The Objective2 (O2) Headphone Amp DIY Project

Soldering Question:
I once read somewhere that cutting the lead first and flat to the pcb before soldering gives better performance. Because there are no sharp edges (from the cut) on the finished solder that can act as an antenna, if i remember correctly. Or is soldering first and cut the leads afterwards better. Option C is neither of the above. So what do the pro's think around here?
 
Little antennas? You'd be up in the buh-jillion Herz range or so fractional of any terrestrial frequency that it's not worth fussing over.

Anyway, if you want them gone (and there's one on the O2 you really need to have flush with the PCB), I'd much rather do whatever it takes to not have to file it down after it's soldered. My dikes have a little ramp to it so I could take some, but not all of the pile. Maybe that's good enough for you too.
 
Soldering Question:
I once read somewhere that cutting the lead first and flat to the pcb before soldering gives better performance. Because there are no sharp edges (from the cut) on the finished solder that can act as an antenna, if i remember correctly. Or is soldering first and cut the leads afterwards better. Option C is neither of the above. So what do the pro's think around here?

I don't know about the "antenna" business. In earlier times it was considered good practice to "clinch" leads, that is, to bend them flush to the PCB before soldering, but I think it was just thought to provide a more solid, secure joint. I don't think anyone would have seriously argued that it produced measurably better performance. But, in the days before wave soldering when everything was hand soldered, I think properly clinched leads were taken as a sign of "professionalism". Of course they also made everything harder to repair.
 
Alexium, whatever opamp you choose for SE conversion, make sure it can slew really fast, not ridiculously fast unless you know how to tame such things as ADSL drivers, but fast enough to cope with the edges of a very fast waveform without distortion or ringing. i see no benefit whatsoever in using such a low output level, what is the point if you have to add it back in gain of the next stage and in the meantime have a much more susceptible low voltage, low current output? the idea of minimalism is to use enough to do the job properly and nothing more
 
Last edited:
Alexium, whatever opamp you choose for SE conversion, make sure it can slew really fast, not ridiculously fast
Could you please make a guess on appropriate slew rate range?
i see no benefit whatsoever in using such a low output level, what is the point if you have to add it back in gain of the next stage
No benefit, except, like I said, I'm using passive I/U converter, and it's resistance should be as low as possible, resulting in low output voltage (about 150mV).
 
I think it's an urban myth :)
LOTS of things in the audiophile world are myth. They usually start from a grain of truth, get convoluted by audiophiles, and are repeated often enough to create a widespread myth. The other posts are correct. At gigahertz frequencies little sharp bits to nowhere can matter--but not even remotely at audio frequencies.

The PCM179x chips are great. Isn't it interesting how most all of the modern DACs aren't as good as the ones created 6-12 years ago?

This makes me wonder how the ODAC is coming along. I'm itching for details.
I need to do more research, but I'm not sure I buy most of "NOS DAC" arguments. It seems you're merely trading one minor set of issues for another. If modern DACs can be entirely transparent in listening tests (nobody can tell when they're added to the signal path), and measure extremely well, I'm not sure what further benefit ancient DAC chips have? Does anyone have any <i>objective</i> evidence supporting 10 year old DAC chips? See "urban myths" above. ;)

As for the your ODAC question, I just posted an update on my blog. I'm still a few more weeks out before the next ODAC article.

I just wanted to say a big thank you to everyone who have posted tips and tricks on how to make a face plate, with only the minimum of tools. I do not think I could have come up with eve half of this on my own.

This is the first time I've tried drilling into aluminum, and I think it proves that as long as you do not expect boutique quality looks and perfect alignment, there's nothing stopping you doing it on your own.
Thanks for posting the pictures. Your panel turned out well. I especially like that you showed it screwed down to the scrap wood. There have been lots of great suggestions in this (now massive) thread.


Could you please make a guess on appropriate slew rate range?
I would suggest following the reference design for whatever DAC chip you choose including using the same op amp. Slew rate really depends on the filtering in the DAC chip, what sample rates it will run at, and if you want to have enough slew rate to cover signals that only exist in test tracks (versus commercial recordings of music). Many DIY, and some "boutique" commercial designs, use op amps that are too fast.

Most really fast op amps are compromised in other ways (higher noise, distortion and/or power consumption with lower stability). Without the test equipment to measure your design, you really won't know what you end up with. You might hear a difference but it could be ultrasonic oscillation. You could mistakenly conclude it's because one V-I converter sounds different than another but it's really that one of your designs is completely unstable.

As others have suggested, trying to use a passive I-V stage is self-defeating as you just have to make up the gain later. It's safe to assume the best performance will be achieved with the manufacture's reference design. They're all trying to out perform their competitors DACs and the reference designs are the ruler by which they're evaluated and measured. If something improved the performance significantly, you can be reasonably sure they would incorporate it in the design.
 
OK RS, so what are your thoughts as to using the ODA/ODAC as a preamp stage? I'm thinking of (eventually) using that for "critical listening" by hooking it directly to an amplifier (with no volume control). Or would the ODAC line out with a passive volume control be better? Let's assume the amp has a fairly high (~20KOhm) input impedance. Because the ODA is so transparent, would there be any harm in having it in the chain? Any benefit?
 
