The 2 philosophies in DAC design, hands off and hands on. Which is better?
- Thread starter muralman1
- Start date
These little shoot-outs are always interesting if run blind. If not, well, not so much. I must admit I'm struggling with the competitive shoot-out as zen thing, though.
P
P there is nothing we do here that won't change the sound drastically. It took me six years to harness my class D amps full potential. The tiny Flatfish transport, just a bit wider than a CD, is constructed of just the essentials on a circuit just millimeters long. Less is more - MUCH more.
Perhaps. I still think you should run the listening tests blind. He who believes he is immune to expectation bias is probably the most vulnerable of all. Less is more? There's one box, about the size of two packs of smokes, between my source and my speakers. I gotcher less right here.
P
Perhaps. I still think you should run the listening tests blind. He who believes he is immune to expectation bias is probably the most vulnerable of all. Less is more? There's one box, about the size of two packs of smokes, between my source and my speakers. I gotcher less right here.
P
Yes, you did
Some NOS still use analogue filters, hence why I did ask (excluding the 1-bit/Sigma Delta/etc).
The 1-bit was relating to my thoughts thinking of ADC mentioned 1st page doh
Which Audio Note you using as their design is really interesting to say the least (appreciate they also do not use analogue filters) ?
Thanks
Orb
Edit:
OK Seems possibly there is still a basic filtering process that is transformer based, was still trying to get my head around not using any kind of filter as it went against principles I learnt about PCM from telephony carriers.
Last edited:
Some NOS still use analogue filters, hence why I did ask (excluding the 1-bit/Sigma Delta/etc).
The 1-bit was relating to my thoughts thinking of ADC mentioned 1st page doh
Which Audio Note you using as their design is really interesting to say the least (appreciate they also do not use analogue filters) ?
Thanks
Orb
Edit:
OK Seems possibly there is still a basic filtering process that is transformer based, was still trying to get my head around not using any kind of filter as it went against principles I learnt about PCM from telephony carriers.
These days, Audio Note DACs have no filters, analogue or digital. My present DAC is an Audio Note 2.1. This DAC has been transformed from a mid-frequency doting DAC to a star performer mostly through replacing four low performing diodes on the digital power supply board with ultra fast recovery diodes. The difference is a totally different sounding AN DAC. Instead of the old lady English sound, this one is open to the extremes in the frequency band. This gives a much truer sound to the music, no matter the genre.
Steve williams
Site Founder, Site Owner, Administrator
Hi
I am wondering how can one not use a filter in PCM decoding ? The Digital to anlaog conversion uses filtering to obtain the signal... ALL DAC use at least a filter of some, ALL. I am not sure I follow your point.. The Audio Note may note oversample but it does have a filter in tis output.. it has to .
I am wondering how can one not use a filter in PCM decoding ? The Digital to anlaog conversion uses filtering to obtain the signal... ALL DAC use at least a filter of some, ALL. I am not sure I follow your point.. The Audio Note may note oversample but it does have a filter in tis output.. it has to .
I am failing to get my point across. I am sure you have had to straighten out a kink in a hose. It's like that. When all the missing highs, lows and faint reverbs are added to the mix, it breaths life into all the performers.
A filter is virtually required in a delta-sigma (OS, PCM) DAC to filter the output noise. It is perhaps less needed in a NOS DAC, but without one HF glitches will be passed on down the signal chain.
