Output stage for DAC - opamp vs tube

Interesting. I'm subscribing to this thread

siva do you want to start a Tube vs SS discussion

Arj schrieb:

siva do you want to start a Tube vs SS discussion ;)

Not really. This is a very specific application of tube.

All though i like the sound of tube over SS, i'm not very sure whether tube is the right choice for this particular application...hence the question

Further, You are not comparing tubes to power mosfet/transistors...here.

Hi Siva,

IMHO, this Does boil down to a SS vs Tube issue.

Maybe a more 'neutral' comparision would be Op-amp output stage with a discreet Transistor Output stage for the DAC ?

What has been your experience..

Frankly, I have no DIY experience in DACs. Even my technical knowledge in not thorough.

That said, from a theoritical perspective, I would list the following Pros & Cons for the OpAmps vs Tubes.

Background: The DAC outputs a Current... a SMALL current, I may add. When this current is applied accross a resistor, we effectively get a Voltage, which is EXACTLY proportional to the current - Ohms Law) IF THE RESISTOR IS CONSTANT & A PURE RESISTOR.

The DAC needs to provide a VOLTAGE Output, for the Pre Or Power Amp.

Given the low current output, a High ( VERY High value preferred, from the low current point of view ) value resistor is called for.

This implies that the Output Buffer ( Op amp or Valve ) must have a Very High Input resistance.... actually a very high IMPEDANCE )

( Impedance = Resistance + Capacitance + Inductance )

Most Valves provide a very high Input resistance, but some also present a relatively high Input Capacitance ( compared to opamps). High input capacitance implies that the Input Load from the buffer ( Valve ) changes with frequency.

As a result, the output voltage produced will Not Exactly mirror the output current from the DAC...

Therefore a Thumbs down for the Valve on this count, and brownie points for the OpAmp.

In contrast, ALL opamps MUST be operated with HUGE Global feedback... a phoo phoo in contemporary design.

Valve buffers often do not have the DC to Light frequency responses that modern OpAmps provide. As a result, the valve buffer acts as a Low Pass filter, suppressing some digital hash but also killing High Freq extension, and true, extended bandwidth.

(Throwing out The Baby With The bath-water ? )

I am sure that there are other aspects, and Siva and Bombaywalla, and Viren and others could throw some more light ?

Ofcourse, Siva has asked for

What has been your experience..

so please... lets everybody jump in with your experiences... preferably naming the DAC(s) that you have experienced.

Amp_Nut schrieb:

Background: The DAC outputs a Current... a SMALL current, I may add. When this current is applied accross a resistor, we effectively get a Voltage, which is EXACTLY proportional to the current - Ohms Law) IF THE RESISTOR IS CONSTANT & A PURE RESISTOR. The DAC needs to provide a VOLTAGE Output, for the Pre Or Power Amp.

you lost me a bit here, Amp_Nut: you 1st said that the DAC outputs a current & then said that the DAC needed to output a voltage for the pre or power amp.
So, which is it: DAC outputs a current OR DAC outputs a voltage??

Amp_Nut schrieb:

Given the low current output, a High ( VERY High value preferred, from the low current point of view ) value resistor is called for. This implies that the Output Buffer ( Op amp or Valve ) must have a Very High Input resistance.... actually a very high IMPEDANCE ) ( Impedance = Resistance + Capacitance + Inductance )

I'm afraid that I got lost here too!
If a very high value resistor is need to generate a reasonably high voltage then how do you conclude that the opamp or tube needs to have a very high input impedance??

Amp_Nut schrieb:

Most Valves provide a very high Input resistance, but some also present a relatively high Input Capacitance ( compared to opamps). High input capacitance implies that the Input Load from the buffer ( Valve ) changes with frequency. As a result, the output voltage produced will Not Exactly mirror the output current from the DAC...

it seems that the excess capacitance in the tube could result in an excess phase shift onto the (music) signal thereby changing the original phase relationships. This will make the music sound different (if the system is resolving enough to let it thru).

Amp_Nut schrieb:

In contrast, ALL opamps MUST be operated with HUGE Global feedback... a phoo phoo in contemporary design.

i'm not sure that it is global NFB because gNFB implies feedback from the final output all the way back to the input of the DAC (the input i'm referring to is the input from the CDP or DVDP). I do not think that these buffer stage opamps are configured that way. I believe that they have feedback around them so that is tantamount to local/output stage NFB.

