Yes, the rolloff is known, measurable and you can easily hear it. It is a result of the zero order hold, by which the dac holds the sample value until the next sample. The resulting spectrum follows sinx/x as you pointed out. Which means -3dB or so at 20KHz. In oversampling dacs this rolloff happens above the audible band.
You can compensate for this in the analog stage as Thorsten and others have shown. Depending on your speakers and taste, you might not even have to compensate for it.
Best
Alex
Last edited:
I have contacted the distributor.
On the matter of DEM reclocking, I have read John's thread and become confused with the many different DEM reclocking circuits proposed on the thread. Thorsten, could you please point out which one is a working circuit, and which one "sounds" best.
thanks
Erin
Hello Thorsten
Can you say more about that very specific and precise response of the analogue stage for the SAA7220 ?
Did you place a sync or async reclock after the SAA7220 in your LHH-1000 DAC restoration ?
Thank you
Bye
Gaetan
Hi,
If the rolloff is there, of course you will hear it (even 60 Year old codgers can), but there is no reason to have any HF Rolloff with a Non-OS DAC or CD-Player.
HiFi Critic reviewed and measured the AMR CD-777 (the baby brother of the CD-77), you can see the full review here:
Abbingdon Music Research - Information - Reviews and Awards
To quote:
"Frequency responses were flat so there’s no ‘analogue tailoring’ here: 20Hz came in at -0.077dB, 20kHz at -0.5dB for ‘Digital Master 2’. ‘DM1’ was still only -1dB at 20kHz. No significant response errors or distortion variations were seen for any of the sampling options, so the sonic differences were inherent in the processes."
The "sampling options" refers to the various digital filters that may, or may not
be employed at playback.
BTW, martin Collomns also notes:
"After careful comparison with references, a fine sound quality score of 47 points was awarded."
Where this places this essentially "non-oversampling" player can be seen here:
http://www.hificritic.com/colloms/cdplayers.aspx
Ciao T
If the rolloff is there, of course you will hear it (even 60 Year old codgers can), but there is no reason to have any HF Rolloff with a Non-OS DAC or CD-Player.
HiFi Critic reviewed and measured the AMR CD-777 (the baby brother of the CD-77), you can see the full review here:
Abbingdon Music Research - Information - Reviews and Awards
To quote:
"Frequency responses were flat so there’s no ‘analogue tailoring’ here: 20Hz came in at -0.077dB, 20kHz at -0.5dB for ‘Digital Master 2’. ‘DM1’ was still only -1dB at 20kHz. No significant response errors or distortion variations were seen for any of the sampling options, so the sonic differences were inherent in the processes."
The "sampling options" refers to the various digital filters that may, or may not
be employed at playback.
BTW, martin Collomns also notes:
"After careful comparison with references, a fine sound quality score of 47 points was awarded."
Where this places this essentially "non-oversampling" player can be seen here:
http://www.hificritic.com/colloms/cdplayers.aspx
Ciao T
Hi,
There are differences. We must understand that the single TDA1541 will exhibit good channel match, but differences from one chip to the next in actual current output at 0dBFS are appreciable.
Now if just use the chips in parallel this does not matter, however to be honest I cannot see any reason to simply parallel TDA1541, in each and every case I tried this the sound changed, but not into anything significantly better.
Equally, the SNR, DR and THD&N of the TDA1541A are sufficiently close to perfection for 16 Bit that there is little point to try to improve them by paralleling (unlike the TDA1543).
An alternative option is if to use multiple TDA1541 in "analogue oversampling" or whatever this is called, basically multiple DAC's are addressed time delayed. Once we appreciate that there is easily a 1dB level difference between the different DAC's (unless carefully selected) and the result will not be as expected.
Of course, we could (at least at CD Speeds) use the four DAC's not just timedelayed as analogue oversampling, but also operating in effect as a 4 element DEM system (each DAC forms one element and for each sample assigned to a given position in the DAC delay chain we actually use each of the 4 DACs for 1/4 of the sample, so all errors average out), however no-one has implemented such a system and I still fail to see the point TBH.
I understand that many think:
More DAC's = More Better
More Tubes = More Better
More Bits = More Better
and so on.
However I generally find that such simplifications do not make sense.
Sometime more is more, at others less is more and at some even less is even more.
Ciao T
There are differences. We must understand that the single TDA1541 will exhibit good channel match, but differences from one chip to the next in actual current output at 0dBFS are appreciable.
Now if just use the chips in parallel this does not matter, however to be honest I cannot see any reason to simply parallel TDA1541, in each and every case I tried this the sound changed, but not into anything significantly better.
