YES! Quiet those grounds ...
I had an Onkyo DX5700. It was a very well built player for it's time (one of the old Integra series). I spent some time and effort on Jocko's #3 by making the already decent grounds quieter,doing a number of things such as improving (i.e. NOT just adding expensive caps) bypassing. Made a significantly audible difference that to me was much better!
Do it. Don't ask questions. Trust the Jocko. Just do it.
mlloyd1
p.s. And in case you're wondering, the DX5700 laser died (thanks to an abusive loanee - you think I'd learn after a while) or else I'd be listening to it still. And no, I won't part with the carcass, I still have hopes of finding a replacement part via ebay.
p.p.s And also, I don't have a spectrum analyzer. I do, however, have a scope. And, I did spend a lot of time reading about bypassing, breadboarding, testing, measurement techniques, etc. in some of the execellent tech notes and app notes from Linear Technologies (look for anything with Jim Williams name on it), Analog Devices, and anything with Bob Pease' (from National Semi) name on it. I'm sure there are others, these come to mind immediately for me.
I had an Onkyo DX5700. It was a very well built player for it's time (one of the old Integra series). I spent some time and effort on Jocko's #3 by making the already decent grounds quieter,doing a number of things such as improving (i.e. NOT just adding expensive caps) bypassing. Made a significantly audible difference that to me was much better!
Do it. Don't ask questions. Trust the Jocko. Just do it.
mlloyd1
p.s. And in case you're wondering, the DX5700 laser died (thanks to an abusive loanee - you think I'd learn after a while) or else I'd be listening to it still. And no, I won't part with the carcass, I still have hopes of finding a replacement part via ebay.
p.p.s And also, I don't have a spectrum analyzer. I do, however, have a scope. And, I did spend a lot of time reading about bypassing, breadboarding, testing, measurement techniques, etc. in some of the execellent tech notes and app notes from Linear Technologies (look for anything with Jim Williams name on it), Analog Devices, and anything with Bob Pease' (from National Semi) name on it. I'm sure there are others, these come to mind immediately for me.
Jocko Homo said:
Jocko's on the ball
I've not tried the I/V mods, but the clock and PSU mods are BIG winers.
I disagree with this, irrespective of other circuitry, lower jitter is always better - I've not yet found a value below which improvements cannot be heard. A 1ps change is audible.
Additionally, consider the phase noise (jitter) of the oscillator as a modulating mechanism and a whole load of other effects become apparent - the jitter figure does not tell all, the spectral content and purity of the clock is an audible effect, with different spectra having different sonic results.
For PSU's the same modulation mechanism applies, as PSU noise modulates the clock.
A.
I've not tried the I/V mods, but the clock and PSU mods are BIG winers.
I disagree with this, irrespective of other circuitry, lower jitter is always better - I've not yet found a value below which improvements cannot be heard. A 1ps change is audible.
Additionally, consider the phase noise (jitter) of the oscillator as a modulating mechanism and a whole load of other effects become apparent - the jitter figure does not tell all, the spectral content and purity of the clock is an audible effect, with different spectra having different sonic results.
For PSU's the same modulation mechanism applies, as PSU noise modulates the clock.
A.
Two things to look for:
1 Look at the rails. If you see 60 Hz, and its harmonics, then there is a problem with the ground return for the filter caps, in regards to where the 3-terminal regulators get their reference.
2 Look at the output. It should be white noise, without any nasty hash in the audio band. Could indicate problems with the DAC or I/V reference points. These have to be clean. (Just like the "quiet" ground in a power amp.)
Granteed, a well designed unit would not have these problems. Maybe yours won't. But I have seen them.
As for jittter, yes, the frequency and the amount are both things to look at. Getting rid of the noise on the PSU will help bring those numbers down.
Jocko
1
Look at the rails. If you see 60 Hz, and its harmonics, then there is a problem with the ground return for the filter caps, in regards to where the 3-terminal regulators get their reference.2
Look at the output. It should be white noise, without any nasty hash in the audio band. Could indicate problems with the DAC or I/V reference points. These have to be clean. (Just like the "quiet" ground in a power amp.)Granteed, a well designed unit would not have these problems. Maybe yours won't. But I have seen them.
As for jittter, yes, the frequency and the amount are both things to look at. Getting rid of the noise on the PSU will help bring those numbers down.
Jocko
50 (60Hz) and it's siblings
Agreed, or to put it another way, there's a problem with where the circuit is referenced to, with regard to the regulator Kelvin sense points.
