The Objective2 (O2) Headphone Amp DIY Project

My suggestion of the FTDI chip was only if a MCU based system would be the only solution for audio decoding. And just to enable every user to update the software themselves and control it from the computer at the same time because buying a programmer or be limited to preprogrammed chip can be troublesome.
In that case the extra cost of 2232 would be a good candidate for communication and programming only as you said. I assumed that the microcontroller would magically solve the audio part since thats the reason for it...

Still if you can find a good system which relies only on the standard usb drivers thats a really elegant and preferable solution. Even if some digital interface might be required for a rotary encoder which should be easier to solve (connection to a computer would be a nice bonus).

I haven't tested any USB sound card on linux but it has usually got fairly good support for most normal usb devices without any hassle. My guess is that it should work fine but please test to be safe ;)

Regarding alcohol (small amounts only) and soldering it might work. I know that some people used it before on shooting competitions (good old days) but it was a double edged sword. Because it gives a calming effect for a short duration of time but after that it becomes far worse.
My bet would be to stay away from all drugs including alcohol and caffeine since it will only cause trouble in the long run. And the positive effects are very limited at best...
 
My suggestion of the FTDI chip was only if a MCU based system would be the only solution for audio decoding. And just to enable every user to update the software themselves and control it from the computer at the same time because buying a programmer or be limited to preprogrammed chip can be troublesome.
In that case the extra cost of 2232 would be a good candidate for communication and programming only as you said. I assumed that the microcontroller would magically solve the audio part since thats the reason for it...

Still if you can find a good system which relies only on the standard usb drivers thats a really elegant and preferable solution. Even if some digital interface might be required for a rotary encoder which should be easier to solve (connection to a computer would be a nice bonus).

I haven't tested any USB sound card on linux but it has usually got fairly good support for most normal usb devices without any hassle. My guess is that it should work fine but please test to be safe ;)

I believe the audio-widget group will get this solved on Windows too eventually when someone developes the neccessary driver and yes - the solution is MCU based on AtMega32 (I would love to solder one of those again since I have thought of a different approach for the next one ;-)). As I wrote before Linux et al and maybe Apple already has this built in there is no need for drivers - heaven... Now about Microsoft... They are very reluctant to build an UAC2 common driver since their standpoint is that they don't develop hardware so the drivers has to come from the hardware manufacturers. One might ask where the UAC1 common driver came from? This is itchy - otherwise I would be perfectly happy with Linux where dreams comes true (with some lesser effort).

Brgds
 
Last edited:
I believe the audio-widget group will get this solved on Windows too eventually when someone developes the neccessary driver and yes - the solution is MCU based on AtMega32 (I would love to solder one of those again since I have thought of a different approach for the next one ;-)). As I wrote before Linux et al and maybe Apple already has this built in there is no need for drivers - heaven... Now about Microsoft... They are very reluctant to build an UAC2 common driver since their standpoint is that they don't develop hardware so the drivers has to come from the hardware manufacturers. One might ask where the UAC1 common driver came from? This is itchy - otherwise I would be perfectly happy with Linux where dreams comes true (with some lesser effort).

Brgds

apple was the first to champion UAC2, its been standard for years
 
RS!

I can confirm that I have a second one with low battery problems. This one is quiet for about 3s and then goes on for less than ½s. It has the original 2.7M R25.
I believe that this problem occures with fresh batteries especially if you have powerhungry headphones. I use the HD-650's. The headphones that doesn't load the amp and batteries that hard has lesser problems with this I believe.

Brgds
 
In order to ensure you don't exceed max power handling, best use the piece of kit that normally goes between the earpieces.

Headphones usually distort badly before they reach maximum power, which should be plainly audible. In the mid-high frequency range, hearing will probably complain about levels even before that. Moreover, dynamic drivers usually aren't as sensitive to damage as, say, 'stats, which can be ruined by arcing.

A halfway accurate variable impedance power limiter is, well, I wouldn't say impossible to build, but it would be a non-trivial task for sure. You'd need both a model of impedance response and a thermal model somewhere.

If there is a known (fixed) source level, you can choose gain values accordingly, but that'll never be particularly accurate - aim for worst-case (sine) power, and levels with real-life music (higher crest factor) will be limited too much, and then there still is the problem of varying impedance.

