The Objective2 (O2) Headphone Amp DIY Project

Well, I want to use it on my computers' Asus Xonar HDAV outputs.
Because the HTPC has a BluRay drive and the Asus Soundcard has all the necessary licenses and codecs for BluRay playback.
And where I would otherwise have to take the analog signals from the outputs, the Buffalo III allows to grab the I2S signals from the card and produce nice and clean sound from it.
If there were any other DACs capable of this, those would be an option.
I just haven't found any.

And the Buffalo III has (with accessories) the ability to process all the channels at once.
When I only need one Buffalo III where I'd need 3 (5.1 = 3x stereo) other DACs...
I just don't see too much competition.

But we're drifting deeply into the off-topic :)
 
I spent an hour reviewing the nvnavguy website, I am changing my tune, that dscope is amazing and this is the first headphone amp I have every seen developed from groundup with a such a great tool, the documentation is incredile, we actually know where the amp starts clipping with a giving power into different loads.

I have one stupid question. As far as scaling up from a semi-portable to a higher power amp, would it be possible to run the tests for thd vs voltage/load for 15V rails?

The "high much power do we need" section on the website actually rules out a few popular headphones for the 110 dB peak recomendations. It would seem the most elelegant solution to the issue would be to use two boards full balanced, I think this would make it an amp that really would match/beat any other SS design for a fraction of the cost. It would basically increase the power between 2x to 4x (depending on the phones impedane/sensitiviy) and propably handle any dynamic headphone made.
 
The "high much power do we need" section on the website actually rules out a few popular headphones for the 110 dB peak recomendations. It would seem the most elelegant solution to the issue would be to use two boards full balanced, I think this would make it an amp that really would match/beat any other SS design for a fraction of the cost. It would basically increase the power between 2x to 4x (depending on the phones impedane/sensitiviy) and propably handle any dynamic headphone made.
With the usual downsides of requiring cans rewired for balanced driving and an additional inverter stage. Plus minimum sensible load impedance is doubled (so it would mainly make sense for some of these 600 ohm offenders, not at all for orthos).
 
I spent an hour reviewing the nvnavguy website, I am changing my tune, that dscope is amazing and this is the first headphone amp I have every seen developed from groundup with a such a great tool, the documentation is incredile, we actually know where the amp starts clipping with a giving power into different loads.

Welcome to the Dark Side.. :D

I have one stupid question. As far as scaling up from a semi-portable to a higher power amp, would it be possible to run the tests for thd vs voltage/load for 15V rails?

RS has mentioned before that 15V rails might be a bit too much for the 4556's and that he was still planning on using 12 V for the real desktop version.

The "high much power do we need" section on the website actually rules out a few popular headphones for the 110 dB peak recomendations. It would seem the most elelegant solution to the issue would be to use two boards full balanced, I think this would make it an amp that really would match/beat any other SS design for a fraction of the cost. It would basically increase the power between 2x to 4x (depending on the phones impedane/sensitiviy) and propably handle any dynamic headphone made.

In a balanced configuration, is there any extra current draw on any individual channel, as there is in a bridged configuration? I can understand the 2x power, since you are driving the transducers in a positive and a negative direction with a pair of amplifiers, but how does it get to 4X?
Also, what is the real gain in sound quality from a balanced config? I know this has been discussed at adnausum elsewhere, but given the low cost of the O2, the cost of comparing unbalanced to balanced would be quite low.

Hmm.. you've got me thinking, but I guess RS would have to weigh in on whether it would work, but I wonder if 2 boards would fit in a B3 case.. it might if you left off the batteries, since they're the tallest components on the board..
There would be the issue of how to handle the volume control for 2 channels..
 
Last edited:
@ jokener, the ackodac AKD18P is another 8 channel sabre 9018 that has actually been out for longer but is lesser known, with much more comprehensive control (ipad, pc, wifi remote control of ALL registers), input options (hires usb, coax, proper aes not fake, hdmi) and signal routing than the b3; impedance controlled pcbs and more appropriate digital interconnections (all are on smd mini bnc, actually designed for signals in that range, unlike pinheaders). not a simple or cheap build though.

@ mrSlim, yeah thats a funny one. people seem to think its automatic, my understanding is yes there is the potential to x4 due to the impedance seen by the output being halved and thus draws more current, but that also means you have a lot tougher time to put as much voltage across it and first the power supply must be capable of it. I prefer 'a balanced/bridged amp will put double the power into double the impedance'

btw mr Slim, ive been asked by a member to build an amp for them, can you estimate the time for an experienced builder to stuff the boards or anyone that has for that matter? I have said i wont do casework, although with all things just about being pcb mounted and pre milled panels there isnt much to it. it looks a very straight forward build, but looks can be deceiving. i dont plan to do any others and i'm not even sure if i'll do this one yet.
 
