Gentlemen, instead of arguing pro and contra about the superiority and inferiority of this and that DAC, would somebody be so kind and measure their low-level linearity using the following objective method?
http://www.diyaudio.com/forums/digital-source/113620-dac-linearity-test-cd.html
Sorry but I can not do it myself, because I have no examples of TDA1543, SM5865, PCM1702 etc. Could be interesting the result, though, especially in comparison. Effect of PSU noise, I2S crosstalk and others could be visually evaluated, no need for expensive FFT, just an oscilloscope and a good low noise measuring amplifier.
http://www.diyaudio.com/forums/digital-source/113620-dac-linearity-test-cd.html
Sorry but I can not do it myself, because I have no examples of TDA1543, SM5865, PCM1702 etc. Could be interesting the result, though, especially in comparison. Effect of PSU noise, I2S crosstalk and others could be visually evaluated, no need for expensive FFT, just an oscilloscope and a good low noise measuring amplifier.
Oshifis, it has been done on post #9
It was the 1541 non A but I would not expect any wonders...
I copy the whole post
oshifis,
thanks, very useful files.
Track 15 seems to be good to (pre-)adjust MSB very fast, did not check how accurate.
Another useful files would be single bits turned on continiously to adjust bits of discrete DAC with a DVM.
Or single bits turned on and off with same frequency like posted files
to adjust bits of discrete DAC with analyzer automatic Vpp measuring function.
Analyzer with 500mV full scale and 40x preamp:
PCM56 8x os Yamaha CDX-5000 good adjusted MSB track 15
PCM56 8x os very bad adjusted MSB track 15
I measured also PCM63, less noise but no better linearity.
TDA1541 4x os Philips CD960 MSB track 15
TDA1541 bad bits ( various tracks )
It also has a few good ones
__________________
R2R nonos / analog filter = Diamond
Multibit delta sigma / DF = Zirkonia
Standard delta sigma / DF = Glass
It was the 1541 non A but I would not expect any wonders...
I copy the whole post
oshifis,
thanks, very useful files.
Track 15 seems to be good to (pre-)adjust MSB very fast, did not check how accurate.
Another useful files would be single bits turned on continiously to adjust bits of discrete DAC with a DVM.
Or single bits turned on and off with same frequency like posted files
to adjust bits of discrete DAC with analyzer automatic Vpp measuring function.
Analyzer with 500mV full scale and 40x preamp:
PCM56 8x os Yamaha CDX-5000 good adjusted MSB track 15
PCM56 8x os very bad adjusted MSB track 15
I measured also PCM63, less noise but no better linearity.
TDA1541 4x os Philips CD960 MSB track 15
TDA1541 bad bits ( various tracks )
It also has a few good ones
__________________
R2R nonos / analog filter = Diamond
Multibit delta sigma / DF = Zirkonia
Standard delta sigma / DF = Glass
Hi John,
thank you for feedback about ad1862.
In Your old DEM clock injector You feed inverted BCK to pin 16 and non inverted to pin 17. In the current version You don't feed any clock signal to pin 17. Would You post a schematic? Don't you use higher value decoupling caps any more?
Do you have experience with lower clock rate for DEM circuit. BCK is a way over the normal frequency of the circuit.
I don't understand the principle idea behind the DJA circuit. It is a sort of lowpass/integrator circuit. How is jitter on bck minimized? Doesn't this filter mess up with raising edge of bck pulse?
thanks
Hi omainik,
200mVpp (1.4112 MHz) bit clock > 13K0 resistor > BAT43 > pin16 (cathode of BAT43 connects to pin 16)
No I don't use higher value decoupling caps at the moment, in fact I am experimenting with no decoupling caps at all.
This prevents feeding interference (GND) into the bit currents, similar as with disconnecting Vref on the TDA1543.
This is only possible when DEM clock = bit clock.
Advantages are:
- No injection of interference (GND) into bit currents.
- Eliminating effects of decoupling cap properties (sound coloring).
- Fewer components, simplified circuit.
First function of the DJA is reducing bit clock amplitude and providing low output impedance. BCK dynamic duty cycle fluctuations (jitter) are translated to minute DC fluctuations, using an RC integrator. The integrator signal dynamically shifts BCK DC level in such way that the TDA154x can trigger slightly sooner or later. This counter-acts the effects of jitter, effectively reducing it. Correction range of this DJA circuit is very limited, so BCK input signal still needs to have low jitter.
The edges of the DJA output signal are not distorted or "messed up", this can be easily verified with an oscilloscope.
200mVpp (1.4112 MHz) bit clock > 13K0 resistor > BAT43 > pin16 (cathode of BAT43 connects to pin 16)
No I don't use higher value decoupling caps at the moment, in fact I am experimenting with no decoupling caps at all.