Equally, the SNR, DR and THD&N of the TDA1541A are sufficiently close to perfection for 16 Bit that there is little point to try to improve them by paralleling (unlike the TDA1543).
An alternative option is if to use multiple TDA1541 in "analogue oversampling" or whatever this is called, basically multiple DAC's are addressed time delayed. Once we appreciate that there is easily a 1dB level difference between the different DAC's (unless carefully selected) and the result will not be as expected.
Of course, we could (at least at CD Speeds) use the four DAC's not just timedelayed as analogue oversampling, but also operating in effect as a 4 element DEM system (each DAC forms one element and for each sample assigned to a given position in the DAC delay chain we actually use each of the 4 DACs for 1/4 of the sample, so all errors average out), however no-one has implemented such a system and I still fail to see the point TBH.
I understand that many think:
More DAC's = More Better
More Tubes = More Better
More Bits = More Better
and so on.
However I generally find that such simplifications do not make sense.
Sometime more is more, at others less is more and at some even less is even more.
Ciao T
Hi,
The SAA7220 was designed to pre-distort the frequency and phase response so that when combined with the analogue filter specified in the Datasheet the resultant response would be flat to 20Khz. It's all in the datasheet really.
Didn't have to. Marantz (who build the LHH-1000 combo for Philips and sold it themselves as CD/DA-12 as well) had kindly included a secondary PLL (all discrete, very low noise too, a real beauty), reclocking after the SAA7220 AND optical isolation to the SINGLE TDA1541 (which was of course a double crown, note).
So other than replacing obsolete and expired 'lytics with Sanyo Os-Cons and a few BG NX-HiQ and going wild with SMD Cermaic bypass caps and adding more LC decoupling with extra ferrite beads there was not a lot to do...
On the DAC/Analogue side BTW I used ton's of Elna Silmics, OPA637 in "trick" circuit (see AD797 Datasheet) as I/V and BUF634 as Filter (IV lowpass and 2nd order Sallen Key filter combined for the correct response) and output Buffer (final buffer jumpered out).
The transformer coupled balanced output (with a funky feedback scheme) got more OPA627's.
That was mostly it, just a restoration, no real circuit changes.
Ciao T
The SAA7220 was designed to pre-distort the frequency and phase response so that when combined with the analogue filter specified in the Datasheet the resultant response would be flat to 20Khz. It's all in the datasheet really.
Didn't have to. Marantz (who build the LHH-1000 combo for Philips and sold it themselves as CD/DA-12 as well) had kindly included a secondary PLL (all discrete, very low noise too, a real beauty), reclocking after the SAA7220 AND optical isolation to the SINGLE TDA1541 (which was of course a double crown, note).
So other than replacing obsolete and expired 'lytics with Sanyo Os-Cons and a few BG NX-HiQ and going wild with SMD Cermaic bypass caps and adding more LC decoupling with extra ferrite beads there was not a lot to do...
On the DAC/Analogue side BTW I used ton's of Elna Silmics, OPA637 in "trick" circuit (see AD797 Datasheet) as I/V and BUF634 as Filter (IV lowpass and 2nd order Sallen Key filter combined for the correct response) and output Buffer (final buffer jumpered out).
The transformer coupled balanced output (with a funky feedback scheme) got more OPA627's.
That was mostly it, just a restoration, no real circuit changes.
Ciao T
Hello Thorsten
That secondary PLL are very interesting, do you have the LHH-1000 schematic ?
Thank you
Bye
Gaetan
Hi,
Somewhere stored with my other papers at a friends house (and now likely water damaged) I had the xeroxed schematic of the nearly identical DAC of Marantz's next lower model including annotations about differences and some hand drawn schematics, that is about it.
This secondary PLL was probably revolutionary for the time a quarter of a century ago, but it is nothing THAT special by 2011 standards. It was just very nice to already have it and at a level that was massively better than what was in the CS8412 everyone used to swear by (or the CS8420 everyone used to swear at) at the time...
It was basically all the usual stuff, loop filter, parts of the phase detector, oscillator and so on implemented in discrete components, very nice LC oscillator with Varicap diode tuning IIRC, but no patch on a good modern VCXO. And not easy to do DIY either...
If you want a decent 2ndary PLL in 2011 use a pair of VCXO, a PIC with build in DAC to generate the tuning voltage for the VCXO in discrete, stable steps and to use as phase detector and use a good really low noise OPA with a very low corner frequency low pass on the tuning voltage and good quality supply for the VCXO's...