Last edited:
qusp:
RS told me that it takes him about an hour to build one (board stuffing really since he didn't do face plates etc), so I've been planning on being able to do it in about that much time. Of course, that doesn't account for all the parts ordering, etc and I'm doing some "baby" desktops too, which are probably going to be the fiddliest part of the whole thing since I'll be doing RCA's, 1/4 inch jack and the volume pot on flying leads off the board..
 
thanks,yeah thats about what i figured, although i'm not looking at getting much practice, i spend enough time sniffing solder fumes as it is. yeah i'll probably be doing a baby desktop myself with one of the pcbs i got as a present for a friend, he doesnt need a portable amp, certainly not one as ermm substantial as this one, but will love one to connect to his mixing decks, i plan to use the usb output on his mixer and add an opus and simple instrumentation amp for bal-se, perhaps even some onetics line transformers
 
thanks,yeah thats about what i figured, although i'm not looking at getting much practice, i spend enough time sniffing solder fumes as it is. yeah i'll probably be doing a baby desktop myself with one of the pcbs i got as a present for a friend, he doesnt need a portable amp, certainly not one as ermm substantial as this one, but will love one to connect to his mixing decks, i plan to use the usb output on his mixer and add an opus and simple instrumentation amp for bal-se, perhaps even some onetics line transformers

heh, I know what you mean about solder fumes.. With the number of O2's I'm building, I've been debating building up some kind of DIY fume extractor.. (saw one on MAKE I think)
 
First link on google search for "solder fume extractor" :cool:

Solder Fume Extractor

yeah.. I saw that, and did a little more research.. from what I understand, to do it right, you need a Hepa filter to filter out the particles of flux that are actual irritants.. the Carbon filters don't do anything.. Just a fan to blow the fumes away is probably just as useful..
 
The "high much power do we need" section on the website actually rules out a few popular headphones for the 110 dB peak recomendations. It would seem the most elelegant solution to the issue would be to use two boards full balanced, I think this would make it an amp that really would match/beat any other SS design for a fraction of the cost. It would basically increase the power between 2x to 4x (depending on the phones impedane/sensitiviy) and propably handle any dynamic headphone made.

For the very rare headphones that really need more than 20 volts peak-to-peak (7 Vrms) I would suggest a different amp than the O2. 99% of headphones I know of can hit 110+ dB SPL with the O2. If you're the 1% looking for hearing damage or own some really rare uber power hungry headphones, I would suggest looking into other amps. I cannot recommend the O2 for higher than a 24 volt (+/- 12V) supply or in balanced/bridged mode.
 
For the very rare headphones that really need more than 20 volts peak-to-peak (7 Vrms) I would suggest a different amp than the O2. 99% of headphones I know of can hit 110+ dB SPL with the O2. If you're the 1% looking for hearing damage or own some really rare uber power hungry headphones, I would suggest looking into other amps. I cannot recommend the O2 for higher than a 24 volt (+/- 12V) supply or in balanced/bridged mode.

Hey I bought the boards not so quick to give up on it yet :eek:.

I am definately not looking for hearing damage but found your chart on "more Power" sept blog interesting and have given it a lot of thought.

Ok maybe I am not interpreting you chart for power requirements and recommendations correctly. You recommend 115 dB peak for listening to high dynamic range classical music. Very popular headphone: the Beyer DT880 would be "borderline" with this amp as 7vrms*7/600= 81 mW. Also the AKG Sextettes, Also the AKG 240-DF's, the AKG340's, many of the orthos[help=]%[/help]

Maybe I am reading the chart wrong instead of using the 7vrms, would it make more sense to look at it from Vpeak to peak so instead of Prms on chart those are Pp-p? Or mabey there is just very very little music that requires these 115db peaks?

Not trying to stir controversy, just thought the chart was very interesting and deserved a little discussion as it applies to the )2.

thanks
 
I believe you're looking at it right. I asked about this directly to rocketscientist a few weeks ago, and I suggested running in bridged mono mode with 2 boards in order to get twice the voltage swing as I am running K340s myself to get to the 115dB maximum in case I ever decided to need it (I'd have to be listening pretty loud compared to what I'm used to). He noted that it would be possible, but with certain side effects (inability to drive lower impedance headphones, increase in output resistance? - who cares you're driving high impedance headphones anyways. I think he also mentioned exactly "You also need a way to invert the input signal to half of each amplifier which means adding an op amp. " which I'm assuming is for single-ended setups since i am using a buffalo.

Hey I bought the boards not so quick to give up on it yet :eek:.

I am definately not looking for hearing damage but found your chart on "more Power" sept blog interesting and have given it a lot of thought.

Ok maybe I am not interpreting you chart for power requirements and recommendations correctly. You recommend 115 dB peak for listening to high dynamic range classical music. Very popular headphone: the Beyer DT880 would be "borderline" with this amp as 7vrms*7/600= 81 mW. Also the AKG Sextettes, Also the AKG 240-DF's, the AKG340's, many of the orthos[help=]%[/help]

Maybe I am reading the chart wrong instead of using the 7vrms, would it make more sense to look at it from Vpeak to peak so instead of Prms on chart those are Pp-p? Or mabey there is just very very little music that requires these 115db peaks?

