, you only need about 1.7 Vrms which is 0.048 watts. The O2 can provide 5+ Vrms which is good for 124 dB SPL with the LCD-2. And that's running from battery power.

      If you really wanted to push the headphones to 133 dB you would need 14 Vrms which is 3.2 watts not 400 watts.

      Delete
  68. Thanks,

    obviously their was a mistake somewhere, to be fair to him we only talked vaguely and briefly. The logic I think was using a loudspeaker amp for full size room speakers (typical 8 ohm speakers) and would provide a lot less power into 60 ohms compared to 8.

    So the voltage is fine with your amp, what about current required for the LCD-2? Forgive my lack of tech knowledge.

    thanks! :)

    ReplyDelete
    Replies
    1. This is all explained in both the More Power article and the O2 Design Requirements.

      But, long story short, even at an ear splitting 124 dB SPL that's a peak voltage of 7 volts. Ohms law gives a peak current of 7/60 = 117 mA. The O2 has a peak current capability of 200 mA per channel--nearly twice what's required.

      Current with the LCD-2 is where many headphone amps will fall on their face--especially OTL tube amps, Cmoys, and anything using a typical wimpy virtual ground or 3 channel design.

      Delete
  69. Hi NwAvGuy,
    While troubleshooting my o2 build, I had Q1 reversed and was reading -18V on the negative supply line. I desoldered & reversed Q1 but have yet to test it. Do you think I will I need to replace it (and any other damaged components)? Thanks....

    ReplyDelete
  70. Hmmm.... if the regulators (U5 and U6) are working correctly the power supply lines should be +/- 12 volts regardless of Q1 and Q2. So I'm not sure what you measured?

    Generally the main issues with Q1 and Q2 involve ESD damage to the gates of the MOSFETs. Simply installing them backwards would only compromise the turn on and turn off behavior of the O2. If you really measured 18 volts at the op amps that's a different problem. If you really somehow managed to deliver 18 volts to the op amps, I would replace all the op amps and U2 along with U5 and U6. But, honestly, that doesn't seem a likely possibility.

    ReplyDelete
    Replies
    1. I was following the testing procedures so thankfully didn't have any opamps installed. the -18V on the neg rail was measured right after 7912.

      At one point i had U2 installed, and was doing the resistance measurements, and strangely came up with something strange (230k? 330k? i forget) for R24, although the installed resistor is a 270k. Not sure if this is related.

      Actually i should not have even plugged it in to check the voltages when i noticed the resistance problem, my bad. but i figured the only thing i could damage would be U2 or the mosfets. I need to do more testing now. appreciate your help!

      Delete
    2. PS: my rail measurements after the regulators were +12 and -18. i.e. it is only the neg rail that is affected. i will measure again with the proper Q2 installation tonight and lyk what i find.

      Delete
    3. With respect to ground if you're seeing more than +12 volts or -12 volts on the supply lines you need to make sure U5 and U6 are working properly. I would suggest following the voltage measurements in the O2 Troubleshooting section if you haven't already. The voltages at D1 and D5 are especially important.

      Delete
  71. update: with no opamps in place, and Q2 in proper orientation now, i still measure +12/-18vdc just past the regulators for pos/neg rails respectively. i will revisit the troubleshooting when i have the time. thanks.

    ReplyDelete
    Replies
    1. Odds are good U6 is either in backwards, the wrong part, defective, or perhaps most likely, you have a solder bridge between two of the pins on the top or bottom of the board.

      Delete
  72. Hehe build went all out and got myself machined some stainless steel front panels for the O2: http://yzrandome.blogspot.com/2012/04/objective2-headphone-amplifier-ss-front.html
    What do you guys think?

    ReplyDelete
    Replies
    1. That looks very nice! Stainless is way more difficult to machine. Yours is the first I've seen. It's great to see all the ways everyone are making their O2 unique.

      Delete
  73. This might seem like a silly question but do you think it's OK to leave the battery O2 plugged into mains 24/7?

    ReplyDelete
  74. @Divvy, yes you can leave the mains plugged in 24/7. If you're not going to use the O2 from battery at all for a month or longer, you should remove the batteries.

    ReplyDelete
  75. Is it normal to have a little bit of channel imbalance on the O2 at extremely low listening volumes? I can hear the right channel cut off first when the volume knob is almost at the 0% position but otherwise I don't notice anything at regular listening volumes.

    ReplyDelete
    Replies
    1. Yes, it's normal if you're talking so low you can barely hear the music anymore (around -50 dB). If it's happening at levels you want to listen at, then your O2 is configured with too much gain. Use the low gain mode (button out) if you're not already. And if you're already using low gain, consider changing or clipping the gain resistors (see the gain section of the O2 Details article). That lets you use more of the volume control's range and you should not hear any channel imbalance.