This will be smaller, more compact and higher performance than what Marantz did in the mid 80's. Alternatively try the 74HCT9046 PLL Chip with suitable VCXO's.
Ciao T
Somewhere stored with my other papers at a friends house (and now likely water damaged) I had the xeroxed schematic of the nearly identical DAC of Marantz's next lower model including annotations about differences and some hand drawn schematics, that is about it.
This secondary PLL was probably revolutionary for the time a quarter of a century ago, but it is nothing THAT special by 2011 standards. It was just very nice to already have it and at a level that was massively better than what was in the CS8412 everyone used to swear by (or the CS8420 everyone used to swear at) at the time...
It was basically all the usual stuff, loop filter, parts of the phase detector, oscillator and so on implemented in discrete components, very nice LC oscillator with Varicap diode tuning IIRC, but no patch on a good modern VCXO. And not easy to do DIY either...
If you want a decent 2ndary PLL in 2011 use a pair of VCXO, a PIC with build in DAC to generate the tuning voltage for the VCXO in discrete, stable steps and to use as phase detector and use a good really low noise OPA with a very low corner frequency low pass on the tuning voltage and good quality supply for the VCXO's...
This will be smaller, more compact and higher performance than what Marantz did in the mid 80's. Alternatively try the 74HCT9046 PLL Chip with suitable VCXO's.
Ciao T
Hi,
First, look up Nelson Pass's D1 for a simple secondary PLL
Second, don't use the oscillator on the PLL Chip, use VCXO
Third, asynchronous reclocking is something that I class as "it does not work", as in "it does not remove jitter" (note, I said ought about sound)
Fourth, true synchronous reclocking from a low jitter clock is hard to beat, all AMR Digital Products use exactly that method and of that reason
Ciao T
First, look up Nelson Pass's D1 for a simple secondary PLL
Second, don't use the oscillator on the PLL Chip, use VCXO
Third, asynchronous reclocking is something that I class as "it does not work", as in "it does not remove jitter" (note, I said ought about sound)
Fourth, true synchronous reclocking from a low jitter clock is hard to beat, all AMR Digital Products use exactly that method and of that reason
Ciao T
If you choose to pick out random quotes without reference to the the whole discussion you will tend to find them stupid.
I read the whole thread
My comment was directed at Thorsten who was saying that the TDA1541 with the optimal implimentation as used in the AMR77 was fandiddly-tastic, to which I enquired if he could share his recipie and he declined, to which I said that without the fandidly-tastic optimisations that modern dacs would beat a non fandangled TDA1541. And by beat, I mean in a multitude of ways which inclues the cost of the circuit, and accesabiltiy to ordinary people - who are not able to buy an AMR77. Sound quality vs cost.
This is DIY Audio - the place were we ( some of us
Now if you came to pick on me because I said one thing against the TDA1541 you really have chosen the wrong person
Pick on you ? NO - how could I - you have infinitely more electronics knowledge than I do, you understand the figures and can make comparisons.
I think I get a bit fed up of listening to everyone else's opinions on sound. How on earth can anyone tell me this ? I am a Pianist - I may listen in a different way - when my system improves because of something I do I absolutely know it's an improvement. I don't then force my opinions on the rest of the members.
FWIW, I dont like the sound of TDA1541 when it is used with a SAA7220.
So, that is why I run my TDA1541s in NOS mode. To me this sounds better.
I tried NOS for a while but went back to a well sorted 7220...see....we're all different
... my experience comes from what I have heard. If you think my experience is not good enough. Well OK, there is nothing I can do about that - you will just have to ignore my posts.
Mine too - I had the misfortune of working in a high end retailer for several years. The crap in there is incredible - some of the nice stuff was equally incredible. It never leaves you.
Your experience is not in question - we are music lovers FIRST - not the other way round.
The argy bargy that goes on regarding sound is really quite funny. If you hear a good sounding sound system, its a good sounding system. It does not matter what makes up the system as long as it all works well together.
This is the best part of your reply - could not have worded it better myself.
In truth, I felt a bit queasy watching Thorsten getting questioned in such a way - he doesn't need me to defend him of course - he's got the answers, experience and knowledge which he very eloquently demonstrated every time.
It was however like watching someone tell Nelson Pass he's wrong, deaf and can't solder
Erin - it was not personal and do accept my apologies if it seemed that way
If you choose to pick out random quotes without reference to the the whole discussion you will tend to find them stupid.