Interestingly, the Audio Note site lists the AD1865 dual 18-bit DAC as used in the current DAC 5 Sig; a DAC rated for 16x OS by ADI although they do not appear to be using it that way (they say "1x oversampling" which I assume means no oversampling). It is a conventional segmented (not delta-sigma) DAC that is laser-trimmed (of course) for accuracy. ADI is phasing it out; it is currently on "last time buy" status. Wonder if Audio Note is using the on-chip amps or their own? They may be taking the DAC chip's current outputs straight into the output transformer. The DAC itself is single-ended, though I think Audio Note says the transformer is balanced so balanced outputs can be provided. The ADI data sheet says "low glitch" but does not specify the glitch energy. http://www.analog.com/static/imported-files/data_sheets/AD1865.pdf
Frantz, the output transformer undoubtedly provides some HF filtering and that may be all they use/need. There's always a chance of HF spikes getting into the rest of the system with that approach. - Don
Interestingly, the Audio Note site lists the AD1865 dual 18-bit DAC as used in the current DAC 5 Sig; a DAC rated for 16x OS by ADI although they do not appear to be using it that way (they say "1x oversampling" which I assume means no oversampling). It is a conventional segmented (not delta-sigma) DAC that is laser-trimmed (of course) for accuracy. ADI is phasing it out; it is currently on "last time buy" status. Wonder if Audio Note is using the on-chip amps or their own? They may be taking the DAC chip's current outputs straight into the output transformer. The DAC itself is single-ended, though I think Audio Note says the transformer is balanced so balanced outputs can be provided. The ADI data sheet says "low glitch" but does not specify the glitch energy. http://www.analog.com/static/imported-files/data_sheets/AD1865.pdf
Frantz, the output transformer undoubtedly provides some HF filtering and that may be all they use/need. There's always a chance of HF spikes getting into the rest of the system with that approach. - Don
Don
Entirely agreed and while I was reading your post, it came to me that the output transformer, could/might be used in this fashion, I will not go too deep into the design of their DAC. Not the forum and frankly, my knowledge of DAC subtleties is limited, yet I would have a hard time performing D to A without a filter of some sort. HF are not glitches when it comes to sampled signals they are part and parcel of the process and need to be removed ideally, entirely .. An inductance is far from the theoretical ideal and this is science here ... Results follow theories ... Allow me to continue to have my doubts ...
Entirely agreed and while I was reading your post, it came to me that the output transformer, could/might be used in this fashion, I will not go too deep into the design of their DAC. Not the forum and frankly, my knowledge of DAC subtleties is limited, yet I would have a hard time performing D to A without a filter of some sort. HF are not glitches when it comes to sampled signals they are part and parcel of the process and need to be removed ideally, entirely .. An inductance is far from the theoretical ideal and this is science here ... Results follow theories ... Allow me to continue to have my doubts ...
Audio Note does not oversample. A filter is not needed. Audio Note says there are no filters. All their DACs use the AD1865 chip. That is a non-oversampling chip.
Just read the above posts. Audio Note accomplishes this feat with their circuitry.
I sure am glad I got this DAC going. I noticed an AD1865 chip sold quickly. I wonder what AN is going to do???
Frantz, without knowing anything about the transformer it's hard to say what it's doing, what sort of bandwidth it has, etc. I suspect the stray capacitance around the PCB and transformer plus load is enough filtering. This makes it rather cable-sensitive, which may explain the tale of trial and error muralman1 related. It's not too hard to envision a simple third-order filter (effect) just from those elements (PCB and transformer primary capacitance, transformer inductance, transformer output and cable/load capacitance). In fact, by default it is impossible in a real system as described to not have such a filter; the only question is of bandwidth. Sampling (quantization) noise at the output of a DAC can be pretty broadband so such a simple filter may suffice. That justifies their claim of "no output filter", I suppose. I am guessing, of course...
muralman1, ADI does not list a replacement, only a 16-bit version still in production. The "similar components" list has only delta-sigma DACs; I suspect the conventional style DAC is too expensive (most of the cost is in trim and test) to be competitive, and volume from "purists" like Audio Note is simply too small to justify keeping it around as they move on to other process technology. What AN will do is a good question! There may be other good Nyquist (NOS) DACs out there from e.g. TI, National, Maxim, Wolfson, etc. that AN can use; I have not looked.
muralman1, ADI does not list a replacement, only a 16-bit version still in production. The "similar components" list has only delta-sigma DACs; I suspect the conventional style DAC is too expensive (most of the cost is in trim and test) to be competitive, and volume from "purists" like Audio Note is simply too small to justify keeping it around as they move on to other process technology. What AN will do is a good question! There may be other good Nyquist (NOS) DACs out there from e.g. TI, National, Maxim, Wolfson, etc. that AN can use; I have not looked.