      Delete
  76. I’m looking for a suitable power supply that will work in Europe/Germany. What do you think of this:

    http://www.henri.de/stromversorgung/steckernetzteile/steckernetzteil-9v-ac/19430/wechselstromnetzteil-9v-12v-15v-18v-24v-ac-schaltbar.html

    My main concern is that it says “not stabilized ac-ac converter”. What does not stabilized mean? Does that have any effect on the O2? The other issue is whether the plug will fit into the JDS Labs O2. Their specs say 2.1x5.5mm for the plug, whereas this is 2.1x5.0mm.

    I realize this power supply is oversized, but it’s one of the few viable options I have here. The other power supplies capable of at least 14V have >20A output and thus are much, much, more expensive.

    What voltage setting (15 or 18) should I use given that this power supply is able to output more than enough for the O2 either way? Does it even make a difference? Or does it have any drawbacks that the voltage setting is switchable?

    ReplyDelete
    Replies
    1. Yes, that should work nicely set for 12 VAC or 15 VAC. In fact, I should add it to the BOM and article as 15 Euro is a good price. the "not stabilized" part is OK and true of all AC wall transformers. The regulators in the O2 provide the stabilization.

      Delete
  77. Oh, wow, I didn't expect that kind of reply :)

    If you do include it into the BOM, you may also want to put a note there about shipping charges - 6 Euro within Germany, 15-30 Euros for the rest of Europe. So buyers may want to check for a retailer in their country first.

    After a bit of googling I found this spec sheet with a model number by the manufacturer, which could be helpful when searching for stores selling the power supply:

    https://www.distrelec.at/ishop/Datasheets/230924AC_eng_tds.pdf

    ReplyDelete
  78. It seems that J1, as described in the BOM, is no longer in production. What is an acceptable replacement? Mouser suggests Kobiconn 163-7620E-E, but it won't be shipping until June. I was hoping there was another part that could replace it that is in stock.

    ReplyDelete
    Replies
    1. I'm working on this issue. The CUI jack listed from Digi Key will fit but you have to trim down a couple of the pins slightly. I've ordered a Kycon jack from Mouser as a sample.

      This may require a revision of the O2 PCB to fix properly. In the meantime it may require some minor modification of another jack. Mouser had around 25000 of the jacks in stock so it's a surprise they now have none.

      Delete
  79. Just a bit of constructive feedback from me:
    I don't know if it's just me but it feels as if the Kycon 3120 3.5mm female jack doesn't really "hold in" the standard 3.5mm male plug quite as strongly as most other ones. It's not really a big deal but I wish my headphones would stay plugged in a bit more firmly as I've accidentally pulled them out on occasion.

    ReplyDelete
    Replies
    1. Could it be the front panel isn't letting the plug go all the way in (i.e. the jacks are not flush with the panel)? I don't have that problem at all. In fact, the jacks hang onto plugs better than a lot of other 3.5mm jacks I've used. If it's not the panel it might be you have a 3.5mm plug that's not made properly or perhaps your jack is worn out from lots of use.

      I am looking into options for the ODA for higher quality jacks. The 1/4" jack will likely be fro Neutrik.

      Delete
    2. Nah it's all the way in and my 3.5mm plug is from Neutrik so no issues there...either my jacks are a somewhat looser batch I suppose or my reference is skewed since my computer and phone seem to hold in the plugs extremely well.

      1/4" Neutrik jack for the ODA would definitely be a plus.

      Delete
  80. I have some questions about re-locating jacks and switches for placing the O2 in a different enclosure that I don't remember seeing addressed before. I might have just forgot about it though...

    Would something shielded like miniature balanced mic cable have any benefit over the normal twisted pair you've recommended before?

    Also, what switches/jacks would be "safe" to reroute without compromising performance? Does anything besides the gain switch need to stay put?

    ReplyDelete
  81. IMPORTANT UPDATE: Mouser had over 20,000 of the Kobiconn 163-7620E power jacks when I designed the O2 last summer. The part was listed as current with no "EOL" (End Of Life) flags. Less than a year later it's been discontinued and Mouser has sold out of them.

    Unfortunately, coaxial power jacks differ slightly in their PCB footprints. The Kobiconn 163-179PH is in stock at Mouser, and it will fit by trimming down two of the leads (tabs) slightly. The same is true for the CUI PJ-002A from DigiKey listed as a substitute in the O2 BOM(tabs) to fit the O2 PCB.

    Mouser has a replacement for the original 163-7620-E on order but it's not clear if that will fit correctly either. I'm tentatively planning to wait until the new replacement version is available and, if it fits, that solves the problem.

    If the Kobiconn replacement part is not going to work either, I may slightly revised the PCB to fit the 163-179PH and CUI PJ-002A jacks. At the same time I would move the via out from underneath the gain switch so that's no longer a potential short (causing one channel to have lower gain than the other).

