Building the ultimate NOS DAC using TDA1541A

Hi ecdesigns,

Greetings!

I am new to the forum and was looking for a word on DACs, ESS Sabre, Wolfson, Bur Brown. Which is the best to be used with a Vacuum Tube Amplifier?

I couldn't read all the 500 pages but I did go through your first post, what a pleasre to see the DAC's finished picture. I would like to ask, what is the cost to build this DAC project and is there any kit available for this unit?

I have never had a home audio DAC before, I only have experienced a Bur Brown 24/192 DAC in my car.

Thanks for your support.

Regards
LR
 
After a while, I got used to the SD cards...16GB for the lazy. :D

Dear Youknowyou,

It seems you are in position to tell us how my future Mosaic will sound :cool:

Cheers,
M.
:), Im still waiting for my mosaic dac. will report back once received. I cant wait!

The Mosaic DAC can now directly drive a beyerdynamic DT-990 premium studio Headphone with 250 Ohm impedance.

Now we basically listen to the music produced by electronic switches and a passive resistor matrix. Then one hears how much distortion and degrading is introduced by active circuits like (pre) amplifiers.
Wow, so that means we wont need a headphone amp anymore?
 
Last edited:
Hi ecdesigns,

Greetings!

I am new to the forum and was looking for a word on DACs, ESS Sabre, Wolfson, Bur Brown. Which is the best to be used with a Vacuum Tube Amplifier?

I couldn't read all the 500 pages but I did go through your first post, what a pleasre to see the DAC's finished picture. I would like to ask, what is the cost to build this DAC project and is there any kit available for this unit?

I have never had a home audio DAC before, I only have experienced a Bur Brown 24/192 DAC in my car.

Thanks for your support.

Regards
LR


I have read these tread troughout 3-4 times!!! I have learned a lot and improved many of my cd player to a decent level so I can enjoy listening them.

I advice you to do the same.
Since only you who knows what change you desire or can be build by you.
You can try out different mods, upgrade step by step and learn.

Regards D
 
ecdesigns, congratulations for your work.

I just found out by accident this threat is full of TDA1543 discussions and to be honest, I'm very glad. I'm a big fan of this chip.

A few days ago, I also warmed the hearts (and ears) of several audiophiles with much experience, who were stunned by natural, liquid sound this chip gives.

My DAC is a bit on the extreme and funny side. 16 paralleled chips immersed in paraffin oil for lazy cooling. Shunt Salas regulators, choke input power supplies, neotech single crystal copper wire, XMOS USB board, pine chassis with grain oriented, Neotech 3PS25 power cable with Valab plugs, Neotech rhodium RCAs. Today I tried putting Duelund copper foil caps on the output

attachment.php


There are a few things that don't give me a break.

-I read more and more than a single chip sounds better than paralleled chips. I am willing to try impatiently . Can you define more precisely? If this DAC has an issue, is it's lack of a wider soundstage
-The chip temperature is very audible. The problem with my implementation is that, the oil heats very much and the sound gets somehow slow, lacks engagement, lacks punch and sounds weirdy. Then if I drain the hot oil and pour a cool batch, first the sound is sterile, but after a minute it gets amazing! And after 10-15 minutes it starts degrading.
-So it will be easier for me to try 1 chip first.. and the cooling will be easier anyway.
 

Attachments

  • 11800563_1025457354154505_5625173758016880204_n.jpg
    11800563_1025457354154505_5625173758016880204_n.jpg
    84.5 KB · Views: 2,153
Lord Raven
Your question is a tough one to answer. A little like asking what can I eat for a great lunch :)

I think you have come to one of several great spots.

Some would say that the TDA1541a cannot be beat if you listen to 44.1k music. NOS of course. It is also quite good on hires when well done.

Some like the ess chips for hi res. Others claim they are dry and lifeless. Cannot comment. I have never heard one. I am in the 1541a camp.

Well worth looking also at DDDAC home page for ideas.