Not trying to stir controversy, just thought the chart was very interesting and deserved a little discussion as it applies to the )2.

thanks
 
Regal, I backed off to 110 dB SPL in my reply for exactly the headphones you mentioned. Some of the AKGs (in the hard to drive high impedance versions) are also long since discontinued.

This issue has been discussed several times in this (admittedly long) thread. I believe the best solution is to use a different amp rather than turning the O2 into something it's not designed to be. I'm not sure yet, but my upcoming desktop amp might also be a better candidate for those needing more output voltage.

The main issue is a DIP8 op amp, like the 4556, cannot safely dissipate enough power with higher voltage rails when used with some headphones. And if an op amp fails it might take an expensive pair of headphones with it. Power dissipation increases as the square of voltage. So going from 12 to 15 volt rails increases dissipation over 50%. And bridging will increase it by a factor of 4X (2X per IC).

The best solution is to use an amp like The Wire, National LME49600 reference design, etc. using heat-sinked output buffers with higher dissipation capability.
 
Last edited:
Attached - LT Spice sim of one O2 board set up for balanced output into 600R with 500pF in parallel for the headphone cable. Driven with out of phase 3.2Vpeak (2.26Vrms) with 2.5x gain, +/-12V supplies. That input level is as high as I could go in the sim to preserve a good waveshape with the balanced config.

Disclaimer(s): I can't claim to know squawk about balanced systems, never have messed with them much, so I may have mucked something up here. Also just to re-iterate RocketScientists' warnings about hearing and equipment damage! So not recommended that anybody actually do this, just academic curiousity. :)

The key is at the top of the plots. Green and red are the 180 degree out-of-phase 3.2Vpk inputs to the two channels. Dark and light blue are the 1st stage outputs of the two channels. Magenta is the difference voltage across the 600R load, which would be equivalent to putting a DMM or isolated scope probe directly across it - what the balanced headphone would see.

With the 3.2Vpeak differential input, the difference output shown is 15.6V peak = 11.0Vrms. All is sinusoidal, of course, if anybody hasn't seen it yet check out RocketScientist's good recent blog post on music power. :) That would be 202 mW into 600R.

The remaining two plots are the same thing done at 10Hz and 20kHz inputs.
 

Attachments

  • O2 balanced 600R out - balanced out of phase input.png
    O2 balanced 600R out - balanced out of phase input.png
    41.4 KB · Views: 424
  • O2 balanced 600R out - 2kHz 3.2Vpk input.png
    O2 balanced 600R out - 2kHz 3.2Vpk input.png
    31.9 KB · Views: 409
  • O2 balanced 600R out - 2kHz 3.2Vpk input - closeup.png
    O2 balanced 600R out - 2kHz 3.2Vpk input - closeup.png
    25.1 KB · Views: 399
  • O2 balanced 600R out - 10Hz 3.2Vpk input - closeup.png
    O2 balanced 600R out - 10Hz 3.2Vpk input - closeup.png
    28.8 KB · Views: 386
  • O2 balanced 600R out - 20kHz 3.2Vpk input - closeup.png
    O2 balanced 600R out - 20kHz 3.2Vpk input - closeup.png
    27.7 KB · Views: 393
Last edited:
I should also add I'm planning a future blog article to help dispel many of the myths surrounding balanced headphone gear. In most applications, balanced gear has more disadvantages than advantages and many of the alleged benefits are myths. But ever since publishing the 3 channel article I have received lots of questions about balanced designs.

On top of risking frying the O2's output ICs and damaging your headphones from the resulting DC, balanced (bridged) amps also require cutting the plug off the end of all your headphones or otherwise re-wiring them for a 4 wire connection. One would hope people would only do that for some really good reasons, but it turns out, those reasons are usually more mythological than real.
 
Thanks JCX. There certainly is measurable benefit from using a 4 wire connection on a conventionally grounded 2 channel amp in terms of crosstalk. If the difference between say -65 dB and -85 dB crosstalk is audible is another question. But my main point above is a balanced amp forces re-wiring headphones. Being able to use only 4 wire heapdhones is a significant compromise without any objective evidence "balanced" circuitry offers meaningful performance improvements.

At least with a conventional 2 channel amp and a proper ground scheme you can have both 3 and 4 wire headphone connectors if you want to gain the crosstalk benefits of a 4 wire plug. But the 3 wire option goes away completely with a balanced amp.
 
Hey I bought the boards not so quick to give up on it yet :eek:.

I have an idea for the power-mongers out there lusting after 115dBA that consider the O2 hasn't enough grunt. Why not fork off RS's design with your very own O3? Call it the 'Ozone' because where the O2 has just two paralleled 4556 stages in its drive stage yours will have 3. That'll allow you to raise the supply voltage to 15V without compromising the thermals too much.