    Contrary to some rumors, I'm NOT planning a major revision of the O2. The above two minor fixes are the only things I'm aware of. The idea is to keep the changes simple so prototype boards are not required before publishing the new PCB files.

    If others have found a power jack that fits the existing board, please post the part number and source here. Thanks.

    @Maverickronin, I tried to address relocating components in the Circuit Board Construction section of the O2 Details article. In general you can relocate anything as long as it doesn't ground to the panel (except the input jack which can) and you use proper short twisted pair wiring and keep it away from the power supply section. Some have even relocated the gain switch but most just turn the board around which puts the gain switch at the back of the enclosure.

    ReplyDelete
  82. Hello, just wondering, is there any sound quality increase or faster battery charging times if I use a 16/18VAC 400/500ma adapter over my 13.5VAC 200ma one?

    ReplyDelete
    Replies
    1. No, 13.5 VAC should be fine unless you have some rare headphones (like HiFiMan planars). The power supply "rails" of the O2 are regulated at 12 volts. As long as the adapter can supply enough power to maintain 12 volts when the amp is working hard, a higher output adapter will only make the regulators run hotter.

      There's more info in the O2 Details article in the Components section.

      Delete
  83. I have spent two weeks subjectively testing Hifiman's He-500 on a JDS Labs O2 and a Schiit Lyr. I purchased the Lyr before I had heard of the O2. Once I read this blog I just had to try it for myself so I bought the O2. I have tried my best to pick the Lyr over the O2. I have gone back and forth, back and forth, over and over. I read this blog again, and most everything I could find on the internet pertaining the Lyr. I purchased higher quality tubes to improve the odds that I would decide the Lyr was better for me. I am not what I would call an audiophile, but I do have a few years experience (57 to be exact) of listening to some decent quality audio of the speaker kind. I know what I like, and I can hear a difference. I will be putting the Lyr and the tubes up for sale. I've convinced myself the O2 is for me. Well done NwAvGuy. Add my name to the list of the converts.

    ReplyDelete
  84. hello,
    i just finish the O2,and i works fine,but after several time,there is a problem:
    when i restart the O2 after shutdown it,there is a big clip in the Hphone
    I mesure the pin4 and 8 of the ICs,+17,6v but 1or2v pos on the -rail
    while testing,it appears that by putting the DMM on the pin4 of the 2903,the- rail is delocked,and i obtain +17and -17 and the O2 works fine
    i switch off and on and the -rail is locked again
    Can i suspect the 2903 or tne N mosfet
    many thanks
    patrice

    ReplyDelete
    Replies
    1. Please see the O2 Troubleshooting section. The MOSFETs are nearly always the problem from ESD damage during assembly. If you need further help, please post to the DIYaudio O2 thread.

      Delete
  85. Hi nwavguy,
    Regarding battery for O2, I come a cross a few 9V battery of capacity higher than 250mAh. Something like this even reach 400mAh(NiMH still).
    http://www.ebay.com.sg/itm/2-x-9V-9-0-V-Volt-400mAh-Ni-MH-6F22-PP3-17R8H-Rechargeable-Battery-Ultracell-/130688964962?pt=US_Rechargeable_Batteries&hash=bin1e6dab5962#ht_2995wt_939

    There are even Li-ion ones having more than 500mAh,but i don't think it's safe to charge it using O2 circuit.
    I am wondering whether are they legitimate and safe to use in O2 as I didn't notice you mentioning about 9V battery more than 250mAh.
    My apology if yuo have answered it somehow.

    Regards,
    CCE

    ReplyDelete
    Replies
    1. There are lots of fake claims for batteries being sold on eBay. I would stay away from all of them. If the best battery companies in the world, can only manage around 250 mAH, it's not reasonable some no-name Chinese battery will have 400 mAH. I have some 3000 mAH AA Chinese no-name batteries that test at only 500 mAH. They're junk.

      You're correct you don't want to use Li-Ion or Li-Poly. It has to be NiMh.

      Delete
  86. Thanks for the clarification nwavguy. I noticed you mention about it in an earlier comment.
    Since it's from Ultracell(a short google turn out to be a reputable UK company,although the battery is not in the product line) so I started to believe. On second thought, it's not unusual for ebay to have all sorts of ridiculous fakes.
    So thanks again for the advice, I'll stick to reputable 250mAh from reliable source.

    Regards,
    CCE

    ReplyDelete
  87. Hi nwavguy, my amp is now built, and i am in the troubleshooting stage. 2 issues i could use your help with. note that i have added (2) 1/4" inputs and added a parallel 1/4" st. output as well. the input mod was intuitive to me and although I tried to preserve your design of shunting to ground when no input present and your grounding scheme, i may have overlooked something. please view at http://www.whatsound.net/o2/my%20o2%20input%20jack%20mod.png

    1. significant buzz & low-sub-bass hum gets louder as volume is turned up. buzz at all times when amp is turned on especially when touching the enclosure, increases with volume pot. verified that one of the input jacks is tied to chassis and measures 0.2 ohms (almost a complete circuit). i have tried to keep this the only point of contact with the chassis (i may have failed?)

