AMB Laboratories DIY Audio

[NwAvGuy]

This is, in my opinion a good post. I'm not knowledgeable enough (yet) to know if it's right, but it's well worded, and easy to understand.

Thanks. I'm trying

Here's what I think: You should design a small, 2-battery, 2-channel "portable"-size amp like the one you describe.

I might, you never know But, honestly, I measure products like the FiiO E5 and E7 and think "why bother?". You can't buy the front panel of the Mini3 for the price of the FiiO E5. And I can't do anything about the price of low volume front panels. Yes, I could certainly design something that outperforms both of them, but would it sufficiently outperform the $120 E9? I don't know... It's easy to beat China Inc when they get it obviously wrong, but when they get it mostly right, it's very hard to beat them unless cost is relatively unimportant (and for most of us that's rarely true).

When you say you'd spend the money saved on having only 2 channels on a second battery, do you mean you'd have one for +, one for - and the virtual ground inbetween? Since the batteries will discharge separately, won't you end up with poor long term performance (if one battery is at, say, 9V and the other at 8V)

See my last post. With two batteries, it's not a "virtual ground" it's a REAL ground. No active distortion inducing, current limiting, battery munching, non-linear (the OPA690 is all these things) parts required.

As far as "what ground channel amp would you use" question, it's not meant to be a loaded question, but I do see your point. The "trick answer" is that you wouldn't use a 3-channel design. But let's say there's a huge market for these newfangled 3-channel amplifiers

I really haven't thought about it. It's like asking a Porsche owner which chrome exhaust tips he'd buy at Walmart for his new 911. He'd probably reply with something like "None because I'm not into fake chrome exhaust tips". He'd buy a real exhaust system. Real is generally best. Edit: I'm still waiting for someone to explain if there's some sort of benefit to 3 channel designs I've missed?

Hmm, if you leave the ground channel unpopulated (like what the balanced (2-board) mini^3 people do) would it perform better? I feel like that's easy to test.

Hmmm... I'd have to think about running the Mini3 bridged if that's what you're suggesting? You would need either a balanced source or modifications to invert the signal. Bridged amps, in some ways, suffer from some of the same problems as fake grounds. The load "sees" two amplifiers instead of one. That generally means more distortion, degraded specs, etc. It would roughly double the maximum output swing much like 2 batteries and you could get rid of the dreaded OPA690 and fake ground. But it's still not as clean of a solution as 2 batteries.

[DorkVader]

Hmm, if you leave the ground channel unpopulated (like what the balanced (2-board) mini^3 people do) would it perform better? I feel like that's easy to test.

Hmmm... I'd have to think about running the Mini3 bridged if that's what you're suggesting? You would need either a balanced source or modifications to invert the signal. Bridged amps, in some ways, suffer from some of the same problems as fake grounds. The load "sees" two amplifiers instead of one. That generally means more distortion, degraded specs, etc. It would roughly double the maximum output swing much like 2 batteries and you could get rid of the dreaded OPA690 and fake ground. But it's still not as clean of a solution as 2 batteries.

No, I mean leaving it completely unpopulated, and running the return wire directly into the virtual ground like this: http://i51.tinypic.com/344sg88.png

[amb]

No, I mean leaving it completely unpopulated, and running the return wire directly into the virtual ground like this: http://i51.tinypic.com/344sg88.png

No. The TLE2426 could only source or sink 20mA. The 100uF caps help a bit with the transients, but you can't count on the TLE2426 to sustain it for long.

[agdr]

I thought using dual batteries like that creates dangerous DC offset due to the batteries draining unevenly.

That is why Tangent says that he gave up on using dual batteries and switched to virtual grounds:

http://tangentsoft.net/elec/vgrounds.html

With two batteries you would probably want want to provide some protection against failure modes such as one battery terminal becoming loose and momentarily leaving it powered with just one rail. Something like a solid state relay that would insure the whole thing goes down if just one rail fails.

Dual DC-to-DC converters are another way to do it, like this fellow did:

http://www.diyaudio.com/forums/headphones/187501-nelson-pass-inspired-headphone-amp.html

[nertia]

for some reason, I was reminded of this xkcd comic today

lol, reminded me of a fairy tale...


goatseye

[bcg27]

3 channel topology is popular in headphone amps for a reason.

Perhaps I'm missing something, but I think it's "popular" for one (or more) of 3 reasons:

1 - It's useful in single battery designs to avoid an output cap or DC-DC/charge pump power supply.

2 - It no doubt offers better performance for people who don't understand how to route ground signals.

3 - It's different, like sonic bricks, and therefore some think it must be better.

If someone here is aware of other benefits, based on sound engineering principals, I'd love to hear them?

As a side note, I'm aware that some large pro sound amps (like you'd find splitting eardrums at rock concerts) use various forms of "active grounds". Crown calls theirs Grounded Bridge. These amps usually put out thousands of watts and the topology is used to improve the power supply efficiency which is critical when you have racks full of high power amps drawing hundreds of amps of line current. None of those things apply to a headphone amp.

There is a thread over at head-fi where somebody else was asking questions about the value of the active ground and amb produced simulation results that show a significant amount of ripple being injected into the power supply rails from a waveform, possibly a 100 kHz square wave? I don't quite remember the details. The simulation from the active ground showed no such ripple. I tried to find the thread last night but was not able to. I will try to spend a little more time searching for it tonight and will post the link here if I find it. I believe that the active ground also doubles the slew rate of the amplifier.


PS: xkcd is an awesome comic!