This old thread illustrates your some of your points very graphically (http://www.diyaudio.com/forums/showthread.php?threadid=985&perpage=15&highlight=317&pagenumber=1). Whilst I've moved on a bit from here, most of the techniques presented are valid.
With supplies as noisy as those in the CD63 KIS I measured, it's never going to sound any good.
There's also a good ilustration of the 60Hz (or 50Hz in my case) problem - worth a look for others that think that even a supposedly tweaked player is well-engineered.
I'd like to have a chat with Mr. Ishiwata...
Andy.
P.S. On the soundcard front, I've just upgraded the adequate, but severely limited SBLive to an M-Audio 24/96 card - fantastic as a front end to the FFT analyser, and much cheaper than an AP... now I can get above the audio band a bit - before I was blind
Agreed, or to put it another way, there's a problem with where the circuit is referenced to, with regard to the regulator Kelvin sense points.
This old thread illustrates your some of your points very graphically (http://www.diyaudio.com/forums/showthread.php?threadid=985&perpage=15&highlight=317&pagenumber=1). Whilst I've moved on a bit from here, most of the techniques presented are valid.
With supplies as noisy as those in the CD63 KIS I measured, it's never going to sound any good.
There's also a good ilustration of the 60Hz (or 50Hz in my case) problem - worth a look for others that think that even a supposedly tweaked player is well-engineered.
I'd like to have a chat with Mr. Ishiwata...
Andy.
P.S. On the soundcard front, I've just upgraded the adequate, but severely limited SBLive to an M-Audio 24/96 card - fantastic as a front end to the FFT analyser, and much cheaper than an AP... now I can get above the audio band a bit - before I was blind
I did not discus that lowest jitter is not important. I assume you measured accurately down to 1ps and you have done extensive listening tests. Up to my knowledge 1ps is the measurement limit of modern advanced (and very expensive) equipment. And yes, some go a fraction lower, but not much. But if you can manage a jitter on the L/R clock pin of the DA-chip of 5 ps (RMS) you did done a pretty good job.
The SB live is pretty good up to 20 khz but only at 48 kHz sample rate
I am using the M-Audio 2496 too and yes it is better than the SB Live. You can mod the M-Audio by extra filtering the + and – 12V lines. I put 22 ohm resistors in the lines and added 220uF ultra low esr capacitors paralleled with 220 nF chip capacitors after the resistors. It gave me a 8 dB lower noise floor. With averaging this card shows a noise floor below –120 dB.On the regulator front: Has any one tried modern LDO regulators? They look promising. I am considering of replacing the 7805’s with 7806’s or LDO 8V regulators and putting a LDO 5V regulator behind it.
Sound cards
Sorry - slight thread hijack
"You can mod the M-Audio by extra filtering the + and – 12V lines. I put 22 ohm resistors in the lines and added 220uF ultra low esr capacitors paralleled with 220 nF chip capacitors after the resistors. It gave me a 8 dB lower noise floor. With averaging this card shows a noise floor below –120 dB."
I was thinking of trying this - you've just made my mind up for me!
Any chance of a pic to save me a bit of time, it would be greatly appreciated.
I also feel those 5532's could be bettered, not so much in noise, but in dynamic performance, maybe some AD8610's...
Regards,
Andy.
Sorry - slight thread hijack
"You can mod the M-Audio by extra filtering the + and – 12V lines. I put 22 ohm resistors in the lines and added 220uF ultra low esr capacitors paralleled with 220 nF chip capacitors after the resistors. It gave me a 8 dB lower noise floor. With averaging this card shows a noise floor below –120 dB."
I was thinking of trying this - you've just made my mind up for me!
Any chance of a pic to save me a bit of time, it would be greatly appreciated.
I also feel those 5532's could be bettered, not so much in noise, but in dynamic performance, maybe some AD8610's...
Regards,
Andy.
Hi Andy,
I don’t have a digital camera at hand at the moment. Maybe next week. I will see if I can draw them in on the digital picture of the card from the midi-man site to speed it up.
Replacing the NE5532’s seems to me not a good idea. They are well-regarded op-amps used in many high-end mixing consoles and very low noise. You do a load better by keeping the signal lines far away from noisy digital lines AND most important the power cord! I did do a distortion measurement in loop trough ( with a 10 cm cable). Distortion is much lower than the codec specs from the data sheet. And in the end I can compensate for it in my software when needed.
Peter
I don’t have a digital camera at hand at the moment. Maybe next week. I will see if I can draw them in on the digital picture of the card from the midi-man site to speed it up.