Amp_Nut schrieb:

Valve buffers often do not have the DC to Light frequency responses that modern OpAmps provide. As a result, the valve buffer acts as a Low Pass filter, suppressing some digital hash but also killing High Freq extension, and true, extended bandwidth.

And this is precisely the wave that the valve-wallas are riding!!
You know everyone has phoo-phooed digital for years & years saying all kinds of derogatory things (1000+ page books can be printed with all those comments!) one of which being that digital output is harsh & that it gives the ears "digitis"! So, people discovered tubes to roll off the high freq hash naturally & yet provide a more analogy sound. People have been so happy w/ this mode of operation that they have forgiven the roll-off of the hi freq. It has spawned a # of CDP companies Audio Aero, Lector, Canary Audio, etc to name a few that use tube output stages.

Amp_Nut schrieb:

I am sure that there are other aspects, and Siva and Bombaywalla, and Viren and others could throw some more light ? Ofcourse, Siva has asked for What has been your experience.. so please... lets everybody jump in with your experiences... preferably naming the DAC(s) that you have experienced. :prost

unfortunately I do not have any experience w/ tube output stages in CDPs or DACs. I try to avoid such units as I personally believe that a tube does not belong in this particular part of the audio chain. I, otherwise, like tubes a lot!
All the DACs & CDPs I've been exposed to have s.s. output stages. And, I really like all of them! I cannot find too much fault in any of them!

Dear Bombaywalla,

As always you have approached the post with much Gusto...

I will presume that your post was the result of an enquiring mind so I will respond .....

1.

you lost me a bit here, Amp_Nut: you 1st said that the DAC outputs a current & then said that the DAC needed to output a voltage for the pre or power amp. So, which is it: DAC outputs a current OR DAC outputs a voltage??

the DAC ( Digital To Analog Converter) is usually an IC (intergrated circuit, consisting of an R + 2R network, who's output is really a current. This current is applied to a resistor, accross which a voltage is generated. The outside world ( Hi Fi Chain after the DAC) needs a Voltage, so the voltage generated accross that high value resistor, needs to be Buffered ( Isolated ). THAT is the function of the DAC Buffer.

To answer your question, the DAC IC outputs a current, but the DAC UNIT, ( as sold in a stand alone Hi Fi Product ) needs to output a Voltage, to the rest of the Hi Fi Chain.

Hence the DAC as a Hi Fi Product, consists of a DtoA IC + The Buffer Stage mentioned by Siva.

I did a quick Google and can recommend this link for some basic reading :

http://www.ikalogic.com/dac08.php

2.

If a very high value resistor is need to generate a reasonably high voltage then how do you conclude that the opamp or tube needs to have a very high input impedance??

The Buffer's Input Impedance appears in PARALLEL to the external resistor ( also called summing resistor) of the DAC IC.

Low or a fluctuating input resistance will mess up the Summing resistor's action, because the input impedance will swamp the summing resistor.

it seems that the excess capacitance in the tube could result in an excess phase shift onto the (music) signal thereby changing the original phase relationships. This will make the music sound different (if the system is resolving enough to let it thru).

I agrre... yours is a nice, alternate expalanation.

Your reasoning is "Time Domain" associated, while what I said was Freq Domain associated. ( Yah... Hi Brow terms .. )

Most non engineers are not exposed to the concept of Phase, so i choose a frequency domain expalnation !

i'm not sure that it is global NFB because gNFB implies feedback from the final output all the way back to the input of the DAC (the input i'm referring to is the input from the CDP or DVDP). I do not think that these buffer stage opamps are configured that way. I believe that they have feedback around them so that is tantamount to local/output stage NFB.

Here we are referring to the Output stage, so Global feedback for the buffer, refers to 'Globally To The Output stage.

The point that I was making w.r.t feedback differences in an Op Amp and a Valve Buffer are :

1. OpAmps typically have a gain for 1 million i.e. 1,000,000. Feedback is used to bring this down to a gain of just 1 in a buffer stage.

So there is OODLES of feedback. ( 100% feedback, actualy)

The valve stage has an intrinsic gain of maybe 100, or or so. Local feedback drops it to a gain of 1 in a Buffer application.

