R/E/P > Dan Lavry
Why longer is generally better for an S/PDIF Digital Cable
danlavry:
Ronny wrote on Tue, 11 July 2006 07:28
wild pitch wrote on Sat, 08 July 2006 16:29
danlavry wrote on Sat, 08 July 2006 12:22
Unless one needs to have a given length for a design reason, shorter is better.
How about if that DESIGN REASON was that someone wanted to record from their hifi to the computer, which for convenience's sake they positioned near to where they would normally sit when listening to music- maybe as far as fifteen or twenty feet away.
What would be the safest digital output to utilise (e.g firewire, AES, SPDIF) from the AD converter (positioned next to the hifi) going those twenty feet away to the recording device (computer)?
Or would it just be a safer bet to make sure the computer was nearer to the converter by using something more portable, like a laptop, for example?
AES3 on STP recommended max length is 100 meters, a little over 300 feet. SPdif is 10 meters, a little over 30 feet. AES3id which is closer to SPdif in output voltage specs 1V versus SPdif .5V output and also uses 75ohm like SPdif and coax not STP cable, whereas AES uses 110ohm and output levels between 2 and 7V, AES3id with minimum input level at .32V and output level at 1V, can go up to 1000 meters. AES3 and SPdif minimum input levels are the same at .2V. AES3id is also unbalanced like SPdif and AES3 is balanced. Max current is 64mA AES3, 8mA SPdif and AES3id is 1.6mA.
Firewire supports up to 63 devices, 4.5 meter cable max length before a repeater with 16 connections allowed and 72 meters max with repeaters. In the US that would be about 14 feet max FW recommendations between two devices before a repeater is necessary.
USB does have a max length, typically given as 5 meters, with 5 connections, max total of 25 meters or a little over 75 feet max with repeaters. 15 foot max between two USB devices without repeater.
That said much of it depends on the transmission and reception devices, these are general recommendations. There are some devices using sysconcept SPdif cables that claim going 87 feet without a repeater. RME claims 45 feet on their devices with SPdif I/O, about 1/3rd more length than is typically recommended. I haven't tested sysconcept nor RME's claims as I'm in Dan's camp of keeping cable lengths only as long as the application requires. This may be an old analog habit that is hard to break, but it's never done me wrong by going with the shortest signal path for digital or analog transfers and connections. ITR, I was quite stunned to hear from an EE that shorter was better with SPdif, but after reading the rest of the tripe and bull, I understand that the guy is out for the dough and using power of suggestion and car salesman speak to dupe audiophools into purchasing something way over priced and way under needed.
Good post!
There is a false tendency to view the "robustness" of a transmission line (such as cables) as a function of voltage swing. The fact is: The robustness is a a function of the transmitted POWER.
An ECL line driver (about .8V swing) driving a 50 Ohm line, beats a cmos 5V driver any time.
Say you have an AES 3V signal into 110 Ohms. The power is about 40mW (3^2/110 for half the time when the voltage is 3V, and 0 the rest of the time voltage when the voltage is 0).
Compare it to SPDIF at .4V into 75 Ohms, which is about 1.07mW....
Much of it goes back to "signal power" to "noise power" ratio, the old Shannon information theory study.
Of course I do not discount the other factors such as rise times, cable characteristics, enviormental conditions...
Regards
Dan Lavry
UnderTow:
Pure speculation on my part of course:
audioengr wrote on Mon, 05 June 2006 17:39
I do a demonstration in my suite at CES each year. I start with a stock commercial CD playing in a stock CD transport with stock DAC. Then I make the following changes:
1) rewrite the CD using a low-jitter writer (battery powered) and play this copy instead
IOW, burn a CD with different versions of the song that have decreasing amounts of bad processing compared to the first version played and play that back.
Quote:
2) change the transport to an identical one with digital output mods and Superclock3 installed
IOW, skip to track two on the burned CD which has slightly less bad processing than version 1 and the other CD first played...
Quote:
3) change the transport to a computer-driven USB to S/PDIF converter
IOW, skip to track three on the bruned CD which ...
Quote:
4) change the USB to S/PDIF converter to a USB to I2S converter
IOW, skip to track four ...
Quote:
With each of these 4 changes above, EVERY listener is able to hear an obvious improvement.
I believe you.
Of course it could just be that his audience really is so easily convinced into hearing stuff that isn't there ...
Alistair
lfeagan:
Andy Peters wrote on Thu, 15 June 2006 18:02
AndreasN wrote on Sat, 03 June 2006 08:58
There seems to be a minimum length issue at least with some cabling. Stumbled upon this while building a new computer:
SCSI: .. "The shortest cable must not be less than 30 centimeters (12 inches) between each device. For example, if you have 3 device connected to the host adapter, the minimum lenght of the cable should be 1 meter (3 feet)." http://www.technick.net/public/code/cp_dpage.php?aiocp_dp=gu ide_scsi
IDE: .. "80 wire cables also can't be any shorter than 254mm (ten inches), by the way." http://www.dansdata.com/rcables.htm
Both are ribbon cables with simple unshielded parallel conductor/earth/conductor/earth wiring to avoid crosstalk.
