Hi all
So I decided to build myself a DAC. First of all: this is my first DIY project of this kind.
After considering all the options I've come to the conclusion that... that I have no conclusion. Mind that a few weeks ago I haven't even thought about this, I was almost dead-set on buying a NOS filterless DAC w AD1865, so please forgive any really stupid stuff I may write. But long is boring so I'll skip to the subject.
1. I think I will go for NOS and filterless. Why? Just for the sake of it. Consider those "project requirements"
Nonetheless, further tweaks are not only welcome, but planned. As for others "requirements":
- no tube I/V or anything scary like that. No no no, DIY tubes is not for me (yet), I'm not ready to take that step.
- I'd like both I2S and S/PDIF input
- less than (roughly) 300$ in parts
- balanced would be nice, my (future amp) has balanced in
2. AD1865 is much praised but there are some naysayers. Considering that I have the possibility of listening to this DAC http://www.rockna-line.com/diy/dac/dac.htm I though that I should go for something else, even just for comparing. What chip would the gurus recommend?
3. I lack the expertise and the time (at least now) for "designing" my own. All my personal touch will be the choice of miscellaneous blocks (oscillator, I/V, power supply, etc). I'm calling for the gurus to help me.
- I was thinking about using the discrete opamp by Mr. Pass (D1) for I/V, just to find out that I'm not the first one. Any opinions on this?
- other thing is the clock. I know about the Guido Tent clock, the Kwak-clock, anything else I should find about?
- power supply - things I read about 78xx regulators make my skin crawl, any other better chips or discrete designs out there?
- what input S/PDIF decoder?
- anything else I might have missed
Last, but not least, please note that this is not one of those "what is the best ... out there?". All I want is to be pointed to the right direction by the more experienced, just get me going
Elso Kwak, Jean-Paul, Kuei Yang Wang? (I appologize to any guru not included in the list, also I appologize for including non-gurus by accident. just kidding )
Regards,
Calin
So I decided to build myself a DAC. First of all: this is my first DIY project of this kind.
After considering all the options I've come to the conclusion that... that I have no conclusion. Mind that a few weeks ago I haven't even thought about this, I was almost dead-set on buying a NOS filterless DAC w AD1865, so please forgive any really stupid stuff I may write. But long is boring so I'll skip to the subject.
1. I think I will go for NOS and filterless. Why? Just for the sake of it. Consider those "project requirements"
Nonetheless, further tweaks are not only welcome, but planned. As for others "requirements":- no tube I/V or anything scary like that. No no no, DIY tubes is not for me (yet), I'm not ready to take that step.
- I'd like both I2S and S/PDIF input
- less than (roughly) 300$ in parts
- balanced would be nice, my (future amp) has balanced in
2. AD1865 is much praised but there are some naysayers. Considering that I have the possibility of listening to this DAC http://www.rockna-line.com/diy/dac/dac.htm I though that I should go for something else, even just for comparing. What chip would the gurus recommend?
3. I lack the expertise and the time (at least now) for "designing" my own. All my personal touch will be the choice of miscellaneous blocks (oscillator, I/V, power supply, etc). I'm calling for the gurus to help me.
- I was thinking about using the discrete opamp by Mr. Pass (D1) for I/V, just to find out that I'm not the first one. Any opinions on this?
- other thing is the clock. I know about the Guido Tent clock, the Kwak-clock, anything else I should find about?
- power supply - things I read about 78xx regulators make my skin crawl, any other better chips or discrete designs out there?
- what input S/PDIF decoder?
- anything else I might have missed
Last, but not least, please note that this is not one of those "what is the best ... out there?". All I want is to be pointed to the right direction by the more experienced, just get me going
Elso Kwak, Jean-Paul, Kuei Yang Wang? (I appologize to any guru not included in the list, also I appologize for including non-gurus by accident. just kidding
)Regards,
Calin
It's fashionable. It's also rebelious. Be damned all engineering
But would it sound better to You? I was in similar position at some stage. Tried lot's and lot's of different configurations using the darling little 1543. Separate PS for everything; top quality parts. It only got better and better. Just not good enough. I know, i should have tried Thorsten's sinc filter before giving up.
