Blocking caps
The output comes straight from the AK4396 dac through a stack of Dayton foil caps at 3uf to the XLR pins. No output R or filter. I use balanced into my BPA100 3875 amps with balanced stepped attenuators at the amps. My various single ended amps are played with pin 3 left open.
.
Of course, I'm not using any of the Behringer's active stages so it is possible that they are doing something worse, the higher the level as the measurements show, than the digital attenuation losses do.
.
Another point that confounds me are the reports from Audio Circle where the user's are running full Didden/ Ottle modified DCXs in revealing VTL or OB home systems and prefer playing digital sources through their dacs and pre amps into the analog in at -20db over using the digital input. Even with optimized levels they are at best still adding an additional pair of AD/DA conversions.
The output comes straight from the AK4396 dac through a stack of Dayton foil caps at 3uf to the XLR pins. No output R or filter. I use balanced into my BPA100 3875 amps with balanced stepped attenuators at the amps. My various single ended amps are played with pin 3 left open.
.
Of course, I'm not using any of the Behringer's active stages so it is possible that they are doing something worse, the higher the level as the measurements show, than the digital attenuation losses do.
.
Another point that confounds me are the reports from Audio Circle where the user's are running full Didden/ Ottle modified DCXs in revealing VTL or OB home systems and prefer playing digital sources through their dacs and pre amps into the analog in at -20db over using the digital input. Even with optimized levels they are at best still adding an additional pair of AD/DA conversions.
output stage modifications
Hi!
Have you ever seen this site?
Lampizator mods
The guy seems very peculiar, but he did a lot of mods on dacs output stages. Regarding the AKM dac, he suggested the following mod:
What do you think about it?
Ciao!
Paolo
Hi!
Have you ever seen this site?
Lampizator mods
The guy seems very peculiar, but he did a lot of mods on dacs output stages. Regarding the AKM dac, he suggested the following mod:
What do you think about it?
Ciao!
Paolo
Actually the right question is about the way he used the balanced output from the dac to "feed" the tube: is it reasonable?
I had it at home (first version of Lukasz design) and don't know how and why but it works, but considering you need 2 or 3 such stages for the active crossover, that's less convinient than single opamps that could also sound not bad when properly implemented.
Yes, most of us know the Lampizator. I have a set of Lampizator tubes, but have never used them.
It was a similar tube output stage on a GoldStar CD player that got me interested in all this DAC modding, almost 25 years ago. Lot's of fun!
Don't think I've ever seen anyone do it on a DCX, tho.....
I have a set of Lampizator tubes, but have never used them.It was a similar tube output stage on a GoldStar CD player that got me interested in all this DAC modding, almost 25 years ago. Lot's of fun!
Don't think I've ever seen anyone do it on a DCX, tho.....
I had it at home (first version of Lukasz design) and don't know how and why but it works
yeah, that's interesting! I like also the way he draws his schematics
I need one stage for each channel. Well, if I can find a way to include an output regulation in this circuit (I am completely against doing it by placing the potentiometer between the output valve and the amplfier
)yes, but the tube appeal is something irresistible for me!!
Ciao
Paolo
...small world...
he did it on the src2496, they use the same DAC, don't they?
In any case, it didn't solve my problem: where do I put my nice motorized potentiometer?
Ciao
Paolo
I don't like the idea of placing it there..... you loose all the benefits that a buffer stage can give you....
Ciao
Paolo
Well maybe not all... But I see your point.
But where the heck you gonna put it otherwise?
I was thinking that maybe the resistor element of the pot could go between the +/- of the DAC output legs. Then take the signal from the wiper. You'd still need a DC blocking cap before your buffer. A nice low value pot would be nice, to give the chip its preferred load, but it may not matter much.
Make sense? Or should I draw it?
But where the heck you gonna put it otherwise?
I was thinking that maybe the resistor element of the pot could go between the +/- of the DAC output legs. Then take the signal from the wiper. You'd still need a DC blocking cap before your buffer. A nice low value pot would be nice, to give the chip its preferred load, but it may not matter much.
Make sense? Or should I draw it?
??
Then, the min. output would be obtained when the wiper is set at the center of the pot (assumed it's B type pot). When tuning it all the way down like on normal volume pots, you'd get max. output of negative phase. No?
You may be correct. But I find it hard to believe that, in general, digital attenuation has much of a deleterious effect. Given choice between running wide open signal and attenuating it by 10dB, or using 10dB of digital attenuation, I would go for the digital attenuation, especially now that I know that at high output levels from DCX distortion increases.
(BTW, I have now confirmed in additional measurements with ST17000B that signal of 20Khz with 8 volts output from DCX produces 0.11% distortion. However, at lower levels or lower frequencies distortion is lower. I expected to see even higher levels of distortion, but that was about it, under the worst case I could fairly simulate. Impedances down to 330 ohms didn't seem to make any difference in the distortion either.)