If you settle on 1541a, then this is a definitive source of info.

You likely could just buy John aka ecdesign's DAC and be in audio heaven for a long time.

Don't miss Ian Canada's site. You are in for a cornucopia of great ideas to keep you busy for a very long time here at DIYAUDIO.

Have fun.
 
Hi 50AE,

A few days ago, I also warmed the hearts (and ears) of several audiophiles with much experience, who were stunned by natural, liquid sound this chip gives.

One of the issues with DAC chips and connected circuits is phase cancellation. Following youtube video illustrates this effect using microphones:

https://www.youtube.com/watch?v=_N0ER4A73QE

Similar happens in electronic circuits. The phase of multiple signals reproduced simultaneously gets distorted and this in turn can cause clearly audible distortion despite low THD.

The simple, straight forward setup of the TDA1543 chip results in lower phase errors compared to more complex chips like the TDA1541A, PCM1704 or SABRE for example.

This is why a $1 vintage budget chip can still offer more involving sound than a modern, state of the art DAC chip.

It turns out to be extremely difficult to come anywhere near vinyl & tape sound using digital audio. The first CD player was introduced in 1982 and we are still waiting for that “analogue” sound from digital audio.

Vinyl is becoming popular again, perhaps as a result of the disappointing sound quality of very expensive, state of the art digital audio equipment:

Vinyl revival: is it back for good? | Bob Stanley | Comment is free | The Guardian

And turntable development has made progress too:

Laser turntables - when lasers get groovy - E & T Magazine


My DAC is a bit on the extreme and funny side. 16 paralleled chips immersed in paraffin oil for lazy cooling

Paralleling DAC chips increases interference levels & jitter, possible advantage is averaging between repeating bit error paterns that are typical for R2R and related multibit DAC chips. Most important thing with DACs based on emitter scaling is that all on-chip transistors remain at the same constant temperature. Immersing DAC chips in paraffin oil could possibly help to keep chip temperature more constant.

Shunt Salas regulators

Voltage regulators are -in- the signal path and therefore have major impact on sound quality. Every voltage regulator “sounds” different as it results in a unique distortion and noise spectrum. Aim is a low, linear impedance with largest possible bandwidth. This requires extremely fast (global) feedback loops in the voltage regulator. Most integrated voltage regulators have slow, instable global feedback loop and resulting sound is horrible (thin and distorted).

choke input power supplies

One way to attenuate ever increasing mains interference is using chokes in the power supply. This was already used long time ago with tube radios but was replaced by modern equivalents that don’t work properly. The new balanced mains power supply that powers the Mosaic DAC is based on two large chokes the size of the mains transformer for pre- filtering +9V and -9V supplies.

neotech single crystal copper wire

Copper wire is copper wire, exotic copper wire has no audible advantages over plain copper wire. OFC (copper with low oxigen content) won’t corrode from the inside out over time so it offers better long-term performance. New (not oxidised) plain copper wire and OFC copper wire sound exactly the same. Copper oxide has semiconductor properties and this might cause problems. Remedy is using tinned, silver plated, or lacquered copper wire strands.

pine chassis with grain oriented

We live in the most EMI polluted environment ever and it is getting worse fast. So it would be wise to place your DAC in a Faraday’s cage, not on a piece of wood.

Neotech 3PS25 power cable with Valab plugs

It makes absolutely no sense to hook up some exotic mains cable at the end of 10 … 100 meter plain copper wire of the mains wiring in the house. It would be wiser to run a direct connection to the power entry point in your house and use it for audio equipment exclusively.


Neotech rhodium RCAs.

RCA interlinks are unsuitable for high resolution playback because of Eddy current losses in the bulky brass sockets and plugs. The bullet plug offers some improvement but the RCA sockets in the audio equipment still introduces large Eddy current losses. It is better to use audio interlinks with lower mass (lower electrical resistance).

In the good old days we used DIN interlinks that already had low Eddy current losses, but it seems engineers have forgotten about that.

