That's the one I was trying to recall, thanks. What's a factor of a k between friends in jitter anyway?
(nice sig BTW. but JMK is defunct as of now which makes for an interesting problem
That's precisely why I quote it - for the irony factor
You're the first to pick up on that - kudos.I think this is exactly abraxalito's point as it's all dependent on cable construction and properties.
this makes me wonder... how come galvanic isolation with high end DACs is not the norm?
speaking of which, did I see some noise measurements at CA?
Neither of those handles UAC2 - they're full speed [12Mbits] (as opposed to high speed [480Mbits]).
Its from my perspective part of the problem because of this currently fashionable word 'galvanic' - meaning its DC isolation. But DC noise isn't the problem, AC is - isolation at RF is called for which is a different matter from isolation at DC.
Regarding CM noise measurements - I also recall seeing some on CA, they originated from ExaDevices I think.
Last edited:
Common mode rejection is more a function of the USB receiver circuitry, not the cable.
Could be - I have given up on CA except for the section on hi-res downloads.
That's the beauty of a solution based on magnetics (transformer) - it provides both DC isolation and a low-pass filter.
True for common-mode signal (carried on the data pair) but not for CM noise which is carried on the ground too.
I guess some of us might be sourpusses enough to ask for some kind of proof of the validity of your theory...
Ah, yes, so the people selling the stuff report differences. Right....
Bah-humbug; here we go
Actually, there can be differences in cables, but they are identifiable and measurable. I don't need to elaborate on impedence around here, and reactance is measurable. There are EMR and RFI consideration in some cases. Differences when they are not born in a overly imaginary mind are real.
Jim
I don't think anyone is claiming there aren't differences in cables - I think the disagreement is about the audibility of those differences...
...and to Pano's point, the effect (if any) on the analog output signal, the only thing that counts. Even if the effects aren't readily audible, if they're measurable and significant, some of us would implement the improvements anyway, just in case.
You sure did. Tried to measure uv levels after the introduction of a JLH filter after a 9 v. TeraDak. Used VA Analyzer (free software) and a netbook. Readings are tenuous though, because I don't have a clue on +- error, and precise calibration is a problem until I build a resistance ladder to isolate say 25 uv (and even then I have no way of verifying that...).
Jim
Last edited:
If I may, I would like to go back to the original subject of this thread.
To get all of the potential out of the Peachtree Dacit with USB implemented here, as I have written, I have replaced th 9 v. main ps to the Dac, and 5 v. USB with external ps's followed by JLH filters. This has made a big difference in it's performance. All that remains is dealing with the data input via USB to the Dacit.
Performance wise, there seems to be nothing to do relative to bit perfection, since the system performs very well, despite it's adaptive architecture.
Jitter remains a concern. To reclock the USB data stream can be done with a SOTM USB PCI, but I don't want to spend $350, nor am I sure how accurate their reclocking
is.
This is why I asked the question of this thread. I wanted to understand the nature of the USB signal path and the characteristics of this signal so that I could consider a DIY project to assure that the data stream input to the Dacit is corrected for minimal jitter. "No garbage in, no garbage out"
Peachtree claims 3 ps correction for jitter, but this really means they claim 3 ps correction at their clock. There is no published data on what correction actually is for the system.
By way of the kind suggestions of several members here, I have become aware of the I2S FIFO kit, with measured 1.5 ps correction. This appears to be the solution I have been looking for. I have looked at some of the information, it is lengthy.
My concern is what would I have to do to implement FIFO into the Dacit via a USB cable input from the pc? The Dacit has no I2S input, and there is no schematic for the Dacit?
Jim
To get all of the potential out of the Peachtree Dacit with USB implemented here, as I have written, I have replaced th 9 v. main ps to the Dac, and 5 v. USB with external ps's followed by JLH filters. This has made a big difference in it's performance. All that remains is dealing with the data input via USB to the Dacit.
