You are 100% right!! The 22k and 150pF on pin5, should not be there. This is a mistake, since I copy/pasted this part from another project. I will opdate the schematic.
Hurtig said:
if you just stick any transformer into 75 ohm line, then your statement is correct. If not, then is wrong.
You are right.... Not all signals are sine waves. But this do not really matter in this case. Thing is, that if the signal is not a sine wave but some sort of peak, the RMS value is lower than 2V.
If you have a signal from let's say CD player, with an nominal amplitude of 2Vrms, the peak voltage can not get higher than 2V*1,41 = 2,82Vpeak. That is beyond all doubt and woodoo.
The idea is, that the peak voltage of a sine wave at 0dB, represents the maximum peak voltage that the output can ever give. A 2Vrms sine Wave will have a peak value of 2,82V.
If you measure with a scope, you will notice, that most of the time, the output is less than 10% of the nominal voltage, and that it very seldom gets even close to the maximum peak voltage.
The theory that the peak voltage in an audio device can get higher than 1,41*nominal RMS voltage, is the result of a general misunderstanding of signals.... Any engineer should agree to that (or take the signal conditioning course again).
If you have a signal from let's say CD player, with an nominal amplitude of 2Vrms, the peak voltage can not get higher than 2V*1,41 = 2,82Vpeak. That is beyond all doubt and woodoo.
The idea is, that the peak voltage of a sine wave at 0dB, represents the maximum peak voltage that the output can ever give. A 2Vrms sine Wave will have a peak value of 2,82V.
If you measure with a scope, you will notice, that most of the time, the output is less than 10% of the nominal voltage, and that it very seldom gets even close to the maximum peak voltage.
The theory that the peak voltage in an audio device can get higher than 1,41*nominal RMS voltage, is the result of a general misunderstanding of signals.... Any engineer should agree to that (or take the signal conditioning course again).
Well... feel free to suggest a pulse transformer that do not introduce any jitter or any other sonic performance issues.
We have tried quite a few, and the transformerless approach has always been the best sonic performer. Btw...: Crystal Semiconductor actualle suggest the transformerless version in their datasheet.
I agree that is certainly true on "fixed pixel" devices like a DAC. If 2V RMS is the most it will do, then 2.83 volts peak is all you'll ever get at 0dB full scale. Both theory and measurement show this.
And although much music has an average level of 12-18dB below peak, just about all tracks will hit the peak value a few times. (Quiet classical passages being a noted exception).
So I'm not sure what Ian means with the 10V value. Certainly with analog stuff - especially phono - you have to be ready for wild peaks. Maybe he was referring to 1V RMS being the average line level value with peaks being 20dB higher?
But that's one of the nice things about digital, you know right were the max level will be.
And although much music has an average level of 12-18dB below peak, just about all tracks will hit the peak value a few times. (Quiet classical passages being a noted exception).
So I'm not sure what Ian means with the 10V value. Certainly with analog stuff - especially phono - you have to be ready for wild peaks. Maybe he was referring to 1V RMS being the average line level value with peaks being 20dB higher?
But that's one of the nice things about digital, you know right were the max level will be.
The caps in the power supply are all for reserve power and ripple suppression, you can use whatever caps you like at the chips for decoupling. Personally, regarding transient current supply and low ESR (and price) I've always had good experiences with a tantalum/ceramic combo at the chips.
I really liked the idea about using resistors to incorporate a low-pass filter into the power supply, using the first stage smoothing caps. I'm playing around with values right now, and it looks promising.
I think there's something of a majority opinion here that a pulse transformer is good idea. I've always considered this good design practice, and without measurements to go by, would say that a well designed unit shouldn't appreciably increase jitter. There's been a lot of discussion on the forums about which are best. Has anybody had any experience with Lundahl Transformers?
Lundahl Catalog Page
Would be helpful if you stick to tech debate, and less "advertising"
Neither is it good too be negative towards others products, to try and make yourself look better
Neither is it good too be negative towards others products, to try and make yourself look better
Sorry, I wasn't very clear was I😉
As Panomaniac says, I was eluding to the fact that old school music was recorded with a 20dB headroom above nominal to allow for transients and the dynamics of music. These days with everything compressed, you only really need 3-6dB of headroom.
I have seen sources that put out music at 1V average (perhaps that's a better name) and the occasional peak instantaneously hitting close to 10V. So if you're designing an op-amp circuit to handle this and it's driving a 600ohm load then you will take the op-amp out of class A.
It's all moot here, because 1: you know the absolute ceiling of the DAC output 2: you're (hopefully) driving a low impedance load.
Thanks for the link Kurt, I'll have a read of that.
Digital Audio Pulse Transformers
Having seen a presentation and demo by Jon Paul of Scientific Conversion, I believe these transformers have been well designed. I have a few but haven't done any comparative testing yet.