    2. sounds like 1 channel is out of phase. all input & output jack combinations result in a karaoke effect. however, no problem when only the left OR right 1/4" input is present, and the other not present: in this case one side sounds full and normal & is heard equally in BOTH ears. either 1/4" input will produce this dual mono in both ears. i.e. both L&R inputs are mixing somewhere and going to BOTH outputs instead of just their own. need to track this down.

    any help appreciated, especially with #1
    thanks nwavguy! really looking fwd to enjoying this unit.

    ReplyDelete
    Replies
    1. It sounds like you got the pinout of your jacks wrong and you have swapped ground and the signal somewhere. That's causing both the phase reversal and likely the hum (because you have connected the signal to the case and there is no ground). Your schematic looks OK, but I suspect you have one or more jacks wired incorrectly.

      If you haven't already, you need to find an online datasheet for your jacks and figure out which pin is which. Match up Tip Ring and Ground to the same with the O2's 3.5mm jacks. You can't assume from the position of the pins or lugs on the jacks.

      You also should be using tightly twisted pairs for each signal wire with one lead being ground. I pull pairs out of the inside of scrap CAT5 ethernet cable. Keep the wiring away from the power supply side of the board.

      Delete
    2. Following up...

      in the midst of adding 1/4" inputs (normalled to the standard 1/8"ST in) i hooked the ground into an empty hole on the pcb near J2. just had to move it over to one of the ground holes. found it a lot faster thanks to your help.

      This thing sounds great. absolutely no hiss even when volume maxxed with no input. i have heard a lot of excellent equipment, mic pre's, mastering plugins, etc over the years. this box has a very tight controlled sound with full extension in both areas. i am getting more defined low end, a lot less ~300Hz mud and mix clarity, as well as smooth air. comparing to a Yamaha/Steinbery MR816x audio interface hph out jack. as a mix tool the clarity is already helping my rather warm Senn 650's. this will definitely help me mix low end better, and be less fatiguing in the process. it is a more restrained & balanced sound, less overdriven, less blasting low mids in your face...more air, much better imaging, more refined. i'd say it's a big step up for experienced listeners, depending on what you're coming from.

      my background-I have been a semipro/parttime audio engineer & studio musician since the 90s. I make stock music, sample libraries and I wrote Gardner's Guide to Audio Post-production about film soundtrack mixing.

      i am still getting a light hum only when the power cord is plugged in. will try to uncoil & reroute the wallwart cord.

      Big Thanks NWavguy, loving your approach to the art

      Delete
  88. My O2 by Epiphany is missing the ground wire to the enclosure. Where exactly is pin 1 of the input jack? I couldn't see a label for it on the PCB itself as well as in the documentation package. Also, is there any other way for attaching the wire other than soldering it? (My soldering iron is a fairly bulky one, and I have virtually no experience with PCB soldering)

    The other issue was about the enclosure orientation. In your O2 details article you mention the v-shaped grooves need to be on the bottom (this is how my O2 was assembled by Epiphany), yet on diyaudio.com you said the grooves need to be on the top. Which one is correct?

    Thank you!

    ReplyDelete
    Replies
    1. The ground pin on the input jack is the "center" one closest to the front of the board (between the two open holes). And, no, I don't know of a way to reliable attach it without soldering. It's a bit concerning Epiphany is leaving them off. Because the pin is by itself, and not terribly delicate, even with a big soldering iron you should be OK if you're careful.

      Delete
  89. I have a question on opamps, I had two o2's at my disposal at one point and notice one sounded slightly different than the other. One has a "Brighter" cleaner sound with a smaller soundstage than the other. Curious about this difference, I had an extra set of 4556 and 2068 opamps laying around from another build and decided to swap them out with the ones in my build. The 4556's did not seem to make a difference, but when switching the 2068 Op-amps, I notice that with one, the soundstage is a little wider, but not as clean/bright sounding and with the other, the soundstage is smaller but brighter/cleaner sounding. I have also tested this somewhat (I say somewhat because I am expecting there to be a difference) blindly with someone switching out(or not switching out) the opamps and can still tell the difference. Is there a chance that one of them is damaged? or something is wrong with my build that would allow these two to sound different.? Am I just crazy in hearing these differences? I want to believe they should sound the same but after many tries they sound different and I am not sure which one should be the "correct" sounding one for the O2.

    ReplyDelete
    Replies
    1. I strongly suspect you're hearing a difference because you expect to or perhaps you're listening at slightly different volumes. I can't rule out some obscure problem with one of your 2068 op amps, but that would be easy enough to verify if you can get your hands on another 2068 or even another O2.