Replacing the NE5532’s seems to me not a good idea. They are well-regarded op-amps used in many high-end mixing consoles and very low noise. You do a load better by keeping the signal lines far away from noisy digital lines AND most important the power cord! I did do a distortion measurement in loop trough ( with a 10 cm cable). Distortion is much lower than the codec specs from the data sheet. And in the end I can compensate for it in my software when needed.
Peter
NE5532???
Are you serious about this? Ar the time of introduction these were stellar devices but today there are so many much better opamps.
NE5532 & NE5534 are in my oldest Sigenetics databook of 1981. The pages have turned yellow now.... Even a AD712 was a big improvement at the time. High-end mixing consoles? No wonder some records sound bad!
Hi Pjotr,Pjotr said:Hi Andy,
Replacing the NE5532’s seems to me not a good idea. They are well-regarded op-amps used in many high-end mixing consoles and very low noise.
Are you serious about this? Ar the time of introduction these were stellar devices but today there are so many much better opamps.
NE5532 & NE5534 are in my oldest Sigenetics databook of 1981. The pages have turned yellow now.... Even a AD712 was a big improvement at the time. High-end mixing consoles? No wonder some records sound bad!
Hello Pjotr ( Remco ? ),
You must be joking about the NE5532/NE5534 or you haven't been listening to hifi for the last 10 years. Even a NJM4580 sounds better than it !
At its time there simply was nothing better. But times have changed and opamps like OP249, OPA275, OPA2604 etc. etc. are all better in specs and sound quality.
I dare to say that replacement of this old chip always leads to better sound, lower noise, higher slew rate, lower offset etc. etc.
Talking about studio equipment:
a friend of mine changes all the opamps to OPA types in his DDA mixing console because of the lousy sound quality of the used TL72 and NE5532.
You must be joking about the NE5532/NE5534 or you haven't been listening to hifi for the last 10 years. Even a NJM4580 sounds better than it !
At its time there simply was nothing better. But times have changed and opamps like OP249, OPA275, OPA2604 etc. etc. are all better in specs and sound quality.
I dare to say that replacement of this old chip always leads to better sound, lower noise, higher slew rate, lower offset etc. etc.
Talking about studio equipment:
a friend of mine changes all the opamps to OPA types in his DDA mixing console because of the lousy sound quality of the used TL72 and NE5532.
Hi,
Oops, The detailed picture of the printed circuit board is not on the m-audio site anymore. They revisited the site in the meantime. When I have time, I will take some pictures and send them by e-mail to you Andy. I also put a 22 ohm resistor in the 12V line going to the 7805 that is powering the AK4528 codec and a 220 uF cap at the input of the 7805.
Elso and Jean-Paul, I fully agree that there are much better op-amps available at the moment from a sonically viewpoint. But when it concerns the M-audio 2496: It is powered from + and – 12V. Besides that the opa’s are smd duals. Maybe the OPA2604 will be a candidate, but it is noisier than the 5532’s. I used this one a lot for active filtering. But to my experience these do not perform very well at + and – 12V. They do a lot better at + and – 18V.
Anyway I do not like integrated op-amps at all for audio. For the real stuff I’ll stick with discrete build ones. I do not have much time for DIY at the moment, but when it is actual it is worthwhile to start a thread about discrete opa’s and investigate the merits of it and the loop-feedback involved.
I am now using the M-Audio SuperDac for a month hooked up to my CD player. It is full of 5534’s The circuit is a straight copy from the AKM application notes. It is amazing to see (by listening) that this ugly cheap little box outperforms many high-end DAC’s costing many times more, straight out of the box. It doesn’t have that “synthetic” high and mid many DAC’s suffer. It sounds very musical. For those of you who can read Dutch here is a good review of it, although there is some b*llsh*t in it: There is no 40 Mhz super risk audio processor at all in the superdac, there is a tiny pic-microcontroller for handling the front buttons, leds and setting up the receiver and dac control registers.
I am on the way to mod this nice DAC: First I will remove the whole op-amp mess and replace it by good line transformers, the dac chip can drive more than 8V pp symmetrical straight into 600 ohms. Second I will try to improve the power lines. And last but not least, I will put a nice ultra low jitter re-clocker in it with a VXCO.
By the way Jean-Paul, I am NOT Remco
Cheers
Oops, The detailed picture of the printed circuit board is not on the m-audio site anymore. They revisited the site in the meantime. When I have time, I will take some pictures and send them by e-mail to you Andy. I also put a 22 ohm resistor in the 12V line going to the 7805 that is powering the AK4528 codec and a 220 uF cap at the input of the 7805.