This prevents feeding interference (GND) into the bit currents, similar as with disconnecting Vref on the TDA1543.
This is only possible when DEM clock = bit clock.
Advantages are:
- No injection of interference (GND) into bit currents.
- Eliminating effects of decoupling cap properties (sound coloring).
- Fewer components, simplified circuit.
First function of the DJA is reducing bit clock amplitude and providing low output impedance. BCK dynamic duty cycle fluctuations (jitter) are translated to minute DC fluctuations, using an RC integrator. The integrator signal dynamically shifts BCK DC level in such way that the TDA154x can trigger slightly sooner or later. This counter-acts the effects of jitter, effectively reducing it. Correction range of this DJA circuit is very limited, so BCK input signal still needs to have low jitter.
The edges of the DJA output signal are not distorted or "messed up", this can be easily verified with an oscilloscope.
EC,
wouldn't that be a great idea ?
You could show us the linearity of your "life sound" DAC and how it resolves the toggling of the LSB.
Still I prefer FFT but if that is not possible...
http://tube.fw.hu/lintest.zip
By the way, I find it very interesting how a randomly chosen CD player can do it without the ultra low noise power supplies, without the ultra low jitter 180 MHz c-mos reclocking, without the I2S attenuators, using a standard opamp for I/V conversion and finally a DAC chip that does not have the ECL input circuit.
And hey - neither the player, nor the scope did run on a battery power supply
Maybe
this player just has a super linear multibit R2R DAC chip ?
Last edited:
I think it is referenced to the -60dB signal as I do it too, would mean -60 + -47 for S1 or S2 = -107 dB as per datasheet.
He Ecdesigns,
How does it sound? I am very intrigued.
This is clear
I don't understand this function. How should the integrator do that?
How does it sound? I am very intrigued.
This is clear
I don't understand this function. How should the integrator do that?
Dear Berhard wrote:
Then there is no point is researching for a better DAC if a ramdomly chosen CDP can give us "perfet sound forever"...
What a waste of time!
Look, if you think TDA's are not worth it then let it go, man!
You have expressed your opinion several times and we have recorded it; there is no point in posting again and again the same graphs...you are just polluting a thread made for those interested in implementing working DACs with those chips. That's it.
-EC- is a polite gentleman and won't tell you that you stop bothering. For a guy with your skills it should take a couple of hours to build a DAC with all -ECdesign's- recommendations and listen for yourself (and mesure if you want) if his claims of good sound are unreal or not!
Take care,
M.
Then there is no point is researching for a better DAC if a ramdomly chosen CDP can give us "perfet sound forever"...
What a waste of time!
Look, if you think TDA's are not worth it then let it go, man!
You have expressed your opinion several times and we have recorded it; there is no point in posting again and again the same graphs...you are just polluting a thread made for those interested in implementing working DACs with those chips. That's it.
-EC- is a polite gentleman and won't tell you that you stop bothering. For a guy with your skills it should take a couple of hours to build a DAC with all -ECdesign's- recommendations and listen for yourself (and mesure if you want) if his claims of good sound are unreal or not!
Take care,
M.
I was out of the DIY Hi-Fi game for 25 plus years...Worked hard over the 3 years to catch up.
The last year has been the year of learning DACs & CD-Transport.
Built a CD-Pro transport and a 4x TDA1541 DAC. This thread has been very enlightening and critical for me to start doing DACs right.
jk
The last year has been the year of learning DACs & CD-Transport.
Built a CD-Pro transport and a 4x TDA1541 DAC. This thread has been very enlightening and critical for me to start doing DACs right.
jk
-EC- is a polite gentleman .....-ECdesign's- recommendations and listen for yourself (and mesure if you want) if his claims of good sound are unreal or not!
Then there is no point is researching for a better DAC if a ramdomly chosen CDP can give us "perfet sound forever"...
What a waste of time!
Not exactly. This was the reply to EC as he has doubted the accuracy and validity of my measurements with very fragile arguments.
In other words, You need nothing more than standard implementation to tell wether a chip has good linearity or not.
Are you a moderator or something like that ?
The graphs were posted on request of oshifis.
But somehow you're right: Waste of time from my point of view.
It would be very easy for him to verify how good his DAC is.
All that is needed is a soundcard, some freeware, a few free test signals and a preamp.
There have been many explanations, not only from me, try to understand and not just record.
For a guy with your skills it should take a couple of hours to build a DAC with all -ECdesign's- recommendations and listen for yourself (and mesure if you want) if his claims of good sound are unreal or not!
Oh, thanks...
Any contribution to the technical discussion ?
Last edited:
No, I can't contribute to the technical discussion because I have little technical knowledge...nevertheless, I suspect that there are other parameters more important in quality audio reproduction ("live" as John says) that are more important than THD and linearity...