      I've done level matched fully blind ABX testing comparing two O2s with various op amps in each and have not heard any difference. Others have done the same (because they didn't believe op amps can really sound the same) and reported they couldn't hear any difference either even when comparing the 2068 to a 5532, etc.

      Delete
    2. That is what I was wondering and debated bringing up the question on here because of reading the articles on opamps and sound.. I have since ordered replacement opamps that should be here soon to test out. I figured it was worth $5 for the opamps and shipping for me to know if one was bad or not as I suspect one might be because one does seem less detailed/and less clear than the other...so I will switch all out with the new ones when I receive them and know if I am crazy or not. Swapping the opamps in the same O2 without touching the volume I do seem to still notice this difference...but it still may be psychological too...so we'll see. Maybe I'll have someone else "Swap them out" without me knowing if they did or not and see if I notice anything.!
      BTW I'm listening to Allen Toussaint - The Bright Missippi right now via my O2/ODAC combo and a pair of ATH-AD900's and it still sounds wonderful!

      Delete
  90. Hi I would like to use the O2 amp with the XDA-1. And here's my problem. It is supposed to hit 12 V peaks (or at least 9V peaks) in its standalone DAC mode with its volume control bypassed. This gave me trouble with my current HA which is a Xenos 3HA making the audio sound thin and not very good. Will this be a problem if I use the O2?

    ReplyDelete
    Replies
    1. I'm not sure what the 0 dBFS output of the XDA-1 is without the volume control. Because it's also designed to be used as a standalone "preamp" it likely is capable of higher output at full volume to directly drive most any power amp. But with the volume control bypassed, it likely has close to the Redbook standard of 2 Vrms at 0 dBFS.

      It should work fine with the O2 at the default low gain of 2.5X with the O2 running from AC power (which should be OK as the XDA-1 has to run from AC power). On the off chance the output is still too high, you can clip two of the gain resistors in the O2 dropping it to 1X in either gain mode (instructions are in the O2 Details article). You may want to do that anyway if your headphones don't need more than 2 Vrms (many don't). See my All About Gain article.

      Delete
    2. Thank you. Basically what happens is bypassing the volume control is running it at its maximum volume (80). And that is considered it's standalone DAC mode. So I do wonder if they forgot about headphone amps or the proper redbook standard. I currently run my emo power amp using it and it sounds fantastic. But not so much with the headphone amp which leads me to wonder if it is properly implemented. As it's rated as having 12 v swings for the xda-1 (or 9v swings if you look at the ERC-2 soecs) both according to hometheaterhifi. Thank you for the info nwavguy. I am inspired and impressed by what you have been doing and so are a lot of people on the emotiva forums. :)

      Delete
  91. Hey NwAvGuy,

    I really love those charts and figures in your measure, which I believe is the spirit of "reference". I've read a lot of your articles and I become a fan of you. Thank you so much for all you presented here.

    I learned from a main chinese hifi forum that there is a new released portable amp called "SHONYUN-306". It is said to be the best portable amp in its price level. (equiv. 140USD), which is supported by many "expert" claiming how notable the designer is, how great the spec is, how amazing the subjective comparasions are. While, I am not quite convinced of what they talked. It could be misleading if some of them are interest related.

    I am wondering if it is possible that you do measurement on that amp when you can have access to it.

    ReplyDelete
    Replies
    1. I don't have time lately to even publish the last ODAC article on time so I especially don't have time to test Chinese gear right now. If someone is claiming the amp has a great "spec" I would ask for the proof in the form of real measurements made with an Audio Precision, Prism dScope, or R&D audio analyzer. If they can't produce such measurements, it's likely the amplifier hasn't even been properly measured. In which case, how does anyone know it's properly designed?

      Delete
  92. Loving the amp but quick question. I couldn't find the answer anywhere...

    What is the right sequence to turning on/off, plugging/unplugging my amp and headphones?

    I normally start with heapdhones plugged in, turn on source, turn on amp at very low volume. If I want to disconnect the headphone or use a different headphone, I turn off the amp before unplugging headphones or plugging in new ones. Is this right? I've read some people say its better to have the volume low and then plug in/unplug headphones. But I've also read every designer's recommendations are different.

    Thanks,
    Isaiah

    ReplyDelete
    Replies
    1. Ideally you plug and unplug the headphones, and power your source on/off with the amp either off or at very low (or zero) volume. That's all you need to be concerned with. The O2's slight "tick" and "thump" on power on and power off is harmless to headphones.

      Delete
  93. At what level do you believe the o2 with suggested opamp will start distorting and clipping?
    I was looking to drive a stax 202 with the transformer for speaker amps?
    Do you think that amount of power can be taken from the o2?

    ReplyDelete
    Replies
    1. The O2 cannot drive any electrostatic headphones, even with a transformer.