The O2 has no provisions for current limiting resistors inside the feedback loop, which would eliminate the fixed source voltage requirement but leave the other problems.
 
So I'd basically have to cut the design in half (one 4556) and make sure my gain is under 2x, right? That would keep me way under that level... or am I wrong? My cans need up to 125mW at 55 Ohm to ensure I can listen to some really good classical. My reason for wanting the upper limit is because I have young kids in the house who are starting to like listening to my cans. My son almost blew my cans to pieces with my current amp because it can deliver ~600mW into my AKGs which are rated to 200mW.
 
How would I go about modifying the O2 to ensure I can't exceed 200mW into 55Ohms (really 55-125 Ohm according to measurements)?

If you are only running it on AC power with the AC adapter you could change the voltage regulators to +/-9Vdc:

LM7809CT Fairchild Semiconductor Linear Regulators - Standard

LM7909CT Fairchild Semiconductor Switching Converters, Regulators & Controllers

With +/-9V rails the amp starts to clip around 4.7V, as shown in the sim below into 55R loads. That is 4.7V peak = 3.32Vrms, which would be (3.32V) ^ 2 / 55R = 200mW.

This would mess up the battery charging ability though. 8.4V cells need to charge up to about 10.2V and they can't do that if the rails are only 9V. :) So this would turn the O2 into an ac-only unit.
 

Attachments

  • 02 200mW 9V rails plot.png
    02 200mW 9V rails plot.png
    18.6 KB · Views: 289
So I'd basically have to cut the design in half (one 4556) and make sure my gain is under 2x, right? That would keep me way under that level... or am I wrong? My cans need up to 125mW at 55 Ohm to ensure I can listen to some really good classical. My reason for wanting the upper limit is because I have young kids in the house who are starting to like listening to my cans. My son almost blew my cans to pieces with my current amp because it can deliver ~600mW into my AKGs which are rated to 200mW.

This wouldn't work, one 4556 is used per channel (L,R) in parallel to increase the power output. You run one chip and all you do is lose one channel.

Your best bet, if you're really concerned, would be to just use unity gain and leave out the gain resistors entirely. You could try to diy some type of stop to limit the volume pot travel as well.
 
Last edited:
How would I go about modifying the O2 to ensure I can't exceed 200mW into 55Ohms (really 55-125 Ohm according to measurements)?
As has already been suggested, just use 1X gain (or possibly more depending on your source). Very few sources put out much over 2 Vrms. With 1X gain that's only 72 mW into 55 ohms and 32 mW into 125 ohms.

Or, working backwards, 200 mW at 55 ohms is 3.3 Vrms. If you know the highest output source you'll be using, set the gain at 3.3/Vsource. With a typical 2 Vrms home Redbook-compliant source, that's 3.3/2 = 1.6X gain. With an iPod LOD, however, it's 3.3/.5 = 6.6X.

If you're worried about accidents, use the above gain for the High Gain setting and have the Low Gain option be less or the same.

That way, even at full volume, playing a track hitting 0 dBFS, you can't get more than 200 mW out of the O2 into 55 ohms (and less power into higher impedances).
 
How far are we away from a desktop version of the O2?

Any early info to share?
Some of this may be old news to some, but the official name is "ODA" for Objective Desktop Amp. There's some early info, if you haven't seen it already, towards the end of the O2 Summary article (under "WHAT ABOUT A DESKTOP ONLY VERSION?").

In addition to what's linked above, the latest news is the ODA DAC daughter board. As I've mentioned in this thread, the current plan is for a relatively simple, small, single board USB DAC option that supports true high-resolution 24 bit audio at native 44, 48, 88 and 96 Khz over USB (something relatively few DACs can do regardless of what 24/192 DAC chip they use and virtually no DIY DACs support 24 bit USB).

The DAC complies with the USB Audio standard and hence has native driver support, even at 24/96, built into Windows, OS X and presumably Linux. This is in contrast to most high resolution pro audio interfaces, and the open source AVR-based DIY DAC, that require proprietary drivers.

The DAC board will "piggy back" on top of the headphone amp board as a daughter card and there will be a switch to select the line input or DAC as the source.