If you guys could all hear this, you would know why I am sad to see the AD1865 go. There are absolutely no artifacts in this playback. The highs are as pure as can be hoped for.
I am not an AN groupie. They do somethings wrong, and won't ever admit it. The diodes I keep mentioning is a good example. No, it is the main example. If I were to upgrade along their path, I would replace all caps with Black Gate, and resistors with Tantalum foil resistors. I don't know about the Black Gates, but the foil resistors, from what I am reading about their capacitance, are not the way I want to go. The same go for the silver everywhere.
When I do experiment, I will look at outside capacitors. I was amazed at what Sonicap Platinum caps can do in other applications.
I am not an AN groupie. They do somethings wrong, and won't ever admit it. The diodes I keep mentioning is a good example. No, it is the main example. If I were to upgrade along their path, I would replace all caps with Black Gate, and resistors with Tantalum foil resistors. I don't know about the Black Gates, but the foil resistors, from what I am reading about their capacitance, are not the way I want to go. The same go for the silver everywhere.
When I do experiment, I will look at outside capacitors. I was amazed at what Sonicap Platinum caps can do in other applications.
As noted by Don, your DAC still has a filter *in the configuration used in your DAC*. This it the schematic:
The cap on the output prior to the transformer input creates a low pass filter, albeit, one not optimized for the sampling rate at hand. Cascading the standard anti-aliasing together with this one may have caused too much high frequency roll off.
Without knowing the parameters of the components in above schematic is hard to know if the system is a) fully passing the reproduced response or b) there are aliasing components being passed through.
I recognize that circuit. It is what AN is using instead of the chip filter. I can't detect an HF roll-off.
muralman1 -- It's there, but ultrasonic (how much/high, I coudn't say without knowing component values) so as to avoid audible roll-off. There's no on-chip filter IIRC, just an output opamp (not used by AN).
Amir, is that from their website? I didn't have time to dig deeper, but it's interesting to know my guess appears confirmed. I love the market-speak of no filter, then showing the filter elements around the transformer explicitly... I would bet the rolloff is a little above the audio band. The output stage looks a little strange, like the load and power connection of the first tube (gain stage, I assume) is missing, but they show feedback from the output cathode-follower. Hmmm...
Oops -- got it; the follower is actually the active load for the gain stage. Had to redraw it to make sense of it. Must have taken lessons from some of my college profs who would draw a basic circuit in a weird way just to see if we caught on. Maybe they just did it so the drawing is more compact...
Back to muralman1 -- I (and Amir, Frantz, and others) find it very interesting to understand what's going on, and we appreciate good circuit design, but none of this techie-speak detracts from the sound you hear! It just helps us understand it.
Onwards - Don
Amir, is that from their website? I didn't have time to dig deeper, but it's interesting to know my guess appears confirmed. I love the market-speak of no filter, then showing the filter elements around the transformer explicitly...
I would bet the rolloff is a little above the audio band. The output stage looks a little strange, like the load and power connection of the first tube (gain stage, I assume) is missing, but they show feedback from the output cathode-follower. Hmmm...Oops -- got it; the follower is actually the active load for the gain stage. Had to redraw it to make sense of it. Must have taken lessons from some of my college profs who would draw a basic circuit in a weird way just to see if we caught on. Maybe they just did it so the drawing is more compact...
Back to muralman1 -- I (and Amir, Frantz, and others) find it very interesting to understand what's going on, and we appreciate good circuit design, but none of this techie-speak detracts from the sound you hear! It just helps us understand it.
Onwards - Don
Yes Don. The picture is a link to their site. If you keep scrolling down it has that schematic: http://www.audionotekits.com/dac2_1.html. Apparently they make their own custom transformer shown in that picture.
Similar threads