Hence the feedback deployed in a Valve Buffer and an op amp buffer are POLES apart. { no pun intended w.r.t the Bode plot }

I, otherwise, like tubes a lot!

That makes 2 of us !

Thanks for the clarifications, Amp_Nut!
I thought that's what you meant to say but did not want to assume anything hence the "inquiring minds" questions!


From an opamp perspective, I've been on both sides of the fence. As a designer of such circuits FOR NON AUDIO use, the opamp is heaven-sent circuit. That oodles of open-loop gain actually makes the overall ckt work from a output noise, input DC offset, output impedance, distortion perspective.
However in audio, as we know, quality of sound is king & tho the unit measures superbly on the bench, it probably sounds half as good (or maybe even like a POS!). At that time, I agree with you, gobs of NFB around the output stage hurts.

I have to say tho' that this Scott Nixon DAC which has a couple of OPA627 as a buffer stage sounds absolutely dynamite! I think that even Saaya (who owns a different Scott Nixon DAC) is very happy with the DAC's sound.
Like many times before, is it Scott Nixon's implementation?

Amp_Nut schrieb:

Your reasoning is "Time Domain" associated, while what I said was Freq Domain associated. ( Yah... Hi Brow terms .. )

oooohhhh! aaaaaahhhh!
yah! hi-fla-luting terms, Amp_Nut!

Forgot to mention another category here.

Has anyone heard any DAC, that uses neither OPAMP or Tube...but just a transformer to step-up the analog signal from the DAC chip ??

Any thoughts, experience....

siva , is that not how the 47 labs dacs work (progression ? )

Arj schrieb:

siva , is that not how the 47 labs dacs work (progression ? )

I'm not sure about 47 Labs...

Regards
Siva.

sivat said :

Has anyone heard any DAC, that uses neither OPAMP or Tube...but just a transformer to step-up the analog signal from the DAC chip ??

This is an (interesting ? ) and new concept to me.

I would suspect that the Transformer's impedance will vary at different frequencies ... causing poor performance, or atleast a "different" sound.

As I had said in my earlier post :

Low or a fluctuating input resistance will mess up the Summing resistor's action, because the input impedance will swamp the summing resistor.

If a commercial product does use a Transformer as a summing stage, I would be

Very

interested to lean more of it.

Hmmmm.....Amp_Nut,...if what u are saying is true...what about TVC

[quote]If a commercial product does [b]use a Transformer as a summing stage[/b], I would be Very interested to lean more of it.[/quote]


i am not sure if i understand the above..is there any "summing logic" required after the output from a DAC chip ?

Yes, Arj.

The Digital to Analog Converter (DAC) Integrated Circuit (consisting of an R + 2R network,) outputs a CURRENT.

This current is fed to a Fixed, high Value resistor. This in essence, sums the Current output from the R-2R network, and provides a Voltage output.

The outside world ( Hi Fi Chain after the DAC) needs a Voltage, so the voltage generated accross that high value resistor, needs to be Buffered ( Isolated ). That is the function of the DAC Buffer.

I did a quick Google and can recommend this link for some basic reading :

http://www.ikalogic.com/dac08.php

Amp_Nut schrieb:

This current is fed to a Fixed, high Value resistor. This in essence, sums the Current output from the R-2R network, and provides a Voltage output.

Amp_Nut,

This is the I/V conversion. While resistor is the passive approach, there many alternative approches for the same....including active option using opamps. Audio Note has a patent for doing this using a transformer (I do not think this transformer can be procured in the market for any DIY experiments).

However the voltage generated after the I/V conversion is still too low (only a few mV). Hence it needs to be amplified (to somewhere between 500mv to 2V)....this is where a tube buffer or an step-up transformer can help.

Selecting the value of this I/V resistor and design of the tube-buffer or step-up transformer has to be carefully designed so that it does not have any negative impact on DAC (chip) itself....

Regards
Siva.

sivat schrieb:

Forgot to mention another category here. Has anyone heard any DAC, that uses neither OPAMP or Tube...but just a transformer to step-up the analog signal from the DAC chip ?? Any thoughts, experience....