I don't specifically know about the Behringer DSP, but one would think it would work with 24 bits of resolution. At that level, you would have to reduce gain by 8 bits, or 64dB, to get back to the resolution of 16 bits, the same as used by many of the most highly regarded recordings. There is dynamic range to burn in the digital domain.
I don't think much harm can be done with the 15dB digital attenuation on offer in the DCX, and may be measurably superior to other approaches of attenuation.
Listening
That's what everybody says when they "think" it through. And, the stock DCX analog sections might not be the most revealing device to use to demonstrate the deleterious effects of digital attenuation because of their rising distortion, but if you set up a "listening" test with a good dac as I did, you will hear the sound starting to change for the worse by -12db. Then the question becomes does the digital attenuation do less harm than the added volume control which would take it's place. At -0 to -15 the answer may be yes. I have the fortunate situation right now where the reduced output of the direct out gives me the perfect ( just shy of hearing loss) levels when running straight into the low gain 16V/V setting of the Sure 2X100 so I can run without even my stepped attenuators and use only digital volume control to fine tune the listening level. It has been nicely revealing as I work through mods to this new amp module. People using a Squeezbox or Foobar type volume control at -30db just because that is where the levels happen to be in their rig, will improve the sound by installing a fixed pad or adjustable attenuator at the amp. Depending on the distortion versus power profiles of the analog section of their dac.
Last edited:
That peak signal is ~0dBfs.
The average signal should be around 10dB to 20dB below that.
At around 1Vac output the DCX is clean enough, for most uses. It needs improvement for domestic/critical listening.
This is the signal that needs to be attenuated by a further 10dB to 30dB.
This analogue attenuation will reduce the signal and reduce the distortion and reduce the noise coming from the DCX.
Last edited:
so, summarizing, the best way should be: using a transformer to go from balanced to unbalanced, then connect the potentiometer (100k is too high?) and then the triode buffer stage..... nice, and I belive it will sound pretty well!
Ciao
Paolo
Well, that's what I think is the best way! Certainly not everyone will agree.
It seems that Scott has had good luck just picking off one side of the DAC output for unbalanced. He uses cap coupling. I've never tried the one sided output.
I like the differential output because it "should" be lower noise. Noise and digital remains should be mostly common mode, a transformer helps kill that.
Certainly not everyone will agree.It seems that Scott has had good luck just picking off one side of the DAC output for unbalanced. He uses cap coupling. I've never tried the one sided output.
I like the differential output because it "should" be lower noise. Noise and digital remains should be mostly common mode, a transformer helps kill that.
You are free to make this judgement based on tests you have performed. But there is no obvious explanation as to why this should be. It does not follow from standard audio engineering theory.
So long as you are preserving way more useful dynamic range in the digital domain as you have in the analog domain, there should be no audible difference from digital volume reduction.
So I think others should be wary of generalizing from your tests. Personally I don't believe your results, I think that in a properly controlled and blind test, neither you nor I would be able to tell the difference caused by 12db of digital attenuation under typical circumstances. I can imagine circumstances where 30dB of digital attenuation would not be audible, even in the context of audio reproduction of the highest quality.
Generally, except in extreme cases like more than 64dB gain reduction, we would only hear a difference when some other noise or distortion source is increased to the level of audibility to compensate for the digital attenuation. Generally that doesn't even happen with digital volume output reduction, it's actually reducing the effect of analog harmonic distortion, digital quantization inaccuracy, and noise. It can, however, occur when using post-DCX analog attenuation when digital volume reduction would have been more appropriate. And it would not surprise me if many people on this list are making this mistake, as I was.
To be honest, whenever I read a subjective review of a modification I pretty much ignore it. It's not that I think the mod didn't make a difference, but that I have to question if something else is suffering as a result. A good example is in another thread (this one) where it was pointed out by a relative authority (from my perspective) that the mods would have a serious negative impact on the amp, that the only thing that is saving it from total failure is its low power. Yet the modder claims huge sonic improvements.
So when I see comparisons between the output at -12db as opposed to 0db and how one sounds more "revealing" or "open" and that it all goes to hell in a handbasket when the level is adjusted digitally, I think: maybe the modifications had a broader effect than anticipated. Maybe what "sounds" good on one level is actually doing damage at another.
Something to think about.
So when I see comparisons between the output at -12db as opposed to 0db and how one sounds more "revealing" or "open" and that it all goes to hell in a handbasket when the level is adjusted digitally, I think: maybe the modifications had a broader effect than anticipated. Maybe what "sounds" good on one level is actually doing damage at another.
Something to think about.