For anyone interested in such things, 3 channel headphones amps (like the AMB M3, Mini3, B22, Tangent Pimeta, etc.) were something of a mystery to me. After testing a few, I looked more into the claimed benefits, history behind them, etc. and ended up writing an article:
3 Channel Headphone Amps & Virtual Grounds
If I'm missing something, or got something wrong, please let me know?
3 Channel Headphone Amps & Virtual Grounds
If I'm missing something, or got something wrong, please let me know?
These amps are only meant for driving headphones, which electrically, are isolated from the world. So it doesn't matter if the output ground is sitting at half the power supply voltage. In some of these amps (including the Mini3) the input "ground" is also sitting at half the battery voltage along with the entire metal chassis for that matter. But, again, it usually doesn't matter as it's either running from its isolated battery, or an isolated power supply.
But yes, if you were to power one of these amps from a power supply that had a grounded output, and connect the amp to pretty much anything else that was grounded, the virtual ground "channel" in the amp would be seriously unhappy and might meltdown as you suggest.
But yes, if you were to power one of these amps from a power supply that had a grounded output, and connect the amp to pretty much anything else that was grounded, the virtual ground "channel" in the amp would be seriously unhappy and might meltdown as you suggest.
Last edited:
some more discussion on both 'sides' here, for example: B22/Active Ground Query - Head-Fi.org Community
lets just say that a lot of people think that 3ch works; and yet, quite a few do question that topology. I don't think its been settled enough to convince the doubters; but I would not want to read one person's view and conclude that this represents everyone's views.
lets just say that a lot of people think that 3ch works; and yet, quite a few do question that topology. I don't think its been settled enough to convince the doubters; but I would not want to read one person's view and conclude that this represents everyone's views.
you can do both. I have a 3ch b22 that I use for both line-out and for phones-out. on phones-out, I get gnd from the 3rd gnd buffer or channel. on line out, though, I use input gnd as my output gnd. there's no 'load' to have to worry about 'driving' when you are a preamp; its only for loads like phones that you'd care about having a 3rd channel.
so, its not either/or. you pick the right thing for the right target device you are going to drive.
My position is objective, while much of what you're talking about subjective.
I'll gladly discuss any supposed factual, objective, and measurable advantages to 3 channel headphone amps. Even in the thread you reference, I'm not aware of any. Most of what's in that thread is either subjective (i.e. "sonic bricks"), misleading, or just plain wrong.
Perhaps... but saying things like a 3rd channel "lowers the output impedance" is like saying water flows up hill. So, to me at least, it's just plain wrong.
I'm not trying to be an "ultimate judge". I'm just applying well established engineering knowledge--like what gravity does with water. Ti Kan can insist gravity works differently all he wants, but that doesn't make it even a little bit true.
Speaking of 'just plain wrong' the argument given in a reply on your review of the AMB on your blog looks to me just that. You said :
The OPA690's specified THD is relatively high as it's not designed for audio use.
Looks to me to be fallacious reasoning to claim its distortion is high as a result of not being designed for audio. Plenty of opamps not designed for audio use have good specs in the audio band. OPA890 springs to mind with <-106dB at 20kHz and with a low impedance load.
The OPA690's specified THD is relatively high as it's not designed for audio use.
Looks to me to be fallacious reasoning to claim its distortion is high as a result of not being designed for audio. Plenty of opamps not designed for audio use have good specs in the audio band. OPA890 springs to mind with <-106dB at 20kHz and with a low impedance load.
what you claim is well established is debatable.
stop acting like a know-it-all. if you want to explore this topology more, that's fine; but it seems you have already made up your mind (and in a very short time, too; which makes me wonder if you have really thought about it from all angles).
the design is more than just from AMB. perhaps you should question the other 3ch topology supporters.
stop acting like a know-it-all. if you want to explore this topology more, that's fine; but it seems you have already made up your mind (and in a very short time, too; which makes me wonder if you have really thought about it from all angles).
the design is more than just from AMB. perhaps you should question the other 3ch topology supporters.
So debate it then - set out the reasons for, in plain engineering terms.
Here's a suggestion for you. Stop making ad hominem remarks. How does that sound? If you'd like a discussion about engineering by all means go ahead. When people make ad hominems its a sign they've run out of reasonable arguments. Always. No exceptions
Perceptual distortion, obviously.
Ok. The logic of the statement isn't ideal, but the datasheet lists the THD in the -60 to -70 dB region which is considerably worse than most audio op amps I'm aware of. The OPA690 is optimized more for video use. But I agree the two are not always mutually exclusive.
The specified THD of the OPA690 is not audio caliber and my measurements of the Mini3 clearly showed dramatically higher distortion when both channels were in use. It's very likely the higher distortion is due to the shared OPA690 even when it's within its specified current limits.
The main issue here isn't the OPA690 but the general concept of a third channel somehow achieving all the things AMB and others claim:
- Lower Output Impedance
- Lower Distortion
- Lower Crosstalk
the OP has set out to attack the 3ch topology. of course its his right.
but he has concluded, entirely on his own, that its not a valid way to design or run an amp. he says it with authority like there is only one valid conclusion to come to.
when you escalate your viewpoint to that of fact - when its not so clear its a fact - its arrogant.
the OP needs to understand this subject a bit more, I think, before just concluding its invalid. again, its fine to present his views but he should make it clear that its his view and not some industry concensus.
its all in the attitude. its fine to debate the subject but his insistence that his conclusion is 'absolutely correct' is just annoying (amongst other things).
but he has concluded, entirely on his own, that its not a valid way to design or run an amp. he says it with authority like there is only one valid conclusion to come to.
when you escalate your viewpoint to that of fact - when its not so clear its a fact - its arrogant.
the OP needs to understand this subject a bit more, I think, before just concluding its invalid. again, its fine to present his views but he should make it clear that its his view and not some industry concensus.
its all in the attitude. its fine to debate the subject but his insistence that his conclusion is 'absolutely correct' is just annoying (amongst other things).
OK, so help me understand where I'm wrong? Please?
I've provided a nearly 6000 word article, with diagrams, references, test results, some math, etc. What in all that indicates I'm wrong and need to think about it more?
Because its being offered to the video market, the distortion specs are not given over the audio band. As far as I noticed, all the graph left-hand axes on distortion stop at 100kHz - well above audio frequencies. At 100kHz the part doesn't look quite as bad as you're claiming - around -80dB or so.
Here's a handwaving argument that helps to estimate audio-band distortion when its not shown. Go to the open loop gain plot and see where the dominant pole is (fig24). On the OPA690 this looks to be about 30kHz. Above that the gain falls at 6dB per octave. So we could expect a slightly better figure at 20kHz then that shown at 100kHz due to there being more loop gain available.
Yes, its rather obscure to me that an opamp like this would be used as a virtual ground. Particularly because opamps are normally designed to put out high currents when their outputs are close to the rails. In this case the output stays resolutely at 0V meaning high dissipation is the result. Very inefficient use of battery power.
Yes, kudos to you for exposing the BS in this.
Excuse me, how is saying 'stop acting like a know-it-all' not escalating a viewpoint to that of fact? How is that not arrogant?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.