Sivat,
I heard one @ RMAF last year. It was done by my friend who owns Moon Audio (Drew Baird) in collaboration with another friend who owns K & K Audio. The K&K Audio person is the US importer & distrib of (Sweden's) Lundahl xformers. What Moon Audio did was not exactly what you are asking for but close - he modified the output stage of a Chinese CDP called ORIGINAL using xformer output stage.
Here is a link to some very nice pix from his website:
http://www.moon-audio.com/Original%20cd2008mkII.htm

Also there are links to 6moons review.


Now, re. the sound - for a grand sum of $1500 for a modified CDP IMHO I thought that the sound was excellent. Do keep in mind that the environment was that of a show & the room was a regular hotel room.
To Moon Audio's credit, the entire system was very modest -$1500 Shanling integrated amp & $4000 Selah Audio Gravina speakers - but had very, very good sound for the money. Lots of people were impressed.

FWIW.

Bombaywalla, thanks for the info and link.

Yeah, putting an output transformer at the CD player output is certainly an easier task than at the DAC Output!

with 16 bit resolution (1:65536) and a Max output from the DAC of close to 5 Volts peak ( the DC supply voltage), the Transformer will have to handle signals as small as 0.07 milli Volts and a Max of 5 Volts peak to peak

0.07 mV ( ie 70 micro Volts ) Volts peak is less than that of ( most / all ? ) moving Coil cartridges... a position in the Hi Fi Chain where Transformers ( some VERY Expensive... more Expensive than a CD player) have been used for decades.

Sivat, I agree with yr post, just a small point... the passive summing / I/V conversion is conceptually identical with an op Amp.

The Op amp's input impedance acts as a virtual summing / I/V conversion resistor, with the op amp providing the Current gain ( buffering) and / or Voltage gain too.

Amp_Nut schrieb:

Bombaywalla, thanks for the info and link. Yeah, putting an output transformer at the CD player output is certainly an easier task than at the DAC Output!

Amp_Nut,
your statement above confuses me - Moon Audio gutted the Original's output stage & replaced it w/ a xformer based output stage. (what I did not in my prev post was that the Original CDP's output is analog i.e. post D-->A conversion).
The output stage he gutted was the electronics following the DAC, which is exactly what sivat is asking about!!

So, how is using an xformer at the CDP output stage any different (from a electronics execution point of view) from that of a dedicated DAC chassis?
It should be the same thing, no?

Siva,
I recently got interested in the non oversampling DAC,since it offers more anolog sound than that of 24 bit/192KHz OS DAC's, per many online reviews,for opera consonance linear cdp's.(Since OS DAC's introduce some noise.)
http://www.sakurasystems.com/articles/Kusunoki.html
Very intersting article.
The Opera consonance line of CD players offer these NOS dacs,in a commercial,afforadble line of cdp's.
Again this is a debatable issue,since many will attest for improvement with OS dac while many will not distinguish the sound with OS.
No easy answers here..listening preferences,I believe.
Thanks.

Moon Audio gutted the Original's output stage & replaced it w/ a xformer based output stage.

Thanks for the clarification.... I did not interpret yr earlier post as that ...

I tried all of the following with the PCM1704

1> OPAMP (OPA627 for I/V and OP2134 doing the filter)

2> Tube buffer (with a 100ohms resistor acting as passive I/V convertor and a 6DJ8 doing the amplification job)

3> A 1+1:4+4 amorphous core transformer with a 10K I/V resistor at the secondary (roughly translating to about 28 ohms primary imepdance, which should keep the PCM1704 sufficiently happy)

As you can guess, OPAMP sounded the worst (really no comparision with the other two options) and the transformer sounding the best. With the transformer you need no additional circuit for filtering...

However there is a caveat with the transformer. Note that - i have connected a 10K resistor across the secondary of the transformer....and this will also be the impedance seen by the preamp/amplifier. Hence this might not work with all systems !!

Regards
Siva.

hifinovice1 schrieb:

Siva, I recently got interested in the non oversampling DAC,since it offers more anolog sound than that of 24 bit/192KHz OS DAC's, per many online reviews,for opera consonance linear cdp's.(Since OS DAC's introduce some noise.) http://www.sakurasystems.com/articles/Kusunoki.html Very intersting article. The Opera consonance line of CD players offer these NOS dacs,in a commercial,afforadble line of cdp's. Again this is a debatable issue,since many will attest for improvement with OS dac while many will not distinguish the sound with OS. No easy answers here..listening preferences,I believe. Thanks.