Elso and Jean-Paul, I fully agree that there are much better op-amps available at the moment from a sonically viewpoint. But when it concerns the M-audio 2496: It is powered from + and – 12V. Besides that the opa’s are smd duals. Maybe the OPA2604 will be a candidate, but it is noisier than the 5532’s. I used this one a lot for active filtering. But to my experience these do not perform very well at + and – 12V. They do a lot better at + and – 18V.
Anyway I do not like integrated op-amps at all for audio. For the real stuff I’ll stick with discrete build ones. I do not have much time for DIY at the moment, but when it is actual it is worthwhile to start a thread about discrete opa’s and investigate the merits of it and the loop-feedback involved.
I am now using the M-Audio SuperDac for a month hooked up to my CD player. It is full of 5534’s
The circuit is a straight copy from the AKM application notes. It is amazing to see (by listening) that this ugly cheap little box outperforms many high-end DAC’s costing many times more, straight out of the box. It doesn’t have that “synthetic” high and mid many DAC’s suffer. It sounds very musical. For those of you who can read Dutch here is a good review of it, although there is some b*llsh*t in it: There is no 40 Mhz super risk audio processor at all in the superdac, there is a tiny pic-microcontroller for handling the front buttons, leds and setting up the receiver and dac control registers.I am on the way to mod this nice DAC: First I will remove the whole op-amp mess and replace it by good line transformers, the dac chip can drive more than 8V pp symmetrical straight into 600 ohms. Second I will try to improve the power lines. And last but not least, I will put a nice ultra low jitter re-clocker in it with a VXCO.
True, but the bad sound of most CD’s is to my opinion not due to the equipment but due to the sound engineers handling it. Do you recognise that there are not so many CD’s that preserve the correct and full phase relations between left and right channels? Most of it is mixed out of it or there is artificial “spaciness”. There is an old studio saying: If it does not sound we put it louder, if still does not sound, put more reverb in it
By the way Jean-Paul, I am NOT Remco
Cheers
Bad CD-Sound
Bad sound???
I have <B>gorgeous</B> sound from CD!!!
See my points in this thread and also Jocko's three points of course.
Hi Pjotr,Pjotr said:but the bad sound of most CD’s is to my opinion not due to the equipment but due to the sound engineers handling it. Do you recognise that there are not so many CD’s that preserve the correct and full phase relations between left and right channels? Most of it is mixed out of it or there is artificial “spaciness”. There is an old studio saying: If it does not sound we put it louder, if still does not sound, put more reverb in it
Cheers
Bad sound???
I have <B>gorgeous</B> sound from CD!!!
See my points in this thread and also Jocko's three points of course.
Dual op-amps
Sounds like a job for AD8620's.
A bit of noise here could even be beneficial, assuming no dither is added to the analogue inputs to the DAC?
AD86x0's are about the best op-amp I've ever heard, the dual one has stunning channel seperation.
Not sure package size is the same though...
A.
Sounds like a job for AD8620's.
A bit of noise here could even be beneficial, assuming no dither is added to the analogue inputs to the DAC?
AD86x0's are about the best op-amp I've ever heard, the dual one has stunning channel seperation.
Not sure package size is the same though...
A.
Re: Bad CD-Sound
It seems to me that you missed my point completely. I know that there are many <B>gorgeous</B> CD's out there, TELDEC ones for instance. But the mainstream ...Elso Kwak said:
Hi Pjotr,
Bad sound???
I have <B>gorgeous</B> sound from CD!!!
See my points in this thread and also Jocko's three points of course.
RECORDING PROCESS.
Hi,
There is nothing you can do about badly recorded material other than NOT buy it.
The better your player and DAC the more it is going to be obvious.
With vinyl recordings the exact same problem was there too.
If the whole recording is out of phase a change of polarity is required,so such a switch is a must have on the preamp or output stage of a DAC.
Other than that?
Ciao,
Hi,
There is nothing you can do about badly recorded material other than NOT buy it.
The better your player and DAC the more it is going to be obvious.
With vinyl recordings the exact same problem was there too.
If the whole recording is out of phase a change of polarity is required,so such a switch is a must have on the preamp or output stage of a DAC.
Other than that?
Ciao,
Pjotr said:
There may be some truth in that.
I'm playing a CD at the moment -Nora Jones, come away with me-
It sounds gorgous in stereo, or (single channel) but is distorted if I mono it.
Now either they, or my CD player is introducing a variable delay / reverb between L $ R. Most likely they "spacialized" it.
Cheers,
There may be some truth in that.
I'm playing a CD at the moment -Nora Jones, come away with me-
It sounds gorgous in stereo, or (single channel) but is distorted if I mono it.
