How many DIYers actually make the PCB traces for SPDIF 75ohm from the driver to the output connector (and visa versa) and how many commercial audio designs do they do this...
I cant understand why quantum theory keeps raising it head, if you want to understand digital signal transmission there is plenty of REAL FACTS that you can study, though they generally don't support Audiophile digital beliefs, and follow the laws of physics
So no myths are allowed with digital design...
As to worrying about how exact the line impedances are, it is not really a problem at the speed SPDIF runs at (+/-15-20% isn't going to make that much difference, though the closer the better)., what does cause problems is using drivers with an aggressive current drive and rise times that are to fast. There was some interesting discourse on this a couple of years ago regarding a cheap Chinese interface and the Hi-face, and some RF terminators...
As to getting it spot on, you can only do this by using a full SIV (signal integrity verification software), this though is reserved generally for more critical interfaces such as DDR memory and fast Ethernet...
I cant understand why quantum theory keeps raising it head, if you want to understand digital signal transmission there is plenty of REAL FACTS that you can study, though they generally don't support Audiophile digital beliefs, and follow the laws of physics
So no myths are allowed with digital design...As to worrying about how exact the line impedances are, it is not really a problem at the speed SPDIF runs at (+/-15-20% isn't going to make that much difference, though the closer the better)., what does cause problems is using drivers with an aggressive current drive and rise times that are to fast. There was some interesting discourse on this a couple of years ago regarding a cheap Chinese interface and the Hi-face, and some RF terminators...
As to getting it spot on, you can only do this by using a full SIV (signal integrity verification software), this though is reserved generally for more critical interfaces such as DDR memory and fast Ethernet...
Think about the complete signal chain from the output of the driver chip to the input of the receiver chip.
At the transmitter end:
IC to coupling/matching network - is controlled impedance used here? I've not seen a layout yet that it is done so (Stripline layout). It may exists on some of the USB to I2S boards etc designed by people writing in other threads in this forum.
Coupling/matching network to board connector - dito the controlled layout for this.
Now what connector do they use on the board? Probably a simple 2mm spacing plug and socket. The coax probably has a pigtail shield connected to the ground pin. Definitely not controlled impedance.
Wire from connector to back panel - probably audio coax.
Back panel connector - RCA, look at the wiring. Does a pigtail of shield connect to the ground at the connector? Not 75R. If the shield expands to blend to the shell of the connector, it may be controlled impedance.
At the receiving end:
Input connector - RCA. Not 75R
Wire from connector to board - pigtail braid again? Not 75R
Connector on board - pigtail braid to connector pin? Not 75R
Connector on PWB - again 2mm two pin plastic connector. Not 75R
Traces from connector to matching network - no stripline, no controlled impedance.
Traces from matching network to receiver chip input - again no controlled impedances.
SO, why are people getting so anal about the interconnect cable when NOTHING else in the system is impedance controlled?
At the transmitter end:
IC to coupling/matching network - is controlled impedance used here? I've not seen a layout yet that it is done so (Stripline layout). It may exists on some of the USB to I2S boards etc designed by people writing in other threads in this forum.
Coupling/matching network to board connector - dito the controlled layout for this.
Now what connector do they use on the board? Probably a simple 2mm spacing plug and socket. The coax probably has a pigtail shield connected to the ground pin. Definitely not controlled impedance.
Wire from connector to back panel - probably audio coax.
Back panel connector - RCA, look at the wiring. Does a pigtail of shield connect to the ground at the connector? Not 75R. If the shield expands to blend to the shell of the connector, it may be controlled impedance.
At the receiving end:
Input connector - RCA. Not 75R
Wire from connector to board - pigtail braid again? Not 75R
Connector on board - pigtail braid to connector pin? Not 75R
Connector on PWB - again 2mm two pin plastic connector. Not 75R
Traces from connector to matching network - no stripline, no controlled impedance.
Traces from matching network to receiver chip input - again no controlled impedances.
SO, why are people getting so anal about the interconnect cable when NOTHING else in the system is impedance controlled?
I have checked a few more examples of DC bias in such interconnects.
Both HDMI (2.5V) and Display Port (1.8V) maintain a DC bias between
the differential pair and the shield. So if made to take advantage, there
need be no zero crossing of the dielectric. Maybe this is significant?
If we go way back, XLR with phantom power held this same advantage.
Now, I don't know with any degree of certainty that dielectrics used for
cables suffer non-linearity as they cross zero. I do know that hundreds
of caps I've been asked to measure do severely, but those are caps...
It may not be fair to compare, and I don't think I could measure the
actual capacitance of cables due to the stub impedance, but maybe I
could measure to see if presence of DC changes it any... My LCR only
does Zero, 1, and 1.5VDC bias, but does so up to 1MHz. Should be
sufficient to tell if DC has a measurable effect or not.
Both HDMI (2.5V) and Display Port (1.8V) maintain a DC bias between
the differential pair and the shield. So if made to take advantage, there
need be no zero crossing of the dielectric. Maybe this is significant?
