Guys,
I have done it! I got the idea from a web site that I forget the URL too. I built some conversion circuits, toslink and coax to TTL, which I fed into an SN75176. It was fed over twisted pair telephone wire using RJ-11 jacks and connectors to interconnect. At 10ft, the sound is perfect. The receiver circuit has an SN75176, it's received output connects to a circuit with converts TTL to coax, which then is fed into jwb's DAC. With a good tranceiver (I'm getting SP1485ECP which operates up to 20Mbps on Monday) reflection and jitter is minimized, and allows transmission of S/PDIF over thousands of feet with no noise or degradation. At work, we have over a mile of cable run that transmits data at 9600bps. Even though S/PDIF uses much faster transmission rates, I am almost positive it will work well. Why hasn't this method of transmission been used? I think it works pretty damn well.
I have done it! I got the idea from a web site that I forget the URL too. I built some conversion circuits, toslink and coax to TTL, which I fed into an SN75176. It was fed over twisted pair telephone wire using RJ-11 jacks and connectors to interconnect. At 10ft, the sound is perfect. The receiver circuit has an SN75176, it's received output connects to a circuit with converts TTL to coax, which then is fed into jwb's DAC. With a good tranceiver (I'm getting SP1485ECP which operates up to 20Mbps on Monday) reflection and jitter is minimized, and allows transmission of S/PDIF over thousands of feet with no noise or degradation. At work, we have over a mile of cable run that transmits data at 9600bps. Even though S/PDIF uses much faster transmission rates, I am almost positive it will work well. Why hasn't this method of transmission been used? I think it works pretty damn well.
No replies? I'm sure someone would have found some sort of error with it. Anyone have any input? I had the circuit running pretty much all day and jwb's DAC is still sounding good. The transmitting IC gets slightly warm, but still pretty much perfect operation. I'm looking fot your input because I'm a new commer to S/PDIF transmission, and don't really know much about it.
Really? Odd. For the life of me I can't kill one. At work we use them for long distance serial transmission, and they get lightning transients and high voltage spikes, shorts, ESD, and even with atleast 8 years of operation, I never had one die. They seem quite resilliant to me. I can probably go thousands of feet and still get good transmission. I think everyone should use them.
dhaen said:Modern Rs422 tranceivers are slew rate limited, and will not be able to transmit even the fundamenal.
Twisted pair is used for a similar audio standard described in AES / EBU spec's. This runs for "miles" though proper cable, and quite a distance through "improper cable" (RS422)
I once worked at place that needed to move audio around the building in the days when your choice of AES rx was the BBC AESIC or some unobtainable Sony device. Proper AES/EBU spec cable was used and the stuff was hideous. Halfway through the building A4 size eq/boosters had to added. In contrast the 270MHz SDI over coax was a breeze.
Not all RS485 drivers are slew-rate limited. Have a look on the Maxim site for parts with digital pre-emphasis for high bit-rates over fairly long cables.
It is certainly possible to carry SPDIF over CAT5 twisted pair for 100 metres. (As a comparison I am running 2.048M bits/sec over a single CAT5 pair here at work over 500 metres, although I am using a simple first-order C-R equaliser at the receiver).
It is certainly possible to carry SPDIF over CAT5 twisted pair for 100 metres. (As a comparison I am running