O2: High Performance Portable/Home Amp

[Marvin the Martian]

I just found that thread, and posted a in it just to subscribe.....should be interesting to follow.

[CrowdAds]

[NwAvGuy]

Thanks Marvin, that's ideally the sort of product I'd like to find if it measures well. It's entirely possible to build a product like that, or the Xonar U3, etc. with decent performance than can sell for under $50. I just have to hopefully find such a product

[Soaa]

I'm actually really curious about the Xonar U3 now. Hurry with the measurements!

[mikeaj]

re: noise (please correct me if I say something stupid; it's 3am and maybe I ought to get to bed)

I think in many practical scenarios, the O2 or another external amp can effectively reduce the noise of your playback system.

Noise is mostly a concern when you've got one of those 125 dB SPL / 1V or higher IEMs. Typically out of a laptop or other potentially noisy source, you would have the volume pretty low for such sensitive IEMs, like maybe 10 units of volume out of 100 if not lower, whatever that means. It depends on how they map those volume units into output voltage levels. If you turn the volume up all the way, this should increase the signal level without changing the noise much, depending on what's causing the noise. Thus the SNR is increased by turning the volume up.

So turn the volume up, plug the output into an amp, and turn the volume control on the amp considerably down, and you can get back to to the same signal level corresponding to what was 10 units out of 100 in the example. Since the entire signal+noise from the source has been attenuated by the amp, the noise level is lower than it was originally, while the signal level is kept the same.

Right? It's been a while since I've used sensitive IEMs though, so I forget how the noise level of noisy sources scales with the output level.

[NwAvGuy]

mikeaj, you bring up a good point and I should revise my article on headphone amps to include what you suggest. If you always use the volume control on the amp and leave the volume up on the PC, there could be some improvement in certain circumstances. As you say, it depends on the origin of the noise in the signal chain in the PC.

But, just to be clear for everyone, an amp can't change the actual Signal to Noise ratio for a given PC volume setting. If your headphones are not too loud with the volume all the way up on the PC, and you still hear noise, an amp isn't going to help the problem at all. With some of my laptops that's exactly the situation using typical dynamic (not Balanced Armature ultra sensitive) headphones.

If you're using B.A. IEMs, and can't get the volume past say 50% without it being too loud, then using a higher volume setting and an amp could help the noise situation. But it will only help if the majority of the noise is relatively constant with different PC volume settings. You also, ideally, would want an amp configured for 1X gain otherwise you'll be stuck using a vary small part of the volume control's range and will likely have some channel balance tracking issues. The O2 can be set for 1X gain just by clipping two resistors.

In general, for PC users wanting an O2 anyway, I would suggesting starting with just the amp. If you're happy with the sound direct from your PC life is good. If you're not, then consider adding a USB DAC.

[Soaa]

Would it be desirable to implement variable gain using a pot?

[mikeaj]

Originally Posted by Soaa

Would it be desirable to implement variable gain using a pot?

There is variable gain using a pot, apparently implemented in a better way than most--see design docs. There's just also two fixed baseline gains that can be switched between with a switch on the front. Those fixed gains can be set to whatever sane value by the builder, by choosing appropriate resistor values. Nominally the "default" values are 3x and 7x.

The total gain can thus be adjusted by a combination of the gain switch and volume pot, giving a wide range of operation.

[Soaa]

What I meant was, rather a gain switch with fixed gain values, what about two knobs, one for gain and one for volume? That way, gain could be turned down to 1x for the most sensitive headphones (not sure if there would be an advantage in amping but who knows) up to 7x for example, for less sensitive full size cans.

[NwAvGuy]

Originally Posted by Soaa

What I meant was, rather a gain switch with fixed gain values, what about two knobs, one for gain and one for volume? That way, gain could be turned down to 1x for the most sensitive headphones (not sure if there would be an advantage in amping but who knows) up to 7x for example, for less sensitive full size cans.

It would have some obvious benefits but what you suggest is generally considered bad practice for a variety of reasons. The feedback loop of a gain stage (which sets the gain) is the most critical part of the amplifier. It's very sensitive to noise issues, stability problems, what are known as parasitics, etc. And trying to route those signals through a dual (stereo) pot would be a significant compromise. It would also create additional channel balance errors.

There are what's known as Voltage Controlled Amplifiers (VCAs) that do allow you to continuously adjust their gain. But they add significant complexity and generally still don't perform as well as a fixed gain stage as used in the O2.

[Marvin the Martian]

Thought I should post this snip of my Head-Fi post, before it gets edited by anyone.