If we go way back, XLR with phantom power held this same advantage.
Now, I don't know with any degree of certainty that dielectrics used for
cables suffer non-linearity as they cross zero. I do know that hundreds
of caps I've been asked to measure do severely, but those are caps...
It may not be fair to compare, and I don't think I could measure the
actual capacitance of cables due to the stub impedance, but maybe I
could measure to see if presence of DC changes it any... My LCR only
does Zero, 1, and 1.5VDC bias, but does so up to 1MHz. Should be
sufficient to tell if DC has a measurable effect or not.
Obviously you and many others don't have a clue...
Gold are a even better solution than silver - for all kinds of audio and video signals - digital or analog.
As you have to get into quantum territories to get the answers you have possibly no way to figure out this.
The old rumblings about conduction / resistance have close to nothing to do with the reasons silver and gold (and other materials) sound better.
However silver and gold do NOT improve the "0" and "1" in digital signals or the wave forms in analog signals given that the metrics used for copper, silver and gold are identical or slightly compensated so the specs are identical.
Traditional measurements will most certainly not reveal any significant differences, but trained ears (with a functioning brain in between) will be able to detect the differences - however slightly different from person to person as the brain functions are slightly different from person to person.
If cable dielectric begaviour around zero crossing was any kind of a real issue, there would be one group that would have noticed and would have made sure it was all over the literature....
RF engineers! Linearity is kind of important to much that we do, in fact it is probably THE number one figure of merit in many applications.
The remote sensing guys would also have noticed for certain, as would the folks designing things like MRI rigs.
The real reason LVDS has a bias voltage is that it has to both source and sink current and is supplied from a single logic supply voltage (3.3 often, but 2.5 or even 1.8V on occasion).
P48 is only sometimes present, there are plenty of dynamic mics out there that do not use it, and all that stuff works fine.
Now cables do sometimes have various little infelicities, some of which may even involve the dielectric, things like the triboelectric effect and friends, but these are well known to anyone ordinarily skilled in small signal analogue, so should be no surprise.
Regards, Dan.
RF engineers! Linearity is kind of important to much that we do, in fact it is probably THE number one figure of merit in many applications.
The remote sensing guys would also have noticed for certain, as would the folks designing things like MRI rigs.
The real reason LVDS has a bias voltage is that it has to both source and sink current and is supplied from a single logic supply voltage (3.3 often, but 2.5 or even 1.8V on occasion).
P48 is only sometimes present, there are plenty of dynamic mics out there that do not use it, and all that stuff works fine.
Now cables do sometimes have various little infelicities, some of which may even involve the dielectric, things like the triboelectric effect and friends, but these are well known to anyone ordinarily skilled in small signal analogue, so should be no surprise.
Regards, Dan.
Why is gold used to plate connectors and circuit boards? It has nothing to do with quantum physics or any mysterious property that can only be heard and not measured.
Gold is used to prevent corrosion which reduces conductivity and creates a metal to metal oxide interface which can act as a semiconductor barrier.
Tin oxidizes. Plain and simple. Tin wash circuit boards have a short shelf life. Gold plated circuit boards have an order of magnitude greater shelf life.
There are no mysterious properties to gold. Just well know ones which are applied by engineers every day.
Gold is used to prevent corrosion which reduces conductivity and creates a metal to metal oxide interface which can act as a semiconductor barrier.
Tin oxidizes. Plain and simple. Tin wash circuit boards have a short shelf life. Gold plated circuit boards have an order of magnitude greater shelf life.
There are no mysterious properties to gold. Just well know ones which are applied by engineers every day.
The trouble with threads like this is that you can never quite tell the satire from the true believers.
I hope some of the 'high end' sites are set up as spoofs, but you can never be sure.
Quantum is a rather good indicator that gibberish or really scary mathematics (occasionally both gibberish and scary mathematics) are immenent.
Mention of Tesla can also serve much the same purpose in some circles (Generally with less risk of calculus and laplace however), which is a pity becase he really was rather good before he lost the plot.
73 Dan.
I hope some of the 'high end' sites are set up as spoofs, but you can never be sure.
Quantum is a rather good indicator that gibberish or really scary mathematics (occasionally both gibberish and scary mathematics) are immenent.
Mention of Tesla can also serve much the same purpose in some circles (Generally with less risk of calculus and laplace however), which is a pity becase he really was rather good before he lost the plot.
73 Dan.
Could you point me to a reference where this was demonstrated (as opposed to merely claimed)? So far, I've only seen claims with as much evidentiary support as alien abductions with anal probing or channeling of the dead, but it's possible that there could be a reference with actual evidence that you're aware of. Many thanks!
I hadn't been to this thread in several days! Sweet!!!
yes, Ray doesnt mind a spot of stand-up, I cant even formulate a reply with a straight face. Hes consistent though ...
digital also checks itself for errors, it is fast enough so if there are errors it can resend
thats not always true and certainly isnt with spdif. there is error correction, but no resend afaik. not even all USB audio will do this. USB itself, for data will.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.