@FloridaBear, I would very much suggest using the ODA (or some other high quality active preamp) after the ODAC rather than a passive volume control. Passive volume controls seem like a nice idea, and they're sometimes OK for non-critical applications, but they have some significant issues. The most notable problem is they dramatically increase the output impedance at typical volume settings of whatever is connected to them. That can create several problems and usually degraded performance over a decent active preamp. And their performance will vary depending on the components, and sometimes even cables, involved.
 
the best performance will be achieved with the manufacture's reference design.
Initially I thought so too. Then I asked a question on another DIY forum, and gurus of DAC building criticized the reference schematics all they could, saying it's only designed to achieve high S/N and dynamic range. I'm not sure if they are right or not, but they do know something about DACs (and I don't), so they made me doubt.
P.S. I know passive I/U is definitely worse than op-amp based one, I just wonder if I can hear any difference at all. Besides, it's so simple, how can I not try it? :)
 
Last edited:
@Alexium, while there are some genuine gurus lurking around in DIY audio, the vast majority have very little objective evidence to back up their claims. They don't run proper blind listening tests and they lack the right test equipment to verify their claims and designs. Instead, nearly all use highly biased and flawed sighted listening. See: Subjective vs Objective

And, for what it's worth, I don't even consider myself an overall guru. I mostly leverage the work of the genuine gurus--guys like Doug Self, Bob Cordell, Bruno Putzeys, Ethan Winer, Samuel Groner, Walt Jung, etc. All of them back up their claims with credible measurements, blind listening tests, peer reviewed AES papers, engineering calculations, etc. I would suggest looking for genuinely objective evidence supporting claims about I-V stages, or any similar topics. If there is no such evidence, you may want to consider why not?
 
yeah RS got it in one and thats why i didnt make a specific recommendation, because it very much depends on the dac and target output bandwidth you are chasing. but afaik it doesnt just need to be able to cope with the audible music bandwidth, but also to be able to eat up any HF out of band signal and feed it back through the LPF in the loop instead of not coping with the BW and sometimes causing aberrations in the signal. speed for speed's sake is just as silly as RS mentions and having to overcompensate a high speed opamp just to keep it stable in the audio bandwidth or at unity gain as called for in most IV stages; is not what i call a solution.

agreed datasheets are a very good place to start and sometimes a good place to end, particularly if you have no way to quantify the results of other experiments. even if you dont use it in its entirety the information in a good app note is invaluable, but i disagree they are the end game for all dacs by all manufacturers, as there is some truth to the statement that sometimes they are shooting for highest measured performance in a few somewhat dubiously relevant areas. all imo of course

edit, just saw RS's list of mentors :nod:
 
Last edited:
@Alexium, I would suggest digging up some of the application notes, tech papers, etc. by some of the better chip companies. You'll find quite a bit at Cirrus, TI/Burr Brown, Analog Devices, etc. In some cases the user guides for their reference boards can be very informative as they often include measurements, design trade off explanations, etc. Sure there's some bias but there's also a lot of good practical objective info.

I would also suggest Benchmark Media. John Siau and a few others have published several white papers on various aspects of DAC performance. Again, not without some bias, but much of it is purely objective and easily verified by others. Anedio has somewhat more biased, but still useful, measurements and info published as well.

The above companies all have measurement equipment, resources, and usually engineering knowledge, that goes far beyond 99.99% of what DIYers have at their disposal. So the pros are in a much more valid and credible position to research what works best and document the results.

@qusp I partly agree about reference designs not always being the "end game" but for somewhat different reasons. I think the biggest constraint on the DAC reference designs isn't so much focusing on the wrong things, but because they're typically somewhat cost and complexity constrained. Being realistic, so are most DIY designs so that's often not a bad thing.

A semiconductor company doesn't want their DAC reference board to cost $1000 to build in low volumes. Nor do they want to imply you need $980 worth of other parts to make their $20 chip perform well. But, IMHO, they aim well past the point of diminishing returns while keeping the component costs and complexity reasonable. You might be able to get another few percent of ultimate performance out of the chip by spending a lot more money, and/or using a substantially more complex approach, but I'd be really amazed if any such changes could survive a blind listening test.

The other area they may somewhat compromise is "nepotism"--they like to specify their own company's parts for obvious reasons even if they're not always the ideal choice. But, again, I doubt that significantly holds back many reference designs.
 
Just tried to build my O2 last night.

I'm having some problems with the voltage regulators though. I think they may be cooked but would like another opinion before I go buying more just incase I'm wrong.

AC adapter output is 14.5VAC.

Input to voltage regulators is 16.9V and -16.9V for +/- reg respectively.

From output of the voltage regulators though I'm getting -6.9V and 7V. Polarity same but 5V lower than the expected +/- 12V. This is measured at D1/D5 as per RS testing instructions. Is it likely that I have blown both regulators and each failed regulator is giving an output that is still this well matched?

My camera batteries are flat, I can take a pic once they're charged if it'll help ...
 
Doubtful that the regs are cooked. I'd take a real good look for any solder bridges at both top and bottom of the board. Also double and triple check that the power supply components are correct values and installed correctly. I've only skimmed RS's new O2 troubleshooting section of his blog, but so far he's been nothing if not thorough in his documentation. It's likely you'll find help there, aside from anything he may want to add here in the thread.
 
Thanks, that is hopefully good news!

Would anyone mind having a gander at these pics, I haven't spotted anything jumping off the page at me yet.

In addition to the measurements in my last post nothing is getting hot/warm which was another of the checks that RS recommended checking for in his testing guide. Power button is still in the off position.
 

Attachments

  • P1000770.jpg
    P1000770.jpg
    380.9 KB · Views: 334
  • P1000765.jpg
    P1000765.jpg
    361.4 KB · Views: 333