Building the ultimate NOS DAC using TDA1541A - Page 419 - diyAudio
Go Back   Home > Forums > Source & Line > Digital Line Level

Digital Line Level DACs, Digital Crossovers, Equalizers, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 11th February 2012, 06:54 AM   #4181
diyAudio Member
 
-ecdesigns-'s Avatar
 
Join Date: May 2006
Hi Audio curiosity,

Quote:
An idea might be to replace the LED by a low noise NPN transistor used as a Zener with collecter lead clipped off, base connected to negative and emitter connected to positive voltage
Measuring results for some transistors used as zener diode (base and emitter).

2SC2240 9V5, 2SA970 10V8, BC547 9V, BC557 11V, 2N2905 7V3 2SA1360 10V, 2SC3423 7V6, MPSA44 9V4.

These zener voltages are too high to be used as reference voltage for 3.3, 3.6, and 5V shunt regulators.

However, this option is still interesting for the -15V supply. I also plan to test the LM329.


Quote:
More interesting info about LED noise (maybe mentioned before?):
I have seen those measurements and that's one of the reasons I have been using LEDs as reference.

When you look at the schematics you will see that I already applied some LED noise filtering using 220uF electrolytic capacitor and 1uF film capacitor.
  Reply With Quote
Old 11th February 2012, 07:44 AM   #4182
diyAudio Member
 
-ecdesigns-'s Avatar
 
Join Date: May 2006
Hi Tazz,

Quote:
assume the 10pf is the actuall input capacitance of the dac you can see how much is injected into it.
TDA1541A input capacitance (digital inputs) is specified at 12pF (Philips datasheet page 6).

Quote:
I think there might be room for some improvement with regard to the back to back diode solution.
The diodes basically limit signal amplitude to (Vin - 0.6) and +0.6V. So when using 5V TTL input signal, the output signal will swing between +0.6 and +4.4 (5V - 0.6V). Note that SD-player outputs 3.3V output signals. So the signal now swings between +0.6 and + 2.7V (3.3 - 0.6).

The TDA1541A works reliable with varying ambient temperatures when I2S signals swing between approx. +0.4V and +1.8V. So the existing circuit could cause problems at low temperatures (when TDA1541A is powered up cold). In this case the diodes in the TDA1541A input circuit have highest voltage drop and the input signal may fail to offer reliable latching.

Resistive attenuators in combination with TDA1541A 12pF input capacitance could lead to slow transients that could prevent reliable latching too.


The new SD-transport generates I2S signals (WS and DATA) with 1.8Vpp amplitude. I further reduce peak switching currents by using a series resistor (330R). The logic "0" threshold is lifted by a bias resistor between TDA1541A input and +1V8. This way the signal swings between approx. 0.4V and 1.8V, transients are still fast enough for reliable latching.

When the source generates 5V TTL level signals, it's best to use a 5V to 1.8V level translator for WS and DATA.

BCK is now generated directly by the timing module. The synchronous divider runs on 1.8V power supply, so it outputs approx. 1.8Vpp.
  Reply With Quote
Old 11th February 2012, 08:55 AM   #4183
diyAudio Member
 
-ecdesigns-'s Avatar
 
Join Date: May 2006
Hi marconi118,

Quote:
They are 8pin and simple as the TDA1543 but with the 0v output compliance of the TDA1541.
It could be better sounding than TDA1543 but less complicated than TDA1541 DAC.
TDA1543 and TDA1541A are based on CML (current Mode Logic). This offers much lower switching noise levels compared to TDA1387 and TDA1545 that are based on CMOS logic. CML maintains a constant current during switching (constant current sources), this offers very low switching noise levels. CMOS draws almost no current at steady state and peak currents during switching. This leads to high switching noise levels as current changes are now maximum.

Lower switching noise results in lower timing uncertainty at the D/A stage. This in turn leads to lower jitter levels at the D/A stage. Similar, the output signal remains cleaner as less noise is added to it.

This is one of the reasons why I still use the TDA1541A / TDA1543 and not CMOS based DACs that offer similar or better (THD) specs. Speaker distortion is always added to the DAC THD, so minimum THD will be around 1% when using high performance speakers and can increase above 10% when using average speakers. Similar, some of the LSBs are fully swamped by audio set distortion / noise. So even when using 64 bit digital audio source, practical resolution would remain well below 16 bits.

Given an optimized circuit where noise levels have been reduced to minimum practical levels. The limiting factor will then be jitter at the D/A stage and that is basically set by DAC chip properties.
  Reply With Quote
Old 11th February 2012, 01:11 PM   #4184
diyAudio Member
 
Bernhard's Avatar
 
Join Date: Apr 2002
Location: Munich
Quote:
Originally Posted by -ecdesigns- View Post

1) Similar, the output signal remains cleaner as less noise is added to it.

2) This is one of the reasons why I still use the TDA1541A / TDA1543 and not CMOS based DACs that offer similar or better (THD) specs. Speaker distortion is always added to the DAC THD, so minimum THD will be around 1% when using high performance speakers and can increase above 10% when using average speakers. Similar, some of the LSBs are fully swamped by audio set distortion / noise. So even when using 64 bit digital audio source, practical resolution would remain well below 16 bits.