However we can use mini XLR (closest match to DIN interlinks) or 3.5mm Jack interlinks for example as these have low Eddy current losses.

It is also well known that thinner wires within the interlinks offer better sound quality as a result of lower Eddy current losses.

Today I tried putting Duelund copper foil caps on the output

Coupling caps introduce unacceptable distortion. I expermented with many different coupling caps including Sonicap teflon, V-cap TFTF and Duelund copper foil. The best coupling cap is no coupling cap or DC-coupling. The Mosaic DACs now run on balanced power supplies and have fully DC-coupled output with true GND (0V) as GND reference.
 

Copper wire is copper wire, exotic copper wire has no audible advantages over plain copper wire.

My listening experience tells me something different. Some different copper wires sound differently.


It makes absolutely no sense to hook up some exotic mains cable at the end of 10 … 100 meter plain copper wire of the mains wiring in the house.

As much as it looks to make no sense, my listening experience tells me something different.
 
Copper wire is copper wire, exotic copper wire has no audible advantages over plain copper wire. OFC (copper with low oxigen content) won’t corrode from the inside out over time so it offers better long-term performance. New (not oxidised) plain copper wire and OFC copper wire sound exactly the same. Copper oxide has semiconductor properties and this might cause problems. Remedy is using tinned, silver plated, or lacquered copper wire strands.

...

It makes absolutely no sense to hook up some exotic mains cable at the end of 10 … 100 meter plain copper wire of the mains wiring in the house. It would be wiser to run a direct connection to the power entry point in your house and use it for audio equipment exclusively.

dear john,

you know, i have been testing and building many mains as well as interconnect cable during the past years. my findings are that there is big difference between "plane" materials , oxygen-free, large crystalline and mono crystalline materials, regardless copper or silver. the grain size and the interface (impurity at the grain boundaries) has a big impact on the sound quality perception, at least for my ears listing to well tuned systems. these differences are very noticable for all kind of (wired) compontens such as chocks, transformers, etc.

regarding the main cables, i am using a dedicated direct mains connection with so called high-end activated fuse and wall cable ( what ever that means :) ) , but still find big differences comparing different built mains cables. my experience is that the last 1,5m ( mains cable) is very important; i do not have any scientifically and technically proven explanation for that but my feeling is that maybe the "reflection" of the audio signal back in the mains cable or/and the modulation of the signal causes this issue.

please consider my above comments only as just reporting my findings and not objecting your knowledge ! i am always a very big fan with high respect for your excellent works! thanks again!

best wishes
 
dear john,

you know, i have been testing and building many mains as well as interconnect cable during the past years. my findings are that there is big difference between "plane" materials , oxygen-free, large crystalline and mono crystalline materials, regardless copper or silver. the grain size and the interface (impurity at the grain boundaries) has a big impact on the sound quality perception, at least for my ears listing to well tuned systems. these differences are very noticable for all kind of (wired) compontens such as chocks, transformers, etc.

regarding the main cables, i am using a dedicated direct mains connection with so called high-end activated fuse and wall cable ( what ever that means :) ) , but still find big differences comparing different built mains cables. my experience is that the last 1,5m ( mains cable) is very important; i do not have any scientifically and technically proven explanation for that but my feeling is that maybe the "reflection" of the audio signal back in the mains cable or/and the modulation of the signal causes this issue.

please consider my above comments only as just reporting my findings and not objecting your knowledge ! i am always a very big fan with high respect for your excellent works! thanks again!

best wishes

Yes, we are big fans of dear -ECdesigns- but some of us would disagree on this point. I try to use OCC copper wire all around, and some times would use a couple of cm of pure silver wire on digital signals, to great subjective effect.
About mains, I ordered a decade ago some "balanced power transformers" with bifillar secondaries to a local technician, where I connect all my gear from primary and secondary systems. Recently I found that using a smaller "balanced power Tx" (BPTx) for the SD card player and this connected in series with the main BPTx, which powers the amplifier, made also a significant improvement in lowering the noise floor. :cool: even if the SD card player has not balanced power supply...