      Delete
  94. I've seen that you not keen of idea of DC-DC power supply, but for many use cases when you use ODAC +O2 and have +5V and need just 1W at +-12 ...
    What do you think of using Murata-ps mea1d0512sc http://www.murata-ps.com/en/mea1d0512sc.html ?
    With the proper simple LC filter noise can be lowered to 5mV p-p max.

    ReplyDelete
    Replies
    1. It really depends on what headphones you want to drive. The O2 requires 400 mA peak current to drive the full range of headphones. That would require a 10 watt DC-DC converter to prevent the converter from shutting down or overloading on peaks. With only 2.5 watts available from USB, with 80% efficiency, that's only 2 watts out which is only 83 mA total.

      The ODAC requires about 130 mA at 5 volts (.65 watts) so that reduces the total O2 power to 1.85 watts in or only 62 mA of output current with 12 volt rails. From that you have to subtract the quiescent current for the O2. You're left with less than 20 mA per channel of usable output current. That's not enough for a great many headphones.

      Finally, it's not just noise on the DC power rails, but also electromagnetic noise. DC-DC converters throw off relatively large amounts of EMI which readily couples into nearby audio circuitry. So you ideally need a large enclosure with physical isolation of the DC-DC circuitry. If you want an example of what happens when you ignore DC-DC EMI see my Mac Book Air review.

      Delete
    2. Just measured O2 consumption with my D150 250Ohm headphones
      In DC side (at Power switch) on +23.5V DC it consumes around 23 mAmps idling and 26 mAmps on max volume (at x2.5 Gain) that is really so loud that can damage hearing. So we get 0.6W of max consumption...
      BTW I just made another step to portable ODAC + O2 combo http://www.pbase.com/objective3/inbox

      Delete
  95. I would like your opinion about the role of the BUF634 as a high current output stage for driving headphones. It is used in many DIY amplifiers which is why i would value your opinion regarding it especially as a alternative to the NJM4556.

    ReplyDelete
    Replies
    1. The BUF634 is a great output stage for a headphone amp but normally isn't required. It's used in the Benchmark DAC1.

      It's main advantage over the 4556, in the larger packages at least, is higher power dissipation allowing higher voltage supply rails. It also has lower high frequency distortion at high power levels but the 4556 is already well below the audible threshold. And it has somewhat higher peak current capability compared with two paralleled 4556 sections.

      The downsides are higher power consumption, much higher cost (especially in the large packages), you can't have any gain (which is desirable in some applications), and the BUF634 can be fussy for high frequency stability when no output network or resistor is used.

      Basically for the 2% or so of headphones that are outside or borderline for the O2's use of the 4556, the BUF634 is the better choice. You can run it with +/- 15 volt rails which might cause the 4556 to get too hot.

      Delete
    2. Thank you very much for your clarification. I am a Audiophile who like you has found a dearth in high quality *affordable* gear forcing me to design my own.

      I am also a Consultant Physician and I welcome your scientific method indeed.
      I would ask the reader's if they would like it if doctor like me tried to treat them with medications based on "how well the medications taste, and how expensive they are etc", rather than based on evidence of efficacy based on Randomised clinical trials demonstrating efficacy of drugs? If in my medical practice i tried to adopt analogous to a "design by ear" principle of treating patients instead of high quality evidence based medicine i wonder how many patients would stick with me :)

      Sir, do please continue your scientific analysis and debunking of audio myths. Someone needs to do it and you are one of the few people who has the balls and integrity to do it without a vested interest. This is the 21'st century and sadly Audio seems to be still stuck in the dark ages :)

      Delete
  96. NwAvGuy, has anyone you know looked at whether this design can be shrunk to make it more portable? I don't know which components are the limiters but it would be really cool if it could be done. BTW, incredible work!

    ReplyDelete
    Replies
    1. It could be shrunk as a commercial design, with some tradeoffs, but it's not practical as a DIY design. The batteries are by far the limiting factor and, because of the power supply requirements, moving to Li-Ion batteries is not trivial (and potentially not safe for DIY). Also the SMT versions of the output stage op amps have much lower power dissipation creating thermal problems driving some headphones. And, finally, the jacks, volume pot, and switches would still consume quite a bit of PCB and panel space.

      Delete
  97. Hi NwAvGuy! I need your help.

    I just soldered the O2 amp. And some measurments with RMAA dissapoint me. The strange thing is THD level at low frequencis (<100 Hz) with load of HD595 headphones. For 60 Hz tone is -76 dB at 2nd harmonic, -85 dB at 3rd and -106 dB at 4th. Distortions at 1 kHz is absolutley negligible. If I connect amp directly to input of my audio card low freq distortions is absent.

    My setup is Xonar D1 > O2 amp at 1x gain + HD595 headphones > Xonar D1 output.

    What it can be? How can I solve this problem?

    ReplyDelete
    Replies
    1. First of all, all sorts of things can go wrong with RMAA. Please see my RMAA article if you haven't already.