One thing i've learnt the hardway is NOT to believe in half-baked science and schema's available on the internet. They are mostly based on myth and not reality.

With respect to this particular philosophy (non oversampling), i'm not buying it yet and i have my own reservation. But nothing like facts...so i have decided to check it out for myself..

I'm in the process of assembling a new DAC based on TDA1541 using only a I2S interface (so as to avoid any filtering and conversion). Will let you know once i hear the difference between this and PCM1704 ..

Regards
Siva.

Hi Siva,

Many thanks for your report on the Transformer in the DAC.

VERY interesting...

Thanks Siva,keep us posted.
I'm interested in this OS vs NOS,but again,I believe in how it's implemented in the end,that matters more than the technology itself.
24 bit/192KHz has been highjacked by the marketing people to suggest,it cures everything(many things)digital.

siva, was reading up on the audionote DAC1x

he Dac One.1x features Audio Note's Direct from disc 1x oversampling circuit, Crystal CS8414CS input chip and hand wound copper wired Audio Note™ digital input transformer on a spun nickel core. Musical without sounding mushy. In excellent condition.

it appears this is a design philospdhy in AN dacs..is this what you were mentioning ? or is this just a transformer coupled output stage ?

Arj schrieb:

siva, was reading up on the audionote DAC1x he Dac One.1x features Audio Note's Direct from disc 1x oversampling circuit, Crystal CS8414CS input chip and hand wound copper wired Audio Note™ digital input transformer on a spun nickel core. Musical without sounding mushy. In excellent condition. it appears this is a design philospdhy in AN dacs..is this what you were mentioning ? or is this just a transformer coupled output stage ?

Digital Input transformer is used to receive the digital signal from the transport. It stands between the transport and the "Digital receiver" circuity on the DAC.

It can be used for SPDIF ...and it is more or less mandatory for AES/EBU. It helps in both case.

I use one in my DAC....so does Reimyo. Sometimes this transformer is connected in the transport itself...i think some AN transport has it on thier digital O/P (not very sure though !!)

The quality of this transformer is very important....especially the core....basically of high inductance properties (magnetic).

Regards
Siva.

PCM1704 into 28 ohms? Happy? Looks like something wrong with your calculations? It specifies 1K or better? Maybe 28 kohms?

In opamps, much as with tubes, much depends on implementation. Due to the very high open-loop gain, supply decoupling bypassing becomes critical. More so with the 627, which is one of the highest bandwidth 'for audio' devices available.

I'm not going to fault an implementation I haven't seen. However one would like to believe that your supply was the best it could have been, with proper groundplanes and clean supplies.

I would still suggest that the LME49710/20 are worth a spin, and a 10nF cap right across the pins or better, an C0G SMD just on the other side of the PCB along with a good groundplane and a squeaky clean supply (try a super-regulator with a noise shunt). I do not necessarily agree with your choice of opamps, btw. They are difficult to implement, and have a very high input impedance so parasitic reactances are an issue. I know they are reference design parts for the PCM, but I don't know that they are the best.

The only good thing about them is the distortion profile, specially at low output levels. The LME series has a higher distortion at lower output levels, but if you're doing I/V and output filtering only you're anyway running close to 2 volts off the PCM. At that level they should better the distortion figures of the BB opamps. As always, it comes down to the implementation, and IME you wouldn't have those problems as you're definitely not doing it like a small-time hobbyist on a perfboard.

Also (one of the things that is not easy to say) is that you really need to keep subjective preferences out of the equation. Analog warmth is sometimes distortion and digital glare is sometimes detail. It depends on your specific preferences, as (something I keep saying) we are never able to capture the real performance. Never on a 16-bit CD anyway, so it's all illusion.

In the absence of a real parameter then, the only way is to measure, and a good distortion analyser is handy. RMAA is also free, and only a PC is needed.

Transformer I/V can have very good results if the transformer chosen is of the right quality (<1dB insertion loss up to 2x the desired cutoff frequency). I would imagine the best result for this project would be a discrete opamp.

There, I had to throw that one in

If you're building a prototype for a commercial product, look at a solution you can support on a long-term basis. If it's for fun, build and keep for three weeks each, then you find which is the best to live with. All the best with whatever you choose!

http://www.tentlabs.com/Products/diycd/index.html

Check the section on "IV converter and output stage"

On other aspects...thnks for the advice !!.