Performance wise, there seems to be nothing to do relative to bit perfection, since the system performs very well, despite it's adaptive architecture.
Jitter remains a concern. To reclock the USB data stream can be done with a SOTM USB PCI, but I don't want to spend $350, nor am I sure how accurate their reclocking
is.
This is why I asked the question of this thread. I wanted to understand the nature of the USB signal path and the characteristics of this signal so that I could consider a DIY project to assure that the data stream input to the Dacit is corrected for minimal jitter. "No garbage in, no garbage out"
Peachtree claims 3 ps correction for jitter, but this really means they claim 3 ps correction at their clock. There is no published data on what correction actually is for the system.
By way of the kind suggestions of several members here, I have become aware of the I2S FIFO kit, with measured 1.5 ps correction. This appears to be the solution I have been looking for. I have looked at some of the information, it is lengthy.
My concern is what would I have to do to implement FIFO into the Dacit via a USB cable input from the pc? The Dacit has no I2S input, and there is no schematic for the Dacit?
Jim
Jitter's effect on the analog output (the only thing that counts) strictly depends on the jitter of the clock to the DAC chip. Period (bad pun). Get that right and very little else matters as far as errors from jitter go. All the stuff before that makes nice audiophile tales, but doesn't have much, if any, effect on the actual DAC performance.
n0vtz, I think you've hit a typical audiophile issue here.
approach #1: just trust the numbers and relax while listening to the music
approach #2: don't trust the numbers, listen to the music but don't relax and search for a better DAC/reclocker/cable
in my view, as long as the designers of your DAC aren't lying and they know how to do their job, 3ps of jitter is more than enough.
also, I think you noticed already that the guys over at CA don't generally provide an "objective" solution but will rather suggest that you spend some more hard-earned
approach #1: just trust the numbers and relax while listening to the music
approach #2: don't trust the numbers, listen to the music but don't relax and search for a better DAC/reclocker/cable
in my view, as long as the designers of your DAC aren't lying and they know how to do their job, 3ps of jitter is more than enough.
also, I think you noticed already that the guys over at CA don't generally provide an "objective" solution but will rather suggest that you spend some more hard-earned
LOL, 4real you have just justified your well chosen handle.
I've only been at this audio stuff for a couple of years, and am "ok" technically, but not a good or great technician. Despite that, I like to think that I have not been captured by any "audiophile tales", as Sy puts it. I suspect I have been in this case.
After I posted here, a good friend, and exceptional technician came by. I made my case-registry's/PLL's, etc.. He listened to some music, then said my position was a waste of time.
Add to that the opinions I've received here, and it is apparent that I should forget this, and find another project.
A guys got to have something to mess with...
Thankyou all for your input.
Jim
n0vtz, I think you've hit a typical audiophile issue here.
approach #1: just trust the numbers and relax while listening to the music
approach #2: don't trust the numbers, listen to the music but don't relax and search for a better DAC/reclocker/cable
in my view, as long as the designers of your DAC aren't lying and they know how to do their job, 3ps of jitter is more than enough.
also, I think you noticed already that the guys over at CA don't generally provide an "objective" solution but will rather suggest that you spend some more hard-earned
Have really appreciated your input. Hope to see you around again, as well as the others here.
Jim
because the dac is not specified at 3ps, the clock they have used is a 3ps clock, the 3ps is not even measured its just a spec pulled from a datasheet. that is like using a dac chip and claiming (for example) that you dac is capable of -120dB+N simply because the ES9018 is capable of -120dB THD+N. meanwhile you may have in fact used a tube IV, or resistor, which raised it above -100dB
so no they arent lying, they are simply giving correct information that happens to be not terribly relevant/complete
it tells us nothing at all about the actual jitter of the digital input and clock in application.
also it seems you have missed the point of the fifo, it has i2s output, it is designed to connect directly to the dac and that 1.5ps jitter is at the i2s output of the clock buffer, not the clock in isolation. the ES9023 is a pretty DIY friendly chip, its soic, the XO and i2s inputs will be fairly easy to solder to, in fact it would be fairly trivial to add a u.fl connector right near their line termination.