I have contributed though, confirming practically (that is, doing the thing) that every single recommendation that -EC- has made here is worthy from the sound point of view, in a way that one cannot go back.
I could comment further but it is pointless...
I have contributed though, confirming practically (that is, doing the thing) that every single recommendation that -EC- has made here is worthy from the sound point of view, in a way that one cannot go back.
I could comment further but it is pointless...
Another attempt.
A guy has built a preamp.
The guy tells you that the preamp produces live sound.
The guy shows you the schematic.
What do you see ?
The guy uses LM741 opamp as output buffer because the guy believes this is a good opamp. ***
You know the opamp...
You tell the guy that this is a very bad opamp because it has a classB output stage and it is very slow and has very high distortion.
What does the guy do ?
He says, no - this is a very good opamp, no other opamp sounds as good like that, also I have a very good power supply with batteries, that is much more important.
You tell the guy that the opamp will remain a very bad opamp even with a very good power supply.
The guy says no - also I use very good teflon coupling caps, forget about the distortion of the opamp... The Teflon caps are important.
And so on...
You know, if the guy would use a better opamp...
Understand ?
*** Obviously the guy has no equipment to measure the opamp and also he does not believe the datasheet that explicitely mentions the high distortion of the opamp, and also mentions that the opamp is made for kitchen radios, but perhaps the guy believes that the manufacturer of the opamp is stupid and had a bad implementation when testing the opamp.
Anyway the guy never believes other people...
If somebody else tells him that he measured the opamp and the opamp is not good, he will doubt the accuracy of the equipment and tell him that the preamp produces live sound.
A guy has built a preamp.
The guy tells you that the preamp produces live sound.
The guy shows you the schematic.
What do you see ?
The guy uses LM741 opamp as output buffer because the guy believes this is a good opamp. ***
You know the opamp...
You tell the guy that this is a very bad opamp because it has a classB output stage and it is very slow and has very high distortion.
What does the guy do ?
He says, no - this is a very good opamp, no other opamp sounds as good like that, also I have a very good power supply with batteries, that is much more important.
You tell the guy that the opamp will remain a very bad opamp even with a very good power supply.
The guy says no - also I use very good teflon coupling caps, forget about the distortion of the opamp... The Teflon caps are important.
And so on...
You know, if the guy would use a better opamp...
Understand ?
*** Obviously the guy has no equipment to measure the opamp and also he does not believe the datasheet that explicitely mentions the high distortion of the opamp, and also mentions that the opamp is made for kitchen radios, but perhaps the guy believes that the manufacturer of the opamp is stupid and had a bad implementation when testing the opamp.
Anyway the guy never believes other people...
If somebody else tells him that he measured the opamp and the opamp is not good, he will doubt the accuracy of the equipment and tell him that the preamp produces live sound.
Yup .. I upgraded the clock on the CD-Pro and reclocked the I2S. I reclocked all 3 I2S signals.
The standard on board CD-Pro clock is not very good, you will hear a very noticeable improvement just by reclocking the CD-Pro. Then I2S reclock will be heard as well.
I wish my Hifidiy transport kit had a programming function for the control bus. I would like to command the CD-Pro to turn off some functions and modify some others.... oh well, can't have everything.
I set up the linear delay shift register circuits for the 1541s to time on 12 transitions of BCK for the 48 Bits out the CD-Pro. I also have music server that feeds the DAC via a Dir9001 spdif. That is a 64 bit word. That seems to sound just fine as well.
johnk
The standard on board CD-Pro clock is not very good, you will hear a very noticeable improvement just by reclocking the CD-Pro. Then I2S reclock will be heard as well.
I wish my Hifidiy transport kit had a programming function for the control bus. I would like to command the CD-Pro to turn off some functions and modify some others.... oh well, can't have everything.
I set up the linear delay shift register circuits for the 1541s to time on 12 transitions of BCK for the 48 Bits out the CD-Pro. I also have music server that feeds the DAC via a Dir9001 spdif. That is a 64 bit word. That seems to sound just fine as well.
johnk
Interesting JK,
I have a CD-PRO 2M with a D1M (TDA1543; the AKM internal DAC s****s) but I will have to upgrade the clock and use it also to re-clock the BCK...as time allows...
Have you upgraded the system clock?
Bernhard, certainly criticism can be a helpful ingredient to this discussion but so long as it is constructive and not derisive...there are moments when your criticism is appreciated because pushing one to rethink a critical assumption central to one's thesis is generally a good thing, but then at times you become pretty condescending. Measurements have been requested by you as well as others, and John has decided not to produce them for reasons he has previously stated. Barring measurements, it would be interesting to hear impressions from several more enthusiasts who are in a position to compare John's player against other well executed players based on current dac chipsets.