Some things are still mysterious. :)

About DC/AC coupling, while I wait for my experiments with the recommended direct DC output from John, I think I have a cheap and worthy solution, thanks to the old USSR efforts: I use 22uF PETP//1uF PETP//0.25uF Polystyrene, surplus millitary caps. I will add a newly found 1uF polystyrene to the bunch. Looks bad but sounds great!


Best wishes to all,
M.
 
Last edited:
Yes, we are big fans of dear -ECdesigns- but some of us would disagree on this point. I try to use OCC copper wire all around, and some times would use a couple of cm of pure silver wire on digital signals, to great subjective effect.
About mains, I ordered a decade ago some "balanced power transformers" with bifillar secondaries to a local technician, where I connect all my gear from primary and secondary systems. Recently I found that using a smaller "balanced power Tx" (BPTx) for the SD card player and this connected in series with the main BPTx, which powers the amplifier, made also a significant improvement in lowering the noise floor. :cool: even if the SD card player has not balanced power supply...

Some things are still mysterious. :)

Best wishes to all,
M.

if you are also infected by the silver wire virus like me :) , then your should surely try mono crystalline silver wire and you will never want to use anything else :) . too sad that the price which i got last time for the 0.4mm was 150 USD/meter :-((
ps. john is a really busy guy but promised to visit me this summer (that is now almost over :-( ) then i will demonstrate him differences between cables :) .... if not i send him a shunyata anaconda mains cable to test ;-)
 
What is all this absolute rubbish about "eddy current loss" in RCA connectors ??
The currents involved are absolutely miniscule and cause no discernible loss of anything.
This thinking belongs in the snake oil category. I hope you have not succumbed to Bybee
and his bogus products !!!!!!

It is very impolite to enter here with these awful manners and treat the statements of the thread starter as "absolute rubbish".
Not to mention that -EC- knows a bit about what he is talking.
I imagine that you can recognize these facts...

Yours truly,
M.
 
I'd like to start a discussion here about the decoupling capacitors used on the three power supply pins of the TDA1541(A), +5v (pin 28), -5v (pin 26) and -15v (pin 15).

I have been experimenting with different types and values of capacitor on each of these pins and they seem to have a very pronounced effect on the sound.

I have read that the -15v is pure analogue but the +5v and -5v supplies are both analogue and digital.

For the -15v supply I prefer a smooth sounding audio grade cap such as Elna Silmic but on the +5 and -5v supplies I'm not so sure. Currently I'm using the following set up:

-15v Elna Silmic II 100uF 25v
-5v Elna Silmic II 220uF 16v
+5v Sanyo Oscon SEPC 470uF 16v bypassed with 100nF Panasonic PPS SMD

With Silmic on all three supplies I found the sound to be very smooth (perhaps overly so?). With the OSCON SEPC on both the +5v and -5v supplies the sound was super detailed and very crisp but I found it too fatiguing. The combination of Oscon on the +5v and Silmic on the -5v and -15v seems to be a reasonable compromise between smoothness and crispness.

I'm very interested to hear other peoples experiences and which combination of capacitors are their personal favourites.
 
I used Oscon SEPC for years and was happy with the sound. I believe around 220uF. In those days it had a smt organic polymer on the pin. Adding a BG N .47uf right on the pin to my ear was pleasant smoothing the edge without removing detail. More recently I built one with Oscon bypassed with 1uF smt film cap, (same as recommended by John for decoupling). I got in my head that it sounded very detailed but almost plastic. Replaced the oscons with BG NX HiQ 33uf on the +-5v and a BG N 33uf on the -15. To me this is perfect. Detailed yet natural. The supplies up to these are very well filtered with soft recovery diodes, and clclclc filters, shunt regs. Following some of Johns advice the choke filtering added as much as the final caps. I think he advocates a 9th degree filter. While they are hard to find and overpriced, if you happen to get BG NX HiQ caps, they pair very well with this chip. Those that don't believe in sound of caps can likely see effect on noise on scope.
 