      Second, because you're using the same sound card for both source and measurement, you're likely creating a ground loop within the O2. The input and output grounds of the O2 are carefully designed to keep the high current signals in the output stage away from the critical input. But you're connecting those grounds together through your Xonar soundcard and that will greatly degrade the O2's performance.

      Third, the distortion you measured is still easily below the threshold of audibility. Second harmonic distortion is very difficult to detect even in the critical midrange at -60 dB and it's even more difficult to detect at low frequencies. So -76 dB is way below the audible threshold. Likewise, the rule of thumb for odd order harmonics is -80 dB and again your measurements are comfortably below that. If you don't believe these thresholds, please see my What We Hear article where it's explained in more detail. In the Tech Section, you can download a test file and listen for yourself.

      So I'm fairly sure what you're measuring is because you're operating the O2 in a way it will never typically be used--with the output ground connected to the input ground. Headphones normally "float" and are not grounded in any way. This is one of many advantages to using a true audio analyzer (like my dScope). The analyzer's input and output grounds are fully isolated.

      Delete
    2. What level were you testing at ? At high volume, the distortion could easily have come from the HD595 through the output impedance of the O2. Even though the damping factor is about 100, the HD595 can have a few percents of distortion at 60 Hz already at 100 dB SPL (~0.2 Vrms), and is one of the most reactive full size dynamic headphones, so the -76 dB output impedance distortion is not impossible.

      Ground loop issues with sound card measurements can be worked around, but this may require familiarity with DIY electronics, and/or using software other than RMAA.

      Delete
    3. To NwAvGuy,

      Thanks for answer. Can I just add some additional ground to lower this THD? I know that I cant hear this distortions but its interesting thing.

      To Anonymous,

      I tested at very high level. About 2 Vrms. But I think mechanical distortions (that depend on signal level) dont affect electrical distortions that I measured.

      Delete
    4. Ground loop related noise will not be a problem in normal use. Headphones "float" and are not grounded, but when you connect the output of the O2 to a grounded ADC input, you create an artificial problem. Adding more grounds won't help (and will probably make it worse). The problem is you are coupling output ground signals back to the O2's input stage through your soundcard.

      @Anon has a valid point about headphone distortion (which could be significant at 2 Vrms). The O2 has a 0.5 ohm output impedance and the headphone non-linearity could have some measurable, but still inaudible, influence on the output.

      Delete
    5. I suggest reading the Benchmark article about the "sonic advantages of low impedance headphone amplifiers". With any amplifier that has non-zero (even if it is low) output impedance, some of the distortion from the dynamic headphone driver is fed back to the output of the amplifier. 2 Vrms into a 50 ohm HD595 is an extremely loud sound, and could easily result in distortion above 10% at 60 Hz. It is not surprising if about a thousandth of that can also be measured on the output impedance of the O2.
      By the way, note that the distortion of the input of the Xonar D1 starts to rise at about 1.4 Vrms, from less than 0.001% it gradually increases to about 0.1% at full scale (2 Vrms).

      Fixing ground problems with sound card loopback tests is not as simple as adding additional ground, you may want to read the RMAA article (Feb 2011) and comments for more information.

      Delete
  98. This is very cool. Are headphone amps beneficial for "normal" headphones or are they mainly for those high-end, high ohm headphones? I have a pair of $80 Grado SR60s, fwiw.
    Thanks!
    I might have a hard time proving I'm NOT a robot. For all I know, I could be like Daryl Hannah in Bladrunner! JRRRR CHIKA BINGBING ;P

    ReplyDelete
    Replies
    1. See my Feb 2011 articles on impedance. The short answer is: It's complicated.

      A lot of headphone outputs have an output impedance that's too high for even some modestly priced headphones--especially balanced armature IEMs (like those from Shure, Etymotic, etc.). Those devices benefit from a good headphone amp with a lower output impedance.

      The lower-end Grados are relatively easy to drive and if they get loud enough with your current source, you probably won't benefit much from an amp. Still, a lower impedance source may audibly help the SR60s by delivering flatter frequency response and better controlled bass. So it really depends on the output impedance of your source. The typical 7 ohm output of an iPod is too high for many low impedance headphones like your Grados.

      You may also want to check out my "More Power" article from September.