Now either they, or my CD player is introducing a variable delay / reverb between L $ R. Most likely they "spacialized" it.
Cheers,
SPACIALIZED
Hi,
Love this neologism!
If both channels are summed you should not get an increase in distortion levels.
Try this with a number of different recordings and if it still does distort then something is very wrong with the replay system.
An interesting trick is to rotate phase on one channel and not on the other.
On some recordings that can reveal things you never thought were there.
Very funny.
Caio,
Hi,
Love this neologism!
If both channels are summed you should not get an increase in distortion levels.
Try this with a number of different recordings and if it still does distort then something is very wrong with the replay system.
An interesting trick is to rotate phase on one channel and not on the other.
On some recordings that can reveal things you never thought were there.
Very funny.
Caio,
Hi Pjotr,
You are right about the noise. Almost all Fet input opamps ( except OPA627/637 ) are noisier than NE5532. Sonically they are better though. To me noise isn't as important as sonical quality. Reason for manufacturers to use this old workhorse is its price ( and the low distortion maybe ).
You are wrong about the SMD thing, most are still available in DIL.
I am very curious about the M-Audio SuperDAC. I read reviews stating that a RISC CPU was used, but you said it contains a AK chip. With which DAC's did you compare this M-Audio ?
What about built quality ?
Jean-Paul
You are right about the noise. Almost all Fet input opamps ( except OPA627/637 ) are noisier than NE5532. Sonically they are better though. To me noise isn't as important as sonical quality. Reason for manufacturers to use this old workhorse is its price ( and the low distortion maybe ).
You are wrong about the SMD thing, most are still available in DIL.
I am very curious about the M-Audio SuperDAC. I read reviews stating that a RISC CPU was used, but you said it contains a AK chip. With which DAC's did you compare this M-Audio ?
What about built quality ?
Jean-Paul
Hi Andy,
The AD8620 looks promising and it is unity gain stable too. You can probably drop the SOIC R-8 package straight in. If you do, get also rid of the electrolytes on the cinch sockets. I didn’t look into this op-amp because they are hard to get here. When I ask for samples from AD I am forwarded to the Dutch representative. That one is not willing to handle small quantities, even not samples, when it is not backed up by large orders. But if you can hand me 4 samples it is worth the try I am mainly using the M-audio card for doing measurements and not much for extensive music listening. But if the AD8620 also lowers distortion it is definitely worth the effort for me also.
Jean-Paul, I listened to many DAC’s in the past, from X-DAC to Wadia. But they do not involve me into the music. The top Denon DVD players with Apha processing do more and also the Shigaraki DAC to my own amazement. I am not saying the SuperDac is the best DAC ever but it does involve me into the music without tiring. The dac-chip used is the AKM AK4393. The 4 serial digital audio lines go straight to the CS8414 receiver. Putting a re-clocker into these lines is fairly easy. Building the re-clocker itself is not that simple. Build quality is fairly good. Printed circuit board layout is tidy and well done. But oh boy, that 2N3904 muting transistors across the connectors at the output… Look at the review link I gave, what is said about the sound I can only agree with nothing to ad.
Ok Elso
I will send you an e-mail about the fet-xtal oscillator. Good ready made VXCO’s are hard to get. At least it will take some months.
The AD8620 looks promising and it is unity gain stable too. You can probably drop the SOIC R-8 package straight in. If you do, get also rid of the electrolytes on the cinch sockets. I didn’t look into this op-amp because they are hard to get here. When I ask for samples from AD I am forwarded to the Dutch representative. That one is not willing to handle small quantities, even not samples, when it is not backed up by large orders. But if you can hand me 4 samples it is worth the try
I am mainly using the M-audio card for doing measurements and not much for extensive music listening. But if the AD8620 also lowers distortion it is definitely worth the effort for me also.Jean-Paul, I listened to many DAC’s in the past, from X-DAC to Wadia. But they do not involve me into the music. The top Denon DVD players with Apha processing do more and also the Shigaraki DAC to my own amazement. I am not saying the SuperDac is the best DAC ever but it does involve me into the music without tiring. The dac-chip used is the AKM AK4393. The 4 serial digital audio lines go straight to the CS8414 receiver. Putting a re-clocker into these lines is fairly easy. Building the re-clocker itself is not that simple. Build quality is fairly good. Printed circuit board layout is tidy and well done. But oh boy, that 2N3904 muting transistors across the connectors at the output… Look at the review link I gave, what is said about the sound I can only agree with nothing to ad.
Ok Elso
I will send you an e-mail about the fet-xtal oscillator. Good ready made VXCO’s are hard to get. At least it will take some months.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.