At a previous job, I did a lot of work with SCSI, and I never saw a minimum-cable-length spec. Same with IDE.
-a
Not that I agree with audioengr as Dan et al. are quite correct about the length. However, there is a minimum length on ATA cables. For information on this, I suggest reading the http://www.t13.org specs. The T13 Technical Committee are the people who define ATA. In their minutes someone asked about this. Their response:
"The minimum length was set to prevent problems associated with reflection. There is also a maximum distance between two drives at the end of the cable. This requirement is present to solve impedance and capacitance issues."
http://www.t13.org/docs2002/e02144r0.pdf
The older ATA-3 standard working draft has a lengthy discussion of the issues with reflection they faced in the development of the standard. http://t13.org/project/d2008r7b-ATA-3.pdf
danlavry:
lfeagan wrote on Wed, 19 July 2006 09:32
Andy Peters wrote on Thu, 15 June 2006 18:02
AndreasN wrote on Sat, 03 June 2006 08:58
There seems to be a minimum length issue at least with some cabling. Stumbled upon this while building a new computer:
SCSI: .. "The shortest cable must not be less than 30 centimeters (12 inches) between each device. For example, if you have 3 device connected to the host adapter, the minimum lenght of the cable should be 1 meter (3 feet)." http://www.technick.net/public/code/cp_dpage.php?aiocp_dp=gu ide_scsi
IDE: .. "80 wire cables also can't be any shorter than 254mm (ten inches), by the way." http://www.dansdata.com/rcables.htm
Both are ribbon cables with simple unshielded parallel conductor/earth/conductor/earth wiring to avoid crosstalk.
At a previous job, I did a lot of work with SCSI, and I never saw a minimum-cable-length spec. Same with IDE.
-a
Not that I agree with audioengr as Dan et al. are quite correct about the length. However, there is a minimum length on ATA cables. For information on this, I suggest reading the http://www.t13.org specs. The T13 Technical Committee are the people who define ATA. In their minutes someone asked about this. Their response:
"The minimum length was set to prevent problems associated with reflection. There is also a maximum distance between two drives at the end of the cable. This requirement is present to solve impedance and capacitance issues."
http://www.t13.org/docs2002/e02144r0.pdf
The older ATA-3 standard working draft has a lengthy discussion of the issues with reflection they faced in the development of the standard. http://t13.org/project/d2008r7b-ATA-3.pdf
I really do not have the time to read the 180 page (or so) document. But a quick glance showed me that the discussion there is not about properly terminated cables, nor is it it about a single driver and a single receiver, nor is it about a bunch of high impedance input with the last one terminated.
They start with series terminated at the source, then parallel termination at the end. Then they get into what happens when there is no termination, branching, capacitive loading at the middle....
These guys are showing diagrams where a cable has some 25pF capacitive load at the middle, and no termination at the end (and some capacitance). In other words, they have a mess on their hand, and are trying to "balance" things to get a somewhat recognizable waveform...
Such uncleaned signals are often "clipped" with diodes at the receiver, like the old TTL logic. They are dealing with trying to "overcome" a disorderly mess (relative to a clean line with no branching and proper termination). We are talking about a terminated line with no branching.
Regards
Dan Lavry
lfeagan:
Dan, thanks for reading the message. Though, by no means did I intend for you to take your time to reply. It was only meant to dispel the incorrect belief that ATA did not have a minimum length specification.
I am cognizant of the fact that it is not applicable to the SPDIF discussion going on. I debated between merely PMing the poster vs. replying in the thread. My decision to post publicly was based on my desire to dispel an incorrect statement. Though, perhaps it is good that you replied so that no one else would read it and come to the conclusion that the ATA specs were applicable to the cabling discussion. If you think it is overly distracting from the large issues under consideration, feel free to edit the post or do otherwise as you see fit.
Lance
danlavry wrote on Wed, 19 July 2006 14:28
I really do not have the time to read the 180 page (or so) document. But a quick glance showed me that the discussion there is not about properly terminated cables, nor is it it about a single driver and a single receiver, nor is it about a bunch of high impedance input with the last one terminated.
They start with series terminated at the source, then parallel termination at the end. Then they get into what happens when there is no termination, branching, capacitive loading at the middle....
These guys are showing diagrams where a cable has some 25pF capacitive load at the middle, and no termination at the end (and some capacitance). In other words, they have a mess on their hand, and are trying to "balance" things to get a somewhat recognizable waveform...
Such uncleaned signals are often "clipped" with diodes at the receiver, like the old TTL logic. They are dealing with trying to "overcome" a disorderly mess (relative to a clean line with no branching and proper termination). We are talking about a terminated line with no branching.
Regards
Dan Lavry
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version