Scientific Conversion, Inc. - Transformers and Inductors
I have used Lundahl audio transformers in the past and can vouch for their excellence. Per Lundahl knows transformers inside out. Does he also make digital pulse transformers?
Having seen a presentation and demo by Jon Paul of Scientific Conversion, I believe these transformers have been well designed. I have a few but haven't done any comparative testing yet.
Scientific Conversion, Inc. - Transformers and Inductors
I have used Lundahl audio transformers in the past and can vouch for their excellence. Per Lundahl knows transformers inside out. Does he also make digital pulse transformers?
Why don't use Wolfson receiver WM8805??
To USB my suggestion is not user Texas PCM and use the Tenor TE7022L it's 24bit 96khz.
regards,
João Martins
To USB my suggestion is not user Texas PCM and use the Tenor TE7022L it's 24bit 96khz.
regards,
João Martins
Hi
Tantalums is IMHO not good for audio, we´ve tried them out, and I´ve also listened to gear utilizing tantalums with no go - i.e. Linn CD12 and others.
Pulse transformers is also a difficult matter, since a transformer is the very worst audio component at all, only second to disruptions and short circuits.
Measuring the shape of the digital signal in a SPDIF output with and without transformer is a pretty nice experience.
Transformer transmission completely corrupts the shape of the signal.
Hence they introduce incoming jitter.
I agree that Scientific transformers have the very best data of all pulse transformers, which the whole game is about.
Kurt,
I read the thread by werewolf (I remember reading it a few years back now but the refresher was good)
Werewolf says here
http://www.diyaudio.com/forums/digi...nous-sample-rate-conversion-2.html#post348102
that multiple clocks can cause intermodulation and introduce noise into the audio.
It would seem to me that using a 25MHz MCLK for the DAC requires the use of a PLL in the DAC to generate the 128/256*Fs bit clock which in turn introduces jitter as well as intermodulation.
Using the popular 12.288 or 24.576MHz MCLK would allow the ASRC and DAC to be tied to clean clock that has an integer divide down to the output sample rate, at least eliminating intermodulation of clocks and hopefully adding very little jitter to MCLK.
Guess I need to read up a bit on your DAC choice
<edit> at least where the output sample rate is 48K, 96K or 192K
I read the thread by werewolf (I remember reading it a few years back now but the refresher was good)
Werewolf says here
http://www.diyaudio.com/forums/digi...nous-sample-rate-conversion-2.html#post348102
that multiple clocks can cause intermodulation and introduce noise into the audio.
It would seem to me that using a 25MHz MCLK for the DAC requires the use of a PLL in the DAC to generate the 128/256*Fs bit clock which in turn introduces jitter as well as intermodulation.
Using the popular 12.288 or 24.576MHz MCLK would allow the ASRC and DAC to be tied to clean clock that has an integer divide down to the output sample rate, at least eliminating intermodulation of clocks and hopefully adding very little jitter to MCLK.
Guess I need to read up a bit on your DAC choice
<edit> at least where the output sample rate is 48K, 96K or 192K
...and werewolf reinforces that the ASRC cannot "fix" any input clock jitter so therefore the recovered clock in the SPDIF input must be as clean as possible. This is probably the best argument for using a transformer to eliminate the jitter caused by induced currents in the connecting cable.
Hurtig said:
Sonic performanse issues....read that datasheet again and change PLL filter values if you want to to increase performanse issues.
Change phase detector update rate...another sonic performance issues.
Run CS8416 differentially, not SE....another sonic performance issues.
Stick line driver infront of CS8416....another performance issues.
You have tried quite a few transformers..... I suppose you run them into 75 Ohm transmission line? 😀
Link to another Crystal datsheet CLICK ME
Attachments
Kurt von Kubik said:
Disagree.
Wrong impedance is corrupting your signal, not transformer.
Or wrong transformer, if you follow Crystal recommendation 😀
Lowest interwinding capacitance = highest leakage inductance 😡
Last edited:
Compressed or non compressed.... A CD player with a nominal output of 2Vrms, can not give peaks above 2,82V. The average outtput however, will be way below 2V.... But that is the nature of music.
@ Stormsonic:
What we did about the SPDIF is:
- Remove the pulse transformer i the CD-drive, and made a true 75ohm out put.
- Remove the pulse transformer in the DAC input, and terminate with a 75ohm resistor.
This has proven to be the best sonic performer. And trust me... we did try a few...
What we did about the SPDIF is:
- Remove the pulse transformer i the CD-drive, and made a true 75ohm out put.
- Remove the pulse transformer in the DAC input, and terminate with a 75ohm resistor.
This has proven to be the best sonic performer. And trust me... we did try a few...
- Status
- Not open for further replies.