I read the whole thread
My comment was directed at Thorsten who was saying that the TDA1541 with the optimal implimentation as used in the AMR77 was fandiddly-tastic, to which I enquired if he could share his recipie and he declined, to which I said that without the fandidly-tastic optimisations that modern dacs would beat a non fandangled TDA1541. And by beat, I mean in a multitude of ways which inclues the cost of the circuit, and accesabiltiy to ordinary people - who are not able to buy an AMR77. Sound quality vs cost.
This is DIY Audio - the place were we ( some of us
Now if you came to pick on me because I said one thing against the TDA1541 you really have chosen the wrong person
Pick on you ? NO - how could I - you have infinitely more electronics knowledge than I do, you understand the figures and can make comparisons.
I think I get a bit fed up of listening to everyone else's opinions on sound. How on earth can anyone tell me this ? I am a Pianist - I may listen in a different way - when my system improves because of something I do I absolutely know it's an improvement. I don't then force my opinions on the rest of the members.
FWIW, I dont like the sound of TDA1541 when it is used with a SAA7220.
So, that is why I run my TDA1541s in NOS mode. To me this sounds better.
I tried NOS for a while but went back to a well sorted 7220...see....we're all different
... my experience comes from what I have heard. If you think my experience is not good enough. Well OK, there is nothing I can do about that - you will just have to ignore my posts.
Mine too - I had the misfortune of working in a high end retailer for several years. The crap in there is incredible - some of the nice stuff was equally incredible. It never leaves you.
Your experience is not in question - we are music lovers FIRST - not the other way round.
The argy bargy that goes on regarding sound is really quite funny. If you hear a good sounding sound system, its a good sounding system. It does not matter what makes up the system as long as it all works well together.
This is the best part of your reply - could not have worded it better myself.
In truth, I felt a bit queasy watching Thorsten getting questioned in such a way - he doesn't need me to defend him of course - he's got the answers, experience and knowledge which he very eloquently demonstrated every time.
It was however like watching someone tell Nelson Pass he's wrong, deaf and can't solder
Erin - it was not personal and do accept my apologies if it seemed that way
Hi,
I tried several of the ones that where "final" at the time John choose graciously to share them with the DIY-Audio crowd, they all worked. I would expect the most recent versions to be "the best", John is both exceptionally talented and exceptionally prolific, awesome.
Ciao T
I tried several of the ones that where "final" at the time John choose graciously to share them with the DIY-Audio crowd, they all worked. I would expect the most recent versions to be "the best", John is both exceptionally talented and exceptionally prolific, awesome.
Ciao T
Thank you Thorsten. You picked up on the problem I was having, how they were all "final" until the next one......
I will go forth and build the circuit.
Last edited:
Hi,
Was that Srajan's original 2006 review unit? I am not aware of any others ever having gone to Crete. If so, the CD-77 has had a fair few upgrades since, as had the Buffalo.
Ciao T
Was that Srajan's original 2006 review unit? I am not aware of any others ever having gone to Crete. If so, the CD-77 has had a fair few upgrades since, as had the Buffalo.
Ciao T
John's dac doesnt do anything for me, but there is no denying the scale of the project, the detail in which it is covered and the generosity with which it is given away. a chip that needs such babying to sound its best, well, it doesnt make sense to me, but some people like that sort of thing. i suppose i've put almost the same amount of care into my dac because its fun also
however he has published some excellent fet regulators and cap multipliers over the years. the floating fet regs with common mode filter for his circlotron power amp will come in handy.
i will admit i havent searched out an ultimate version using this chip, but i just didnt have the desire to
however he has published some excellent fet regulators and cap multipliers over the years. the floating fet regs with common mode filter for his circlotron power amp will come in handy.
i will admit i havent searched out an ultimate version using this chip, but i just didnt have the desire to
Last edited:
TL, it was a 2nd hand unit (bought from some other European country if I recall) and it was not in Crete. Not sure of versions etc but I don't think it was ever sent back to AMR for any upgrades. It would have been mentioned by the owner, I guess.
Buffalo was a BuffaloII with IVY3. It has since been upgraded with better supplies.
Transports where different of course so this must have also played a role. AMR's was the obvious one while Buffalo played through a laptop equiped with an M2Tech EVO USB transport (with good DIY power supply). We had extracted the same CD to wav.
Buffalo was a BuffaloII with IVY3. It has since been upgraded with better supplies.
Transports where different of course so this must have also played a role. AMR's was the obvious one while Buffalo played through a laptop equiped with an M2Tech EVO USB transport (with good DIY power supply). We had extracted the same CD to wav.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.