Dépends of how is made your PS for each of the rails voltage,

I'm now particulary like at the pins of those rails : just a little decoupling only on the -16 V if you have a fast shunt PS and Nothing after the follower (Darlington made with discrete bipolars for the +&-5 V & -16V) !

Yeap... Only a SMT PPS cap or PPS NPO cap (has to be benchmarked with ears in your setup in relation to the tonal balance of your system)
near the -16 V pin of the TDA. Value between 0.020 and 0.100 uF (benchmarked with ears in your own system with your ears for the best trade off you much like). I would like to benchmark a Black-Gate N cap at 0.1 uF as well ! But leave the -/+ 5V alone!

in that setup main power cap before the shunt PS are important, I prefer Panasonic FS over FM or Nichicons' one for their snappy bass and clearness in medium-highs in my system. Tried in several TDA1541 stuffs and it works for me at home !
 
Last edited:
Sorry Mike, an errata :

I talked about Panasonic FC... not FS, my mistake, advise you 1200 to 1500 uF at main cap in a shunt PS (before the current follower). (you can // it with let say 300 to 600 uF to setup it).

I do advise if your layout PCB is good enough 0.020 to 0.1 uF with smt PPS or COG (try both) only on the -16 V pin as close as you can. Avoid 1 uF. I benchmarked a lot of values on all the 3 voltage rails and in relation to each others 3 rails on several pcb (shame on me, caps despite being an old debat works for me and makes great différences) for the last decoupling cap value (the little bypass).

The setup I testimonie above gives the best result at home : faster transcient and snappier mid-bass but with higher resolution détails... without loss of bass.
 
Disabled Account
Joined 2002
Hello,
In my Curcio dac the 3 caps surrounding the dac chip are all 10mF 50 volts standard grade. There some is a kind of voltage regulation which seems pretty standard to me but no place to put something else without messing up the board a lot.
I did however modify the power supply part that is in front of the regulation part. Bigger transformer, schottky diodes, elna silmic caps and choke input.
I have some of these red B Gate caps. 2*47mF 50 volts, 8* 100mF 16 volts and 2*100mf 6,3 volts.
Maybe the 6.3 vol;ts are to close for the 5 volts supply and the 16 volts are to close for the 15 volts supply. Should i go for two 16 volts for the 5 volts supplies and the 50 volts for the 15 volt supply?
Greetings, eduard
 
Thanks for the replies.

I used Oscon SEPC for years and was happy with the sound. I believe around 220uF. In those days it had a smt organic polymer on the pin. Adding a BG N .47uf right on the pin to my ear was pleasant smoothing the edge without removing detail. More recently I built one with Oscon bypassed with 1uF smt film cap, (same as recommended by John for decoupling). I got in my head that it sounded very detailed but almost plastic. Replaced the oscons with BG NX HiQ 33uf on the +-5v and a BG N 33uf on the -15. To me this is perfect. Detailed yet natural.

wlowes, are you still using the SMT film caps bypass with the BG's or do you just have the single BG caps now? I don't have any BG's but I've got plenty of Cerafines which are supposed to be quite close to the BG sound.

Like you I found the Oscons very crisp and detailed but somehow less pleasing to listen to/less musical.

I should have mentioned earlier that I have recently upgraded from 7805/7905 regs on the +5v and -5v respectively to the Spower super regs from Fidelity audio and they were a very pleasing upgrade adding some lovely smoothness and refinement to the sound. I also have a 15uF Oscon SP from the -5v pin to the -15v pin.

On my previous TDA1541A project (an Arcam Alpha 5) I finally settled on 220uF Silmic for the -5v and -15v with a Pansonic FC 470uF on the +5v. To me the FC's were too bright if used everywhere and the Silmic's too mellow if used on the +5v pin but the combination of Silmic and FC gave a good overall balance.