It may be worth looking at something less revolutionary like the Rakk Dac - and it's right within your price range.
Have a look at my DIY-Site, I have projects described in detail on both non oversampling (8412-1543)as well oversampled dacs (PCM63/PCM1704-DF1704/SM5842). Also a solution for use with a PCM1738 with transformer output...
might help you decide....
good luck with your DAC !
doede
www.dddac.de
might help you decide....
good luck with your DAC !
doede
www.dddac.de
Take a look at Doede Douma's DDDAC1543: nonos - 8*TDA1543 DACs (no balanced outs I'm affraid) with Tent's 01 clock; resistor load as I/V. All for EU115. This is the starter kit to be fed with a battery. You can tweak it as much as you want and it sounds very, very good. (no digital artifact here)
Maybe with 16*TDA1543 and transformer output you can make balanced outs
Good luck
Mauricio
(no digital artifact here)Maybe with 16*TDA1543 and transformer output you can make balanced outs
Good luck
Mauricio
Yup the hype got to me. God knows that first time I found about that I was like "whaaaat?"analog_sa said:It's fashionable. It's also rebelious. Be damned all engineering
Dunno, AudioNote can't be all wrong.analog_sa said:
Ahh, let me think about this... Oops, I think I've blown a fuseMaybe with 16*TDA1543 and transformer output you can make balanced outs
Other option would be to give up on the balanced. This is not a must, but it sounds cool
And my amp has balanced in as said before.As for batteries. Nah, I never liked the idea. But that's not an integral part of the DAC so I guess it's not an issue.
Btw Mauricio, don't you have an UCD180 also? I remember seeing you on the UCD180 Q&A thread, if my memory isn't playing tricks on me. Excuse me if so. Anyway, I have 2 of those, that's the amp I'm going to use.
One other short question: why resistor I/V? Do you think that a more complex scheme is overkill?
And one of those questions from the "which is the best" series (can't help myself): which part of a DAC do you see as the most important? Clock, chip, output block?
Regards,
Calin
Hi mr_push_pull !
Good memory
My 36Ah battery will last 36hrs if I didn't recharge it
And is humm free
Appart for the fact that it reduces/eliminates digital artifacts.
Why don't you go active? I use a cheap Behringer active crossover that makes "servo-balanced outputs". I feed directly (without preamp) the amps with that balanced signal. (Someday I will have a HiEnd crossover). The dynamics of this configuration are overwelming (did I wrote this wright? )
How would I know?
Ask the wise. Ask Doede.
The philosophy is "simpler is better" everytime you add an active stage you introduce distortion. The best opamp is no opamp
The second best is OPA627, maybe.
Hey, I'm only beginning with it! So far I would say:
Output coupling capacitor make a huge difference. Non polar BG 22uF parallel Wima 0.47 uF made the sound very warm and relaxing, with good dynamics but lack of HF extension.
BG non polar 4.7uF in super-E-cap configuration (antiparallel) is very, very transparent and extended but takes eons to burn-in!!! This DAC is up to down detailed.
Add independent regulated supply for Tent clock (even with 7805, said to be very bad, but is the only one that I know how to use) made a better soundstaging (depth) and "presence"
Even wiring makes a change
Good luck, Calin.
Mauricio
Good memory
My 36Ah battery will last 36hrs if I didn't recharge it
And is humm free
Appart for the fact that it reduces/eliminates digital artifacts.
Why don't you go active? I use a cheap Behringer active crossover that makes "servo-balanced outputs". I feed directly (without preamp) the amps with that balanced signal. (Someday I will have a HiEnd crossover). The dynamics of this configuration are overwelming (did I wrote this wright?
)How would I know?
Ask the wise. Ask Doede.
