Review: NwAvGuy's O2 DIY Amplifier - Page 3  

^Indeed.

Very nice review! Sounds interesting. Any chance of a few photos including back panel?

Looking forward to seeing the finished product!

Shike, have you used the O2 with a mixamp? I'm planning on using it with my Q701s on the 360.

Very good review! Are you still going to publish photos?

We want to see your build.

Just take a tall glass from your kitchen and rest the camera on it.  Use the macro setting and you're good to go.  All you need for lighting is regular room light, or put the board on your laptop keyboard and use the light from the screen.  Set the camera for florescent lighting, macro setting, rest on tall glass if you need it, but I found that the laptop light was enough to take very sharp pics hand held...

Here's one I took for knife-Fi with the light from my laptop screen (click to enlarge)

concluding your review, are you saying that o2 is better sounding than Schiit Asgard and Audio-GD Sparrow, shike? if it is, that's pretty interesting..

Both the Asgard and AudioGd products are cases of "Buzzword Design", IMO. Suffice to say that being fully discrete or single-ended has precious little to do with audio quality, so I wouldn't be surprised if they were actually audibly inferior. The only comprehensive measurements I've ever seen for AudioGd products (their discrete opamps) were abominable.

/mild rant alert

What I mean is that when I head on over to the Schitt/AudioGd sites and read the descriptions/FAQs for their products, I become incredibly sceptical of their offerings - mainly where they follow design practises which seem hugely counter-intuitive and justify them incredibly flimsily. For example, to quote the page for the AudioGd Sparrow:

"The Sparrow uses ACSS (Audio-gd Current Signal System, which is a current signal transmission system). ACSS is a non-feedback technology make by Discrete amp. Most people had know the global feedback can offer better specs for the meters, non-feedback can't please the meters but can offer better sound for human's ears . Here is a conflict of the class circuits. But the ACSS arrive a new field, it can offer least coloration sound which is more neutral with very low distortion and high linear . So it can recur the dynamic, detail and neutral sound but not bright or harsh."

This is pandering to audiophile mythology, not solid design practises. Sure, it's possible to get acceptable results whilst avoiding global feedback, but why would a designer begin their design process by handicapping themselves? Perhaps they're referring to feedback used incorrectly in some way? But no, they admit it will bugger up their measurements.

It gets worse - they go on to paint the absence of a voltage switch as some ludicrous crusade to prevent audio degradation and to ensure that their users can hear the effects of upgrading their power cables. At this stage I'm feeling more than a little cynical regarding their marketing literature. Regrettably, they're not done - it's "fully discrete."

Now, fully discrete is all very well in power amps, but in headphone amps it doesn't make any sense. Firstly, it is monstrously expensive to do well, both in R&D and in sheer parts cost. Even then, you're going to have a phenomenally hard time equalling the performance of a cheap little IC opamp. Trying to paint this as "not using cost cutting measures" is utter crap. The only examples of AudioGd's attempts at discrete opamps I've seen measured terribly - doesn't fill me with confidence regarding their designer/s.

This is finally exacerbated by how they describe their products. The one thing that puts me off a company's product range is when they assure me the sound isn't "digital."

Call me paranoid, but I get the feeling that these design decisions are not justified by some incredible crusade for pure audio. This is, as I put it earlier, "designing by buzzwords." Audiophiles have various irrational dislikes of things such as IC opamps and get all excited when somebody shouts "Fully Discrete!" at them. Thus, people don't design for low distortion, or maximum neutrality, or anything like that. They design to pander to audiophile mythology and do so at such a price point and so utterly that you become incredibly wary as to the quality of their products.

Let's move on to Schiit. In some ways (such as basic literacy - I know AudioGd are Chinese, but other small companies pull off rather better translations) they are superior, but they still ultimately annoy me in much the same way. The very first line on the Asgard's web page reels out its fully discrete credentials and follows up with the line "No opamps, no shortcuts" - This is, again, pandering to audiophile mythology, so people on Head-Fi can go "Ooh, fully discrete at such a low price!"

I was recently further annoyed by the publicity spiel for their new DAC:

"The ugly truth about most DACs in this price range is that they sacrifice every single one of your original music samples to get their magic "192kHz" spec"

This is scaremongering crap, playing to audiophile paranoia. The Benchmark DAC1, for example, resamples to 110khz and I'll wager it beats the crap out the Bifrost delivering - god forbid - what's on the recording! With minimal variance! 

Compared to the "Fully Discrete" BS however, it's small fries.

And don't get me started on Burson. Their "The Naked Truth About Opamps" article single-handedly dissuades me from ever touching any of their products with a barge pole - I could make up something better myself. Seriously, raving on about "The size of opamps compared to discrete circuits causes forms of quantum degradation which are only clearly apparent in listening tests." would have been more plausible.

I realise this is reading like an incredibly ill-natured rant against some of the most popular products on Head-Fi. But when a company sells its products to irrational audiophile criteria...on the one hand, I can't blame them. Nobody wants to hear about how little distortion your amplifier has when you can rave on about how it uses no evil "short-cut" opamps - even if thanks to that 0.1% THD is the best you can manage on your own specs - that's not Hi-Fi by the standards of a SS power amp in the 1980s, let alone a modern headphone amp. When I read the list of reasons why I should purchase such company's products, I don't expect them to reel off the number of ways they handicapped their designs.

I understand that it is entirely possible to make an OK product despite such self-imposed stumbling blocks - but when half their publicity literature is about claiming that they are all plus points... Meanwhile, I can't really begrudge companies playing a little to their audience. But likely screwing over their product so the marketing department has something to say? People will say "Oh, but this distortion is inaudible," - but how can all these audiophiles claim to care about the supposedly subtle and unmeasurable when they don't seem to care about the basic, measurable aspects of a product that determines that it is competent?

Good grief, this is getting more rant-y by the second. The really infuriating part is that if you were ever to post something like this in its own thread, I can predict the first 5 responses:

1: OMG, not another measurement wh***, bugger off.

2: I can hear how opamps corrupt the music - you must be deaf.

3. I think designer "X" knows a little more than you do (delivered condescendingly)

4. Fully discrete is always better because opamps are cheap.

5. Well I like how they sound! Some people, always stirring things up...

So the best I can say is that companies who design primarily to buzzwords will never, ever get any of my money. It's a pity, because designing in this way is so common you have to pay hugely over the odds simply to ensure that the designer didn't decide he knew more about opamp design than the entire R&D department of Texas Instruments.

I think Willakan has a sensible point about discrete vs opamp.  The more components in a design, the more variance that can occur between assemblies of the same design.  With discrete there are collections of transistors, caps, resistors, etc from different vendors. Of course, they are spec'd within a tolerance but there is variance.  So it is probable that each amp rolling off the assembly line has a unique signature. On the other hand, opamps are manufactured from monolithic blocks, as far as I know, so I expect they will be self-consistent.

Opamps and buffers from National and others claim incredible low distortion. They provide reference boards (too expensive for DIY) which others vendors can test for quality.

IMO, discrete designs without blocking capacitors are too difficult and risky for DIY. One mistake, and poof goes the expensive headphones.  Stick with opamps.