[NwAvGuy]

Thanks for sharing that Marvin. If Tyll Hertsens tests an O2, as he's offered to do, that should go some ways towards silencing those sorts of criticisms--at least if it measures reasonably well in his testing. Anyone who wants to claim we're both faking the numbers with zero evidence to the contrary is waaaaaay out on a limb. Tyll, if anything, should have the opposite alleged "agenda" (given the advertisers on InnerFidelity).

[mikeaj]

Maybe you should just run it through RMAA (with all the usual caveats) so people have a "reference" figure. Not that I think it's more meaningful than what's written in the first article, but RMAA shouldn't take long.

If other people using RMAA with a cheap interface get pretty much the same results with their line out -> line in, as compared to line out -> O2 (loaded) -> line in, that should be fairly convincing.

[Marvin the Martian]

Originally Posted by NwAvGuy

Thanks for sharing that Marvin. If Tyll Hertsens tests an O2, as he's offered to do, that should go some ways towards silencing those sorts of criticisms--at least if it measures reasonably well in his testing. Anyone who wants to claim we're both faking the numbers with zero evidence to the contrary is waaaaaay out on a limb. Tyll, if anything, should have the opposite alleged "agenda" (given the advertisers on InnerFidelity).

Well, the response to that other post was not what I expected, to be honest....when the "vested interest" comment was made, I immediately thought they were referring to financial aspects. Here is the reply.

[NwAvGuy]

Mikeaj, I thought about that and might yet do it. But it does take a while to do correctly. Because RMAA has no concept of actual levels you have to manually set the levels with other instrumentation, run the tests at the right loads, etc. I also use my Benchmark DAC1 and ADC1 with RMAA and they have to be moved around and re-configured for RMAA duty. And I often end up having to re-run the tests multiple times to finally get RMAA to behave and spit out entirely valid data.

And what you end up with is still difficult for someone else to verify as they probably don't have proper (or at least the same) test loads, they often have no way to properly set the absolute levels, and they're using completely different sound hardware that's very likely far worse than the O2 in many respects so the resulting report is more a reflection of their sound hardware rather than the O2. That's even true for me. The O2, for example, has lower noise than even the Benchmark DAC1 (because of all the DAC1's nearby digital circuitry).

[NwAvGuy]

Originally Posted by Marvin the Martian

Well, the response to that other post was not what I expected, to be honest....when the "vested interest" comment was made, I immediately thought they were referring to financial aspects. Here is the reply.

Their response makes sense and I would agree there's some truth in the reply. Obviously there's a certain amount of pride in a project like this. I don't see that as a bad thing and it's certainly not hidden.

But numbers are numbers. All the pride in the world doesn't cause 1+1 to equal 3. So I'd have to be rather foolish to stick my neck out with this highly objective design and base it on fake performance.

[Satellite_6]

You know what, my integrated sound is pretty darn good plugged into my M-Stage. It sounded so unimpressive straight out of the headphone jack I never bothered trying it before. The MSII is really only a minor improvement. Shows what I know. I suspect I may feel really stupid when I try your amp connected straight to my laptop. We'll see.

[saratoga]

Originally Posted by Marvin the Martian

Well, the response to that other post was not what I expected, to be honest....when the "vested interest" comment was made, I immediately thought they were referring to financial aspects. Here is the reply.

Vested as in investment, so a "vested interest" is literally a financial investment.

I think he probably figured the O2 was for sale or somehow being monetized when he wrote that. Otherwise his comment doesn't really make sense, since hes basically saying that it "sucks" that someone else is designing him good hardware for free.

Well that or if we take him literally he'd prefer that people try to suck at things around him. But probably not.

[coacharnold]

Ok ... I may be missing the boat .... can someone explain ..... is there a possibility that this will include a usb DAC ??? if not can we do that ???

regardless ....how soon the PCBs ???


tim

[Satellite_6]

Originally Posted by coacharnold

Ok ... I may be missing the boat .... can someone explain ..... is there a possibility that this will include a usb DAC ??? if not can we do that ???

regardless ....how soon the PCBs ???


tim

Yes, but not atm.

[Sorensiim]

I finished reading the 3 articles on your blog yesterday and I'm thoroughly impressed. Can't wait for the ready-to-solder kits to show up. My brother is a carpenter and I might just have to get him to build me a custom cherry wood enclosure for my O2...

And by the way: KUDOS for having the balls to call "bullshit" on Head-Fi and the numbers to back it up. Reminds me of the saying "Why offend people with style when you can offend them with substance?" The article about getting banned from Head-Fi also made me lose all respect for Schiit - I didn't have much respect for Head-fi to begin with.