[DorkVader]

No, I mean leaving it completely unpopulated, and running the return wire directly into the virtual ground like this: http://i51.tinypic.com/344sg88.png

No. The TLE2426 could only source or sink 20mA. The 100uF caps help a bit with the transients, but you can't count on the TLE2426 to sustain it for long.

Ohright, I see. Thanks for the explanation.

[NwAvGuy]

There is a thread over at head-fi where somebody else was asking questions about the value of the active ground and amb produced simulation results that show a significant amount of ripple being injected into the power supply rails from a waveform, possibly a 100 kHz square wave? I don't quite remember the details. The simulation from the active ground showed no such ripple. I tried to find the thread last night but was not able to. I will try to spend a little more time searching for it tonight and will post the link here if I find it. I believe that the active ground also doubles the slew rate of the amplifier.

Thanks. If you find the thread, please let me know? The slew rate is a non-issue. Nearly all headphone amps can already slew faster than any signal you can possibly feed them. And more slew rate than needed doesn't help anything. Slew rate is much more an issue in big power amplifiers that can slew 80 volts or more. A headphone amp only has to manage a few volts over the same time period.

The supply rail ripple issue is likely related to what I mentioned about large power amps using active grounds. When you're trying produce killowatts and sending 20+ amps to each speaker the power supply ripple becomes much harder to deal with. In a headphone amp, however, the peak currents are only a few tenths of an amp. And it's easy to keep the supply clean with such low currents.

I suspect this is a classic audiophile case of trying to fix something that's not broken. Sometimes the fixes just cost money but are otherwise harmless. In this case (headphone amps), it seems the cure is worse than the alleged "problem". But, as I said, perhaps I'm (edit typo) missing something?

[egallego]

Thanks for your reply:

2 - It no doubt offers better performance for people who don't understand how to route ground signals.

IMHO that's the main reason, 3 channel configurations are believe to have better noise performance. I'm afraid I cannot tell if that's really true, but active ground being widely in use means a lot of different designers (diy and professional) agreed.

As a side note, I'm aware that some large pro sound amps (like you'd find splitting eardrums at rock concerts) use various forms of "active grounds". Crown calls theirs Grounded Bridge. These amps usually put out thousands of watts and the topology is used to improve the power supply efficiency which is critical when you have racks full of high power amps drawing hundreds of amps of line current. None of those things apply to a headphone amp.

What's your opinion on balanced headphones and amps? (Of course assuming you've got a balanced source)

[NwAvGuy]

but active ground being widely in use means a lot of different designers (diy and professional)

I'm only aware of very few commercial 3 channel designs. Perhaps I'm missing some?

What's your opinion on balanced headphones and amps? (Of course assuming you've got a balanced source)

In theory, at least, I can see the appeal of a fully balanced signal path with a 4 wire connection to the headphones. But, in reality, just the 4 wire connection to a conventional unbalanced amp might provide even better performance with much less cost and complexity.

Balanced audio is another really misunderstood area by many audiophiles and the topic is already on my list for a future blog article. It's another classic case of taking something with some truth behind it, and applying it in ways that don't always make sense. Here's the condensed version of part of the article:

The word "balanced" sounds nice and conjures up notions of symmetry so the intuitive sense is it must be better. But the dirty truth is a fully balanced signal path will usually be more noisy (in terms of the electronics--not external noise) than an unbalanced one. A standard op amp input with a 10K impedance implemented in balanced mode is a whopping 14 dB noisier than the same op amp and 10K impedance in an unbalanced design. That's rather significant--especially in the headphone world where noise tends to be more audible.

The main advantage to balanced audio is much higher noise immunity when running long lengths of low level interconnects as done for recording, studio work, live sound reinforcement, etc. The interconnects are far more noise immune than unbalanced "RCA-type" cables. They also allow bundling many signals together with minimal crosstalk. But in home audio RCA cables work surprisingly well and rarely contribute any significant noise (or any other kind of distortion) to the system (phono cartridge level signals being the notable exception). Audiophile RCA cables are usually single mono cables and crosstalk isn't an issue.

A secondary advantage is helping to avoid ground loops. Balanced interconnects usually don't share the signal ground with the chassis ground helping avoid hum issues. But again, this typically isn't a problem in home audio because most home audio gear "floats" and has 2 prong AC power cords (versus pro gear that's usually grounded).

Balanced connectors are more robust than RCA connectors. But, again, that's usually not a big problem for home gear which isn't being re-configured all the time by stage monkeys.

Some claim advantages to balanced terminations claiming fewer cable reflections, better signal integrity, etc. But these things don't even start to become significant until well above the audio band. Proper termination is much more of an issue for data and RF signals. And the impedances usually involved with audio inputs are not matched to cable impedance anyway--even with balanced inputs. Again, there's a grain of truth but it gets misapplied to audio.

And just how much is really "balanced"? Even really high-end "balanced" gear with four and five figure price tags can contain unbalanced circuitry. The volume control is a common area where otherwise balanced signal paths suddenly become unbalanced. So visions of an "entirely balanced signal path" are often not realized even with gear that has balanced inputs and outputs.

When you add it all up, there's usually little tangible benefit for home audio use. The notable exception would be if you do have to make long runs with line level signals--say from a preamp on one side of the room to active speakers/subwoofers, power amps, etc. located on the other side of the room. But that doesn't apply to headphone listening.

I do, however, see the appeal in 4 wire headphones so you're not sharing the ground connection between the two channels. I'm not sure how audible the difference is, but unlike a lot of audiophile "fixes", it certainly can't hurt.