Building the ultimate NOS DAC using TDA1541A

[Bernhard]

Originally Posted by erin

Bernhard, I read many of your threads and posts where you complain about the performance of various DAC chips. Please advise which DAC chip(s) you do like? It will be enlightening for all concerned.
Thanks

With my favourite PCM56, passive I/V is possible and 50% of chips have very good low level thd after adjustment.

Other DAC chips do not allow passive I/V, or 9 out of 10 have bad low level thd.
Maybe PCM61 or some AD chips like 1865 are an alternative.

I have no experience with modern chips because I stay with non oversampling.

[triode_al]

Bernard, Though very OT by now, why did you not mention PCM63 or is that in your category of bad low level thd?

With PCM63P, the low-noise steps are clearly defined and
increase in near-perfect proportion. This performance is
achieved without any external adjustments. By contrast,
sigma-delta (“Bitstream”, “MASH”, or 1-bit DAC) architectures
are too noisy to even see the first 3 or 4 bits change (at
16 bits), other than by a change in the noise level.

(PCM63 brochure) Funny what they say of the streaming bits at that time a hype.

[Ken Newton]

Originally Posted by Bernhard

With my favourite PCM56, passive I/V is possible and 50% of chips have very good low level thd after adjustment.

Other DAC chips do not allow passive I/V, or 9 out of 10 have bad low level thd.
Maybe PCM61 or some AD chips like 1865 are an alternative.

I have no experience with modern chips because I stay with non oversampling.

Hi Berhard,

This reminds me to inquire whether you have measured THD of the PCM1794A with passive I/V?

As you said, only a very few DAC chips are suitable for use with passive I/V. Most of these are 16-bit R-2R converters, but the 24-bit sigma-delta hybrid PCM179X series seems to be a contemporary exception. I believe that a number of commercial digital playback components use some version of PCM179X with passive I/V, typically of about 200 ohms on each output phase producing a FS signal of up to 1VRMS differentially.

The commercial digital playback components which I believe to utilize the PCM179X with passive I/V are: Audio Research CD-8, Ayre QB-9, Neko D100, and the Raleigh Audio RAKK DAC.

[Bernhard]

Originally Posted by triode_al

Bernard, Though very OT by now, why did you not mention PCM63 or is that in your category of bad low level thd?

(PCM63 brochure) Funny what they say of the streaming bits at that time a hype.

I think it has a clipping diode at the output and would need a high ratio transformer for passive I/V.
Also test results not overly impressing. Finding very good ones becomes expensive.

[Bernhard]

Originally Posted by Ken Newton

Hi Berhard,

This reminds me to inquire whether you have measured THD of the PCM1794A with passive I/V?

As you said, only a very few DAC chips are suitable for use with passive I/V. Most of these are 16-bit R-2R converters, but the 24-bit sigma-delta hybrid PCM179X series seems to be a contemporary exception. I believe that a number of commercial digital playback components use some version of PCM179X with passive I/V, typically of about 200 ohms on each output phase producing a FS signal of up to 1VRMS differentially.

The commercial digital playback components which I believe to utilize the PCM179X with passive I/V are: Audio Research CD-8, Ayre QB-9, Neko D100, and the Raleigh Audio RAKK DAC.

PCM1794A has oversampling inside.

In recent years I was very lazy but now I have a sampling scope with a horizontal resolution of 1ps/div and around 1ps intrinsic jitter.
Could help me to continue working on the digital section of my DAC.

I still have to get familiar with it.

This is 100kHz output from a "function generator":

[Ken Newton]

Originally Posted by Bernhard

PCM1794A has oversampling inside.

Yes, I know that the PCM1794A has an integrated OS digital filter - and, therefore, is not something one would consider for a NOS DAC project. Just the same, I was curious about how the PCM1794A THD performance measures via a resistor I/V because a few high-end digital vendors appear to be using it that way. Passive resistor I/V, while not uncommon in use with older generation 16-bit R-2R DAC chips is unique, to my knowledge, in use with the PCM1794A among modern 24-bit sigma-delta based DAC chips.