the 3ps clock would be disabled and preferably if there is something like an inductor, or anything in the trace on the way to the dac it should be removed so that the output of the clock is not in parallel to the new input clock. Even with no power often there are diode clamps or buffer circuitry that unless on and actually disabled, will stay connected to the circuit and effect the clock wave shape; adding jitter.
if hes looking to tweak this area, in my experience the fifo was one of the more profound improvements I made to my dac, which is already claimed to be jitter immune....
increasing the analogue reference voltage for the es9023 to 3.6v will yield an improvement as well
the i2s fifo is about as far from a fairy dust offering as you will get in the digital line level area, it is very well engineered, extensively developed and well characterised.
but it really depends on the priorities, he may well be better off spending some effort on his speakers or poweramp
Last edited:
qusp, you're right, the peak-to-peak jitter generally only speaks about the xtal but not actual jitter as seen by DAC chip.
what you wrote reminds me of one problem with hifi nowadays. more and more small companies emerge because manufacturing costs are so low. I wonder what it used to take to make PCBs for DIY or semi-DIY quantities back in the 80s? I cared about other things then, such as my toys (I have other more expensive ones now LOL), so I have no idea. but I'd guess it was out of the question for DIY.
nowadays even for DIY a hot air rework station or soldering oven are accessible so companies emerge overnight. I many times wonder how many of them really have the expertise and equipment to make good audio products. some of them apparently rely on dogma (boutique capacitors and the like) but also seem to lack a solid engineering background. obviously, not all.
not to mention - and this is far from a detail - quality control. supplier changes specs without notice (I assume everyone read the disclaimers in the datasheets) and everything is affected.
I have bought a product in the past that suffered from a problem of the type. it does happen. manufacturer said that their product is superior partially because of the use of a linear supply with oversized filtering, custom discrete regulators and the other typical marketing talk. turned out that the product produced such an amount of noise that it was audible from the listening position even when the PC was on. surprisingly, the product it replaced, which was SMPS-based (a wall-wart SMPS, actually) produced absolutely NO audible noise, ear against tweeter. see? much philosophy, little engineering.
if measurements alone mattered, I'd try a product from a well-known manufacturer which is confirmed by third-party measurements. such a product is the Musical Fidelity V-DAC II. measurements are text-book. very cheap and good resell value if it proves not to be sonically acceptable.
what you wrote reminds me of one problem with hifi nowadays. more and more small companies emerge because manufacturing costs are so low. I wonder what it used to take to make PCBs for DIY or semi-DIY quantities back in the 80s? I cared about other things then, such as my toys (I have other more expensive ones now LOL), so I have no idea. but I'd guess it was out of the question for DIY.
nowadays even for DIY a hot air rework station or soldering oven are accessible so companies emerge overnight. I many times wonder how many of them really have the expertise and equipment to make good audio products. some of them apparently rely on dogma (boutique capacitors and the like) but also seem to lack a solid engineering background. obviously, not all.
not to mention - and this is far from a detail - quality control. supplier changes specs without notice (I assume everyone read the disclaimers in the datasheets) and everything is affected.
I have bought a product in the past that suffered from a problem of the type. it does happen. manufacturer said that their product is superior partially because of the use of a linear supply with oversized filtering, custom discrete regulators and the other typical marketing talk. turned out that the product produced such an amount of noise that it was audible from the listening position even when the PC was on. surprisingly, the product it replaced, which was SMPS-based (a wall-wart SMPS, actually) produced absolutely NO audible noise, ear against tweeter. see? much philosophy, little engineering.
if measurements alone mattered, I'd try a product from a well-known manufacturer which is confirmed by third-party measurements. such a product is the Musical Fidelity V-DAC II. measurements are text-book. very cheap and good resell value if it proves not to be sonically acceptable.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.