3) Given an optimized circuit where noise levels have been reduced to minimum practical levels. The limiting factor will then be jitter at the D/A stage and that is basically set by DAC chip properties.
1) A simple test: 1 khz -60dB or - 80dB dithered sine wave displayed on a spectrum analyzer without averaging will show the noise floor. Comparison with other DACs will reveal the truth.

2) DAC related harmonics are always audible through speakers because of different spectrum.
In independent tests, TDA1541A totally failed in reproducing a LSB toggling signal, spitting out garbage, while other chips reproduced it flawless.
Resolution well below 16 bit is true for TDA but not for all.

3) Noise is less critical for listening than harmonics. Noise can mask detail but harmonics will alter tonal balance and can be disgusting.
  Reply With Quote
Old 12th February 2012, 02:32 PM   #4185
diyAudio Member
 
-ecdesigns-'s Avatar
 
Join Date: May 2006
Hi Bernhard,

Quote:
1) A simple test: 1 khz -60dB or - 80dB dithered sine wave displayed on a spectrum analyzer without averaging will show the noise floor. Comparison with other DACs will reveal the truth.
It only shows performance of the DAC when generating 1 KHz sine wave. In practice the DAC has to generate many fundamentals plus harmonics simultaneously. The 1 KHz measurement doesn't show what will happen in this case.

Quote:
2) DAC related harmonics are always audible through speakers because of different spectrum.
In order to become audible, the harmonics must exceed certain threshold level so they are not masked by other (louder) signals.

Quote:
In independent tests, TDA1541A totally failed in reproducing a LSB toggling signal, spitting out garbage, while other chips reproduced it flawless.
DAC linearity test CD

Post # 39 (oshifis):

"I tried this also on a Marantz CD-74 modified to NOS and digital filter bypassed. Here most bits are perfect or nearly perfect, except bit 9 which shows some oscillation-like noise on the bottom level, but not serious. Another reason for NOS?"

It seems like the garbage was caused by the SAA7220, not by the TDA1541A.

Quote:
Resolution well below 16 bit is true for TDA but not for all.
Philips datasheet spec indicates typical Edl of 0.5 LSB for a plain A chip. I have no reason to believe that Philips was not able to perform correct measurements.

When TDA1541A measured resolution turns out to be well below 16 bits, it is likely there is something wrong with the test setup / application.

Quote:
3) Noise is less critical for listening than harmonics. Noise can mask detail but harmonics will alter tonal balance and can be disgusting.
Depends on how this noise interacts with circuit performance, think of masterclock or reclocker / divider power supply noise for example. Noise on the DAC chip logic power supply can also interact with on-chip logic circuits that in turn can lead to higher jitter amplitude and / or different jitter spectrum at the D/A stage.
  Reply With Quote
Old 13th February 2012, 03:37 AM   #4186
jambul is offline jambul  Belgium
diyAudio Member
 
Join Date: Mar 2011
Dear John,

out from building the ultimate dac with TDA1541A,

Can i apply your clock design with CS8416+pcm1794?

thank you

denis
  Reply With Quote
Old 13th February 2012, 10:33 AM   #4187
diyAudio Member
 
Join Date: Mar 2008
Dear John,
You convinced me to build your outstanding TDA1543 DAC with some modifications:

Battery PSU:
use alkaline batteries that have less noise than nicd or lead-acid
4X1.5v D cell in series direct supply to the +5V (it will be +6V but 1543 should survice)

Tap in the midle of the 4x1.5V series to obtain the 3V that goes to the I/V resistors. Or has it to be 3.2V exactly? maybe by changing the value of the I/V resistors (820R) it is possible to have 3.0V as the best bias reference?

Any advice on this?
  Reply With Quote
Old 14th February 2012, 10:35 PM   #4188
diyAudio Member
 
Bernhard's Avatar
 
Join Date: Apr 2002
Location: Munich
Quote:
Originally Posted by -ecdesigns- View Post
It only shows performance of the DAC when generating 1 KHz sine wave. In practice the DAC has to generate many fundamentals plus harmonics simultaneously. The 1 KHz measurement doesn't show what will happen in this case.
You talk about the digital section of the DAC. And that similar, the output signal remains cleaner as less noise is added to it. Wether the measuring signal is 1 kHz or a "complex signal", doesn't matter much because there is always a number of bits switching from sample to next sample. The amount of parasitic energy depends on numbers of bits switching and MSB produces more trash than LSB.
Sample hold was used to suppress this energy with older chips that had bad settling time and produced spikes and other garbage. Same with the digital logic in the chip.
Even if it made a difference, when comparing two different chips, you can not expect that the results turn completely to the opposite when you change the measuring signal.
If I had to question one sort of audio measurement, I would name jitter measured with a analog scope.
One example of noise floor below.