The philosophy is "simpler is better" everytime you add an active stage you introduce distortion. The best opamp is no opamp
The second best is OPA627, maybe.Hey, I'm only beginning with it! So far I would say:
Output coupling capacitor make a huge difference. Non polar BG 22uF parallel Wima 0.47 uF made the sound very warm and relaxing, with good dynamics but lack of HF extension.
BG non polar 4.7uF in super-E-cap configuration (antiparallel) is very, very transparent and extended but takes eons to burn-in!!!
This DAC is up to down detailed.Add independent regulated supply for Tent clock (even with 7805, said to be very bad, but is the only one that I know how to use) made a better soundstaging (depth) and "presence"
Even wiring makes a change
Good luck, Calin.
Mauricio
I/V circuits ain't that scary
Anyway you could feed to an external preamp circuit the I/V signal. The preamp would be similar to a phono preamp except no RIAA equalization circuit. Your favorite circuit that would amplify a 100mV signal up to line levels, 1 to 2V is all you'd need. Could be tube, transistor or op-amp. Anyway, take a look at my web page on modifying the analog section of CD players . The preamp circuit need not be in the same box as the rest of the CD player. For someone starting with tubes, it would be easier to have the tube circuits inside a separate box.
Re: Hi mr_push_pull !
I forgot to thank Doede for his answer, excuse my rudeness. I'll look into it as soon as possible!
I know, I'm insane.
I use it with good results
The only objective opinions that I can collect are the ones based on experiment and comparison... but wait a minute, that's subjective. I guess it's as objective as subjective can get
Thanks for the link.
Later I will post some links I've come accross, maybe some could comment.
Best regards,
Calin Dorohoi
I forgot to thank Doede for his answer, excuse my rudeness. I'll look into it as soon as possible!
Would you believe me if I tould you that I read every single reply on that thread? Speaking about no-life freaks...maxlorenz said:Good memory
Don't get me wrong, but somewhere in my head lurks the idea that one should not depend on battery power. Makes me feel insecuremaxlorenz said:My 36Ah battery will last 36hrs if I didn't recharge it
And is humm free
Appart for the fact that it reduces/eliminates digital artifacts.
I know, I'm insane.Budget? One more thing is that it's probably not worth the effort, at least not now. I'll use 2-way DIY speakers w Morel drivers (somethink like Dynaudio Audience 42), I think that active is justified only for speakers with more that 2 ways. Btw, if I manage to finish the DAC I'll have a 100% DIY system. If the amp counts as 100% DIYmaxlorenz said:
Use this simple rule: when in doubt, there probably is an 'h' after the 'w'maxlorenz said:The dynamics of this configuration are overwelming (did I wrote this wright?
I use it with good results I'm asking everyonemaxlorenz said:How would I know?
Ask the wise. Ask Doede.
The philosophy is "simpler is better" everytime you add an active stage you introduce distortion. The best opamp is no opamp
The only objective opinions that I can collect are the ones based on experiment and comparison... but wait a minute, that's subjective. I guess it's as objective as subjective can get After all, does a decoupling transformer have to be expensive? What about DIY decoupling transformer?maxlorenz said:Hey, I'm only beginning with it! So far I would say:
Output coupling capacitor make a huge difference. Non polar BG 22uF parallel Wima 0.47 uF made the sound very warm and relaxing, with good dynamics but lack of HF extension.
BG non polar 4.7uF in super-E-cap configuration (antiparallel) is very, very transparent and extended but takes eons to burn-in!!!
Rod Elliott says on his page that the adjustable regulators are better. On the other hand, I'm sure there are plenty of better discrete schematics availablemaxlorenz said:Add independent regulated supply for Tent clock (even with 7805, said to be very bad, but is the only one that I know how to use) made a better soundstaging (depth) and "presence"
Even wiring makes a change
Thanks, keep in touch.maxlorenz said:
I guess I'll just jave to see and experiment. So many possibilities, so little moneywa2ise said:
I/V circuits ain't that scary
Anyway, take a look at my web page on modifying the analog section of CD players . The preamp circuit need not be in the same box as the rest of the CD player. For someone starting with tubes, it would be easier to have the tube circuits inside a separate box.