      Delete
    2. Hi NwAvGuy,

      I just wanted to take the opportunity to thank you for all the outstanding information you are sharing on your blog, and especially for designing something as brilliantly performing as the o2. While I had trouble trusting all those feature claims before, I now cannot help but praise your achievements after I just finished my own o2 build last night (got a parts kit from Head'n'Hifi in Switzerland). Ever since I am running around with my cans on the head and a big smile on my face :)

      Delete
  99. Hi NwAvGuy, I've just tried DC-DC bipolar converter to power O2 from USB of ODAC, http://www.pbase.com/objective3/dcdc
    however it is 1W, it works perfectly (no lack of power and noticeable noise). As it can be a nice power option for ODA-ODAC combo, It will be great if you can do the measurements for it with O2 and ODAC combo. I can Mail you ready assembled converter. pls send me the address to chouck-wap(AT)nm.ru

    ReplyDelete
    Replies
    1. Thanks for sharing. There's work already underway on a USB powered O2+ODAC so I'm going to wait and see where that goes, and if it becomes a reality, I'll be measuring it and publishing results. the one caution the O2 can't manage the same maximum output with even a 2W DC-DC converter as it can on battery or AC power. But that doesn't matter for most popular headphones.

      Delete
  100. Using 250 Ohms DT-150s with O2+ODAC I've noticed that even in the car I am using only 1/3 of x2.5 gain. So I've changed gain resistors to 10k to get x1.15 gain. And find it also too high. so voltage wise ODAC is giving more then I need.
    I am not sure if putting in-line resistors in ODAC->O2 connection is good idea?
    Is there any reasonable way to get gain less then x1?

    ReplyDelete
    Replies
    1. If you don't mind changing 4 resistors, you can increase R3 and R7 to around 2.2K, 3.3K or 4.7K and decrease R14 and R20 to 8.2K, 6.8K, or 4.7K respectively. That will reduce the O2's "gain" to around 0.75X, 0.67X or 0.5X respectively.

      And you should remove either R17/R21 or R19/R23 completely (use no resistor instead of the 10K resistors you're using now).

      Delete
  101. Hi NwAvGuy

    i have a qustion since its impossiable to find as transformers here could i just get WAU16-400 with a 220V to 100V step down converter, the mains here are standered 230-240 i could only find a 220V to 100V

    Ps i live in hong kong

    regards sid

    ReplyDelete
    Replies
    1. Yes, a 100 volt input would give about 13.3 volts output from the WAU16. That will work great. There are several European/Australian 220 volt wall transformers you could also use instead of the WAU16-400 and you wouldn't need a step down converter--like the Stontronics AC-1550EU, AC-1550BS, Maplin N57AT, N58AT, and Jaycar MP3020.

      Delete
    2. The Jaycar MP3020 is rated at 12V this seems a little close to the minimum?

      Delete
    3. It should be fine unless you have unusually low AC line voltage. Ideally about 14 volts is ideal and 15-16 volts is also OK but the O2's regulators will run a bit hotter. Because the audio circuitry in the O2 has such high power supply rejection even some noise on the power supply (such as the ripple you get if the AC voltage is too low) barely changes the audio performance in a measurable way and is still very likely inaudible.

      Delete
  102. It seems like my O2 has left and right channels flipped. Did this happen to anyone else?

    ReplyDelete
    Replies
    1. If your O2 is made with the official O2 PCB layout, and uses the PC board mounted jacks, it's not possible for it to flip the channels from the input jack to output jack. Tip of the input jack goes to tip of the output jack. There's nothing hand wired to get wrong.

      Delete
  103. Hello, I found the answer for the resistors i mentioned in the BOM, the ones that are 1/2 watt partnumbers instead of 1/8 watt...page 65 or so on the DIY forums Mr. Slim has a BOM that has the correct resistors that will fit etc..
    Thanks Alex.

    ReplyDelete
    Replies
    1. Those resistors can be anywhere from 1/8 watt to 1/2 watt. The only issue is they be metal film and physically fit the board. The Mil Spec "upgraded" resistors in the parts list have a lower power rating so they can meet much tighter specifications. AFAIK, JDS is supply valid parts.

      Delete
  104. Funny I got all the resistors in using the BOM and the 1/2 watt resistors fit in the board with no issues at all...so no need to order 1/8 for fear of not fitting on the board...lol...now have to wait for 3 backordered 40.2k's and the NJM2068D. These have been hard to find....Aug 20th due in at Mouser. Thanks for the response!!
    Alex

    ReplyDelete
  105. Ok my first O2 amp passed all the checks and voltages....and it works wonderfully. This little amp is so simple and easy to build...pay attention and be careful handling the parts, Q1, Q2 etc...now to figure out a case and maybe an integrated ODAC with RCA's and 1/4 phone jack.
    Thanks again for such a wonderfully honest design!

    All the best
    Alex

    ReplyDelete

PLEASE NOTE! If you see "LOAD MORE..." OR "LOADING..." just above this paragraph, you need to click that link to load the newest comments. It may take several seconds. When all the comments are displayed the Load More/Loading link disappears.

COMMENT APPROVAL: Comments may take anywhere from a few minutes to a few days to be approved. You might want to save a copy (highlight Ctrl-C) of your comments before submitting in case something goes wrong. The approval process is necessary to prevent obscenity, etc. Please keep comments on topic. Comments promoting commercial interests will not be approved and may be reported as SPAM.