Regards
Siva.

Sorry, we're talking current output, so lower is better - should've thought of that.

All the best with your project.

Edit: The TentClock is an amazing upgrade, highly recommended for mid-fi CD players. Guido had spent a long time discussing this over at DIYaudio. I have no CD players so can't verify this. But they're not expensive, and they've evolved into the XO2 and XO3 AFAIK.

I have no CD players so can't verify this.

What is your music source ?

[OT] PC but no DAC, just a cheap sound card. I got rid of my last CD player and equipment a couple of years ago. Now I just listen to a pair of computer speakers through a line out. [/OT]

eric_clapton schrieb:

[OT] PC but no DAC, just a cheap sound card. I got rid of my last CD player and equipment a couple of years ago. Now I just listen to a pair of computer speakers through a line out. [/OT]

oh boy!!
you have taken QUITE a turn Eric_clapton!
Guitar concerts taking too much of you time? Spending too much time @ your Crossroads rehab center?

From the info you have been providing you seem to be someone who has gotten his (I *think* that you are male) fingers dirty with DIY electronics. The info seems to be coming from your personal experience.

Using a PC these days is quite the done thing but listening to computer speakers......
Don't you loose all the dynamics?

Pray why not PC + cheap sound card + integrated amp + bookshelf/floorstanding speakers?

bombaywalla schrieb:

Using a PC these days is quite the done thing but listening to computer speakers...... Don't you loose all the dynamics?

Looks like things are changing. From what I've heard, the Audioengine A5 active speakers are pretty darn good and make the perfect companion to the cheap (8k) ESI Juli@ sound card. Their smaller sibling, the A2, got quite a reception from Stereophile and is available locally for about 10k.

particleman schrieb:

bombaywalla schrieb: Using a PC these days is quite the done thing but listening to computer speakers...... Don't you loose all the dynamics? Looks like things are changing. From what I've heard, the Audioengine A5 active speakers are pretty darn good and make the perfect companion to the cheap (8k) ESI Juli@ sound card. Their smaller sibling, the A2, got quite a reception from Stereophile and is available locally for about 10k.

Particleman,
thanks for the link.
Boy, are things changing or what???!!
there's a whole industry cropping up to service the computer audio enthusiast.

bombaywalla schrieb:

Pray why not PC + cheap sound card + integrated amp + bookshelf/floorstanding speakers?

I prefer live music actually.

I had two sets of floorstanders and self-built amps, a bunch of sources (before the CD days, once CDs came out I did have an entry-level NAD player). I've been in DIY for a long time (about 15 years now), started with low-end material and worked my way up to more complex designs, to the point that I could do a few projects for friends.

Like I said, I got rid of all of of my personal equipment. I developed a very bad hearing defect in my right ear, so I can't hear very well now. I'm not deaf, but my balance perception is totally off. An audiologist confirmed that my right is about 8 dB off the left above 2KHz, and there's a ringing in the bad ear that never really goes away. With hearing like that, anything more than my cheap little Altecs is overkill. You guys don't know how lucky you are to have two fully functional ears. Don't take it for granted.

I play guitar myself - obviously my hero is the *real* Eric Clapton, and I have a singing partner with a voice from heaven. I listen to others (commercial artists) only for inspiration, for which what I have work fine for me. And I love concerts and club gigs, attend every which one I can. Mostly though it's a just a gang of friends sitting and jamming up with a few guitars and a piano, a bluesharp and some small African instruments - a Djembe and a rain stick.

I still design and build some things in my spare time, and one of the reasons I keep talking about measurements is that I can't actually listen to 'air', 'detail', 'clarity', 'glare', 'separation'... If it's on a meter or scope, it's there, otherwise it's your imagination.

So that's my little story, what's yours? And to answer your question, I am male.

We're plenty OT now, poor sivat's discussion has been hijacked.

eric_clapton schrieb:

We're plenty OT now, poor sivat's discussion has been hijacked.

exactly!
Hence I'll keep my reply brief to minimize the hijack.

eric_clapton schrieb:

I prefer live music actually.