Quote:
In order to become audible, the harmonics must exceed certain threshold level so they are not masked by other (louder) signals.
I find this very questionable, but if it was true, the same would even more apply to noise.



Quote:
DAC linearity test CD

Post # 39 (oshifis):

"I tried this also on a Marantz CD-74 modified to NOS and digital filter bypassed. Here most bits are perfect or nearly perfect, except bit 9 which shows some oscillation-like noise on the bottom level, but not serious. Another reason for NOS?"

It seems like the garbage was caused by the SAA7220, not by the TDA1541A.
The Marantz CD-74 uses TDA1540.
He did the other measurement with a CD-94.

Yes, it is obvious that the oszillation/noise trash goes away when removing the oversampling.
But that does not mean it is getting better.
When you scroll down last page of that thread, you can see my measurements of -60dB signal with and without oversampling.
The horizontal line marks 60dB below the signal.
You can clearly see there is more noise with oversampling but less distortion.
Removing oversampling results in less noise but raises the harmonics.
Less noise without oversampling is in line with oshifi's observation.

My statement stands. The linearity of the 1541A is everything but good.

Maybe Philips added noise on purpose to mask harmonics. Nobody knows.
Other manufacturers used many tricks in the digital section to bring down low level distortion.
Which shows, they understood the importance of it !
Teac added digital noise to prevent the dac going below a certain analog level.
Later the noise was removed by canceling with an alternating noise only signal.
Clever people.
The cancellation had to be nulled with a pot.
That gave the opportunity to play around and listen to a noisy DAC.
I clearly preferred noise over harmonics.

Quote:
Philips datasheet spec indicates typical Edl of 0.5 LSB for a plain A chip. I have no reason to believe that Philips was not able to perform correct measurements.
I do. IIRC the specs for the R1 version were better than for the plain A.
But it is well known that the R1 had relaxed specs.
Also, "typical" means, if your chips not hold up to their promise, they are sorry.

Quote:
When TDA1541A measured resolution turns out to be well below 16 bits, it is likely there is something wrong with the test setup / application.
Please don't bother me.
I measured enough chips to know what I'm doing.
1541A low level performance is fair at best.
And this is, to use your own words: Highly critical.
Because it has exactly the same effect like crossover distortion in class B amplifier.
It gives unnatural sound.
I found a few good TDAs, sold them on ebay.

You say it yourself that there are better performing DAC chips:

Quote:
This is one of the reasons why I still use the TDA1541A / TDA1543 and not CMOS based DACs that offer similar or better (THD) specs.
Quote:
Speaker distortion is always added to the DAC THD...
Similar, some of the LSBs are fully swamped by audio set distortion / noise.
I can not agree.


Quote:
So even when using 64 bit digital audio source, practical resolution would remain well below 16 bits.
If the chip performs like this from the start, yes.

Quote:
Quote:
3) Noise is less critical for listening than harmonics. Noise can mask detail but harmonics will alter tonal balance and can be disgusting.

Quote:
Depends on how this noise interacts with circuit performance, think of masterclock or reclocker / divider power supply noise for example. Noise on the DAC chip logic power supply can also interact with on-chip logic circuits that in turn can lead to higher jitter amplitude and / or different jitter spectrum at the D/A stage.
I was not referring to jitter, (sorry if now jitter is your killer argument) but to what you say:
...similar, the output signal remains cleaner as less noise is added to it.

And you use TDA because you believe it has less noise than cmos. Maybe inside the TDA family.
Also you prefer the "quieter" TDAs over cmos that have better THD specs.

My opinion is very different. I believe harmonics are much more annoying than noise because they alter the tonal balance and add stuff that was not there before.

If you are on a holiday, what would you prefer ?

A hotel room close to the disco
or
a bungalow on the beach with noise from the sea ?

Same decibels.

Okay, you take the hotel room and I take the beach bungalow.

Have fun.

Last edited by Bernhard; 14th February 2012 at 11:02 PM.
  Reply With Quote
Old 14th February 2012, 10:58 PM   #4189
diyAudio Member
 
Bernhard's Avatar
 
Join Date: Apr 2002
Location: Munich
Quote:
Originally Posted by -ecdesigns-;
Hi omainik,
AD1862 uses CMOS logic elements, and has no current steering / ECL like TDA154x. TDA154x I2S attenuator (200mVpp output) does not work here.
And here:

Quote:
Originally Posted by -ecdesigns-;
Hi galeb,
No, (P)ECL gives too low output voltage swing to drive the TDA1541A. I use UHS buffers running on 1.8V power supply with series resistors on the inputs:
FAIRCHILD...
Now I'm a bit confused.

A DAC chip that runs on ECL, can not be driven by ECL ?

Last edited by Bernhard; 14th February 2012 at 11:12 PM.
  Reply With Quote
Old 15th February 2012, 10:43 AM   #4190
erin is offline erin  Australia
diyAudio Member
 
Join Date: Nov 2006
Location: http://www.makeitpossible.com/
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

Last edited by erin; 15th February 2012 at 10:52 AM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



New To Site? Need Help?

All times are GMT. The time now is 02:18 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2