I'm currently testing the DAC board and it offers well in excess of 16 bit performance (many "24 bit" DACs don't) providing a significant benefit for those using software volume controls compared to a 16 bit USB DAC. It also has very low jitter and respectably low distortion. Like the O2 and ODA, it's a minimalist design focused on maximum performance for the lowest cost.

As previously mentioned, for programming, licensing, testing, and SMT reasons, the DAC daughter board will only be sold pre-assembled (but not by me and I will not receive any revenue or profit). So you can build your own ODA and then just solder the DAC board on top.

For those wondering why not full DIY, there are no USB solutions that I know of supporting 24 bit operation that can be purchased and used off-the-shelf in a DIY project without requiring at least programming. And the better solutions are only available under license and/or other contractual agreements. That makes a full DIY DAC very difficult as DIYers can't buy the chips, and even if they could, they require proprietary programming before use. The only exception I currently know of is the AVR solution already mentioned but even it requires programming and extra hardware to do so. It also requires a proprietary driver for use.

I'm hoping to publish the first ODA article and start the first thread here in diyAudio in the first part of December (the next few weeks).
 
RS!

"The only exception I currently know of is the AVR solution already mentioned but even it requires programming and extra hardware to do so. It also requires a proprietary driver for use."

No HW needed for programming (ISP), Linux and our Apple friends already have UAC2 built in (no need for drivers) it's only us occational Windows users that have to reely on UAC1. So the Audio-Widget is about what we need... CD quality can be darn good if you choose the right productions!

Even 320kps mp3's is almost quite ok listening to IF the source if of quality...
I'm actually listening to Hans Theessink - Call me - amazing (mp3-320)... I have to buy his CD's or better!

My feeling is that we need to crunch this nut regarding MS reluctance to create a common driver. Now how would we do that? I sent a request to MS about it and got as an answer that MS doesn't do HW so the driveras are of the HW manufacturers concern, not MS. But they kindly hinted about a thread in, now it stands still in my head, the MS dev site - now what it is named?
There is probably things going on inside MS as well - they have a UAC1 common driver so why not a a UAC2 one? Everyone else has one built in...

So the Audio-Widget guys are on the right track IMO.

Brgds
 
Last edited:
@Turbon, around 90% of the desktop/laptop PCs on the planet are running Windows. So, yeah, a tiny minority may not need drivers ;) That will slowly change depending on how quickly Win8 is adopted (or perhaps Microsoft will release UAC2 Win7 drivers as part of an update).

ISP does require hardware--it either has to be built onto the board, or outboard like the AVR-ISP from Atmel. Your average analog audio DIYer isn't set up for MCU programming. Or can the design use a factory USB-based bootloader?

For those who are comfortable working with microcontrollers, and don't mind a "2 box" or custom enclosure/packaging, the AVR design is a more flexible option. But I really doubt it will sound any better. It also may cost more than the ODA DAC board--certainly with both in assembled form.
 
Now RS. Of course when reaching the greater public outside of nerdy sites as diyaudio the widget has the latest driver in memory ready for windows and info for windows update about where to find the latest... But this won't be needed since MS is going to release a UAC2 common driver - the question is how much badwill they are prepared to draw upon themselves over this before release...

I'm not selling anything and will of course with great pleasure try whatever you decide to be the best companion for the O2. The fear I might feel is that the DAC isn't usable to it's full potential outside of Windows...
On the other side You doesn't seem to be one of the evil ones and I run windows whenever I feel to or have to (every day :) ).

Brgds / Nerd and proud of it...
 
Last edited:
Now of course we/they are RS.

So we are expecting 24/48 as maximum? Well, it aint bad with the right source :)

BTW, I'm not going to give any benefits, cash payments or be at any help for those who should show an intrest in in using anything outside of windows regardless that I know that you and only you have to cross that white line painted on the highway. Not everyone is ready for that shift of mind. I'll setle for using whatever that suits me best for the job.

Brgds / Occational Geek
 
Last edited:
So we are expecting 24/48 as maximum? Well, it ain't bad with the right source :)

UAC1 supports 24/96 and so will this DAC.

One thing I'd like to mention is that, last I knew, Windows adjusted the native audio feed bitrate on a song by song basis so that nothing had to be soft-scaled as long as the DAC supported it. I believe Mac doesn't do that yet. Where does Linux stand?