[triode_al]

Originally Posted by Bernhard

I think it has a clipping diode at the output and would need a high ratio transformer for passive I/V.
Also test results not overly impressing. Finding very good ones becomes expensive.

yes I have heard too the PCM63 has BAT-diodes to protect the output (the 1541 unfortunately has not, it costed me one at least in testing an I/V stage ).

But an I/V with really low input (lower than John's 25-30 ohms of the 1/Idss of his 2SK170), say 2-4 ohms = less than 5-8 mV RMS will not be influenced by the remaining impedance of the diodes - this is my conjecture.

And I agree with your observation on the linearity, I saw that the PCM56's I had showed the LSB clearly; but the TDA1541 via the CS8412 will also shows the LSB clearly.
albertr

[Audio curiousity]

Hi John,

Interesting reading. I learned a lot here, thanks for that. It took me a while to read it though.

A few questions:
- Are the latest DAC schematics availeble? I couldn't find or have missed them.
- Is the DAC free to build? I prefer hardwire instead of PCB.
- I have a plain TDA1541 (non A). Can the DEM capacitor be neutralized somehow and is it possible to apply an external DEM clock at pins 16 and 17 instead?
- Could you tell me something about sound of a cryo treated TDA1541A? I know it works fine for tubes, so i'm interested.
- I live in the NL. Could i have a listen at your place?

Some ideas that i hope might be helpful:
- A balanced powersupply may result in about 20 to 30 dB additional interference noise cancelation. More information: Balanced Power (e.g. articles/widescreen.html).
I've applied this technique in my home build tube gear and cd player with great results. No more audible humm and noise, better definition, blacker background, more lifelike e.g. metal sounds etc. The TDA1541A specs dont show a balanced design, so maybe a balanced noise cancellation might be realised by using bleeder resistors and compensate the 'bleeding' power by choosing an equivalant higher VA rating etc.
- By removing (the upper layer of) the TDA1541A chip plastic and applying thermal conducting paste and a heatsink thermal distortion can probably be reduced. In my experience THD reduces also due to absence of plastic capacitance and better resonance suppression by the thermal paste. Needless to say the thermal paste needs to be electrical non conductice. I read something about removing the plastic housing in this thread. Did someone applied a stripped TDA1541A with some results?
- Beeswax and a wood center in coils can reduce small resonance en therefore distortion. Had some good results there in the past.
- Stripping capacitors form their plastic / aluminium housing and wrap them in paper, paint them with a natural varnish like 'schellak' or C39 and fixate them in a wooden (birch wood or enoby / blackwood sounds best imo) housing with beeswax can result in sound improvements. See also || DHTRob - Het modificeren van een electrlytische condensator (1) || (Dutch only, sorry)
- Jensen four pole capacitors after the bridige rectifiers sound qiute good. Better then blackgate WKz and way better than Sanyo Oscon imo. Expensive, around €70 each, but as for ultimate... See https://audio.jensencapacitors.com/products/

Henk

[Audio curiousity]

Hi John,


Thanks for giving me the oppertunity to drop by and listen to your DAC.

It sounds analogue and very lifelike. It's almost like listening to a live performance. I did'nt expect the absence of digital artifacts and had to get used to that, before i realised these artifacts were missing. I think this DAC is comparable with the best availeble in the market, and fyi i'm quite familiar with Audio Not DAC 2 and 3 with maximum modification and the Audiomagic DAC.

I learned a lot and hope to build my own DAC soon.


Henk

[jstang]

Three years after building my 4x 1541 s1 dac with input from John's thread....I am still finding some improvements and still just melt into my chair listening to it....

Recently I added a lot of SHIELDING between sections in my DAC. For example shielded the output from the rest...the power supplies...the linear intorpolation circuits...and on and on .....

Noise floor and jitter dropped...

What a great chip when you get it right....well almost right...I think some more shielding can help. It never really ends, does it....

JohnK