I do too but you seem to be more successful in partaking in it. Domestic situation won't allow so much into "luxury" items like this one.

eric_clapton schrieb:

I've been in DIY for a long time (about 15 years now), started with low-end material and worked my way up to more complex designs, to the point that I could do a few projects for friends.

nice!
I'd like to try too but am afraid that I will not be able to give it the attention it deserves. Maybe someday??
Thus far I've been buying commercially available equipment. I often have it modified for better performance. Over the years I've become more choosy as I discover each time that fewer & fewer manuf really make audio gear to service the music. One day there might be nobody that I can afford & that might be the day I commence DIY!
A friend who builds audio gear has tried to "shame" me into DIY. << this fellow knows me well so I take his ribbing in the correct spirit. >> He has also offered to teach me how to. He definitely has planted a seed in my mind. Seeing that the manuf are deviating from their original focus, DIY might be the only way.....

eric_clapton schrieb:

I developed a very bad hearing defect in my right ear, so I can't hear very well now. I'm not deaf, but my balance perception is totally off. An audiologist confirmed that my right is about 8 dB off the left above 2KHz, and there's a ringing in the bad ear that never really goes away. With hearing like that, anything more than my cheap little Altecs is overkill. You guys don't know how lucky you are to have two fully functional ears. Don't take it for granted.

sorry to read about your hearing. indeed I do take my hearing for granted - tends to happen when one can wake up each day & hear OK in both ears. Your caution is well taken.

eric_clapton schrieb:

I play guitar myself - obviously my hero is the *real* Eric Clapton, and I have a singing partner with a voice from heaven. I listen to others (commercial artists) only for inspiration, for which what I have work fine for me. And I love concerts and club gigs, attend every which one I can. Mostly though it's a just a gang of friends sitting and jamming up with a few guitars and a piano, a bluesharp and some small African instruments - a Djembe and a rain stick.

nice!
I have a few friends who play guitar & I like hearing them jam.
I have brother-in-law who plays the violin for a living. he is (obviously) very proficient at that instrument. Nice sounds come off the violin when he plays.

eric_clapton schrieb:

I still design and build some things in my spare time, and one of the reasons I keep talking about measurements is that I can't actually listen to 'air', 'detail', 'clarity', 'glare', 'separation'... If it's on a meter or scope, it's there, otherwise it's your imagination.

I know what you mean. However, the measurements are surrogates for the audio adjectives you wrote.


OK, we'll try to get back to Sivat's original thread.....

I was very sceptical when i started with the TDA1541A project. I was thinking that it will have the sound of those old Sony CD players (big ES series).

However, i was very suprised. The music is far more harmoniously correct...than the PCM1704.

The difference could be because of two very important reasons.

1) The simpler algorithm used in TDA1541A, compared to PCM1704

2) lack of "digital" filtering/processing/conversions that takes place in an external DAC (the PCM1704 was used in a external DAC configuration with filtering turned on !!).

If you start with Music and arrive at your equipment...then the harmonic correctness of a TDA1541A based solution (Integrated CD Player with I2S interface between transport and DAC) is a good bet.

However if you are the type who start by breaking down the sound to anaylze the performance and then arrive at Music....then the PCM1704 seems have a slightly better (micro) dynamics...hence apprears to have slightly more detailed. Ofcouse there is far less distrorion with PCM1704. (But then ...the tubes also have much more distortion than thier solid state peers)

Another significant difference is the high frequencies. The TDA1541A ... sounds unbelievably natural. The "ssssshhh" sound of louis armstrong sounds far more natural that i've ever heard before !!






Before someone jumps one me and says how ugly this thing looks.......i'm yet to trim and dress the wirings. Excuse : I still need to play around with the player....and a lot more R&D is still pending.


Regards
Siva.

congrats Siva...



I personally have been listening to scott Nixon's DAC and love the NOS Phillips TDA 1543A. (also uses the I2S interface..which is probably one of the better busses around for digital audio).

Its a brilliant DAC (a single DAC as opposed to the stacked ones found in the LiteDAC) and i actually preferred it over the Bel Canto DAC2.

Will soon be getting my hands on my DIY DAC from reignohchaos which uses the PCM1704.

i am now fearful that I will end up preferring the Scott Nixon NOS DAC over it.

the inherent roll off in the NOS DAC is key to the way it sounds is my perception. What do you think?