Thanks for the link.Later I will post some links I've come accross, maybe some could comment.
Best regards,
Calin Dorohoi
Excellent. More stuff to learn about
I think active speakers will allways sound better than passive: more controled/relaxed and better bottom extension. You can look at specs of comercial ones that come in this two options to verifie. Appart that my active crossover costs +/- US$120
and is tweakable. Other option: conserve your 2 ways as they are and build two big woofers two complete a fullrange system. I made mine with Eminence delta15 and 5-7 cm real wood panels, trapezoid shape
, all for US$180.About I/V, I forgot to mention that DDDAC1543's 8*DAC tower drives enough current to kill a mouse
That's why you only need a resistor load to choose your desired Vout (and taste, through different R types). The kit's stock max Vout is 1.6V
Did I explained it well, Mr Douma?
Me too!
I live in DIY/audiophile isolation down here
Thanks God, this is the best Hobby in the world
and the forum is great.Arrivederci
Mauricio
well said Mauricio, although my DAC is not meant to kill any living creatures
I have used several output stages, including opamps and transformers, but with a simple resistor always convinced me ...
doede
www.dddac.de
Hi Calin
Concerning voltage regulators, have a look at this...
Voltage regs....
I'm certainly not a pro, but learned a lot from this site's pages.
As the aspect of why one regulator would by better than any other ?
Consider this : every power pin on a dac, opamp, even discrete transistor circuit can be regarded as an additional input. Most opamps do very well in amplifying PS noise above a few kHz...
HTH
A8
Concerning voltage regulators, have a look at this...
Voltage regs....
I'm certainly not a pro, but learned a lot from this site's pages.
As the aspect of why one regulator would by better than any other ?
Consider this : every power pin on a dac, opamp, even discrete transistor circuit can be regarded as an additional input. Most opamps do very well in amplifying PS noise above a few kHz...
HTH
A8
16, no less
Well, I built my 16*1543 DAC tower yesterday, for my DDDAC1543
I had to set again R load and bias R to obtain the 3.85VDC desired. For R load I choose 120R to match better with my active CX wich I use as preamp. The bias R is a multiturn variable resistor and it was difficult to settle this time. Maybe I'm on the limit of it. I also swapped a few wires to solid silver wich usually takes long to burn-in.
Well, it worked at the first attempt I think my beginner's luck doesn't want to live me
It's too late to tell but I noticed even more relaxed and natural midrange, with better soundstaging and instrument definition.
What's next? I have to change the stock R to Riken-Ohm and must explore super-regulators for independent supplies.
As I said, Doede's DDDAC1543 starter kit is a great project: there's no relation between the price paid and the high sonic quality obtained.
Now, I'm expecting to build a CDPRO-2M transport wich probably will include a direct I2S connection to a TDA1543 tower.
Good luck to everybody and enjoy!
Mauricio
Well, I built my 16*1543 DAC tower yesterday, for my DDDAC1543
I had to set again R load and bias R to obtain the 3.85VDC desired. For R load I choose 120R to match better with my active CX wich I use as preamp. The bias R is a multiturn variable resistor and it was difficult to settle this time. Maybe I'm on the limit of it. I also swapped a few wires to solid silver wich usually takes long to burn-in.
Well, it worked at the first attempt
I think my beginner's luck doesn't want to live me It's too late to tell but I noticed even more relaxed and natural midrange, with better soundstaging and instrument definition.
What's next? I have to change the stock R to Riken-Ohm and must explore super-regulators for independent supplies.
As I said, Doede's DDDAC1543 starter kit is a great project: there's no relation between the price paid and the high sonic quality obtained.
Now, I'm expecting to build a CDPRO-2M transport wich probably will include a direct I2S connection to a TDA1543 tower.
Good luck to everybody and enjoy!