The DIY 1704 uses OPA627s for the filtering and output stages, but the voltage regulators (for the opamps and the interpolating filter DF1704) are kinda noisy which i think contributes to the bright sound that reignofchaos reported when he heard it.
would love to hear your DAC sometime...do keep us posted on further developments and a more in depth report as to how it sounds.

_savyasaachi

Savyasaachi schrieb:

I personally have been listening to scott Nixon's DAC and love the NOS Phillips TDA 1543A. (also uses the I2S interface..which is probably one of the better busses around for digital audio). Its a brilliant DAC (a single DAC as opposed to the stacked ones found in the LiteDAC) and i actually preferred it over the Bel Canto DAC2.

it's a brilliant DAC, I agree!!!
I found this to be a superb find - a gold nugget in the audio world. I find the sound thru it to be effortless & natural (much like what Siva reported).

Savyasaachi schrieb:

the inherent roll off in the NOS DAC is key to the way it sounds is my perception. What do you think?

The roll-off is due to the 44.1kHz signal itself...and the NOS DAC simply does not do anything about it. In the analgoue stage, post DAC, u need to handle this through a filter. This filter will impact the transient behaviour....so has to be designed properly.

The modern DACs take a different approach to handle this (Oversampling, upsampling)...and some says that this is adding "false information"

I somehow missed this query when i read this last week...sorry for the delay in response.

Just thought i'll give an update..

After a lot of R&D and auditions, my final DAC is PCM1704 with reclocking and using a transformer in the I/V Stage. The alternatives i tried are TDA1541A (no oversampling) and AD1865 (with SM5813 digital filter). All the DACs were tried with the same output stage using 6DJ8.

Musically, i think PCM1704 is as satisfying as the other older DACS. However, the lower noise level and solid-control over the sound (excellent transients) makes this the winner.

I'm going go back on my earlier comments .....more the experiments, more we learn. For now, i've capped this project...so my opinion might not change again for a while

Very Informative.

Thanks for sharing.

I am waiting for the full differential 'Buffalo' DAC from Twisted Pear Audio.

Amp_Nut schrieb:

'Buffalo'

Nice name...but why ??

No Idea !

But as a techie, you may find the tech info on the DAC chio VERY intyeresting... its also probably the best DAC Spec chip..

Their home page also gives links to the DIY forum discussions on this Chip.

http://www.twistedpearaudio.com

They also plan to offer a Digital Volume Control for the DAC.

Dont miss the relatively low price ..

I am witing for them to re-stock. I could not order their 1st batch since I was on holiday, and no one to take delivery...

sivat schrieb:

Just thought i'll give an update.. After a lot of R&D and auditions, my final DAC is PCM1704 with reclocking and using a transformer in the I/V Stage. The alternatives i tried are TDA1541A (no oversampling) and AD1865 (with SM5813 digital filter). All the DACs were tried with the same output stage using 6DJ8. Musically, i think PCM1704 is as satisfying as the other older DACS. However, the lower noise level and solid-control over the sound (excellent transients) makes this the winner. I'm going go back on my earlier comments .....more the experiments, more we learn. For now, i've capped this project...so my opinion might not change again for a while :)

Siva,
good to know that you have liked the PCM1704 DAC chip. My Wadia uses the same DAC chip (I believe that the Wadia uses K-grade 1704 DACs).
I also really like the TDA1541A DAC chip in my Scott Nixon non-os DAC. I believe that these 2 DAC chips are classics & probably the best sounding DAC ICs still. IMHO, of course.

Hello Bombaywalla,

Where have U been ? Long Time No Post ???

Could you share your opinion on these 2 Very Diff DACs ?

Are you using a Wdia DAC or CD player ?

I would be interested in hearing from anyone who has heard a Wadia DAC. Is it very finiky about the transport that it is partnered with ?

Also anyone have experience with a USB DAC that yopu have Actually heard or used ?

Bombaywala,

Yes. PCM1704 is a good one. I will rate the AD1865 to be second best..and last the TDA1541A.

A lot depends on how the output stage and power supplies are designed.

The combination of transformer @ the I/V Stage and a tube o/p stage...does seem to do magic. Thanks to AudioNote

Amp_Nut,

Implementing a USB interface would be quite easy using the using PCM2704 or something like that. I'm sure the are "USB to SPDIF convertor" modules available online for a small price..


Regards
Siva.