Mauricio
Re: 16, no less
I'm just curious, don't you resistor I/V afficianados find sonic degradation due to the high output voltage on the DAC? The 1543 is spec'd at only 25mV max voltage compliance, right? I haven't done any tests on it, but on the Burr Brown PCM series with 1.2 mA output I found that the sound got better and better the lower the voltage got, until I was down to 10 ohms, which would require way too much gain, so decided at that point it was better to go with a current mirror since I could easily get below 10 ohms impedance with no feedback and it sounded quite a bit better than my previous 100 ohm resistor and more complex discrete gain stage. And much better than any of the opamp circuits, like the AD811. But I guess it's all about tradeoffs and what works best in your system. Looks like a fun project.
Probably be a good topic for a separate thread, but maybe you've all hashed it out ad nauseum in the past.
maxlorenz said:Well, I built my 16*1543 DAC tower yesterday, for my DDDAC1543
I had to set again R load and bias R to obtain the 3.85VDC desired. For R load I choose 120R to match better with my active CX wich I use as preamp.
I'm just curious, don't you resistor I/V afficianados find sonic degradation due to the high output voltage on the DAC? The 1543 is spec'd at only 25mV max voltage compliance, right? I haven't done any tests on it, but on the Burr Brown PCM series with 1.2 mA output I found that the sound got better and better the lower the voltage got, until I was down to 10 ohms, which would require way too much gain, so decided at that point it was better to go with a current mirror since I could easily get below 10 ohms impedance with no feedback and it sounded quite a bit better than my previous 100 ohm resistor and more complex discrete gain stage. And much better than any of the opamp circuits, like the AD811. But I guess it's all about tradeoffs and what works best in your system. Looks like a fun project.
Probably be a good topic for a separate thread, but maybe you've all hashed it out ad nauseum in the past.
Hi Black Heart:
I hope your nickname does not reflect your mood tonight
Where did you find that number? Never mind, because I'm unable to answer you in a technical way
I can asure you that it sounds very good, whatever may this mean to you.
Maybe having 8 or 16 DAC's in parallel limits the drawback you mention. My max Vout must be around 1.4V now.
Help needed, Doede
I hope your nickname does not reflect your mood tonight
Where did you find that number? Never mind, because I'm unable to answer you in a technical way
I can asure you that it sounds very good, whatever may this mean to you.
Maybe having 8 or 16 DAC's in parallel limits the drawback you mention. My max Vout must be around 1.4V now.
Help needed, Doede
Re: Hi Black Heart:
Not really, but I am listening to a CD that a friend recommended by Matt Elliott called The Mess We Made and is is a little on the melancholy side, but that's usually the way I like it. One of my favorite bands is The Black Heart Procession, so I just decided to appropriate part of their name for my nickname here. Might have to play one of their albums next, most likely the one simply titled '2'. Great stuff
I remembered it from working some with the TDA1541 a long time ago, but just checked and it is in the datasheet for the 1543, page 7 in the one I just downloaded. I'll attach that part of the pdf below but if it doesn't come out for some reason just go to page 7 at http://www.vlsi.informatik.tu-darmstadt.de/student_area/sep/tda1543.pdf ...
maxlorenz said:I hope your nickname does not reflect your mood tonight
Not really, but I am listening to a CD that a friend recommended by Matt Elliott called The Mess We Made and is is a little on the melancholy side, but that's usually the way I like it. One of my favorite bands is The Black Heart Procession, so I just decided to appropriate part of their name for my nickname here. Might have to play one of their albums next, most likely the one simply titled '2'. Great stuff
Where did you find that number? Never mind, because I'm unable to answer you in a technical way
I remembered it from working some with the TDA1541 a long time ago, but just checked and it is in the datasheet for the 1543, page 7 in the one I just downloaded. I'll attach that part of the pdf below but if it doesn't come out for some reason just go to page 7 at http://www.vlsi.informatik.tu-darmstadt.de/student_area/sep/tda1543.pdf ...
An externally hosted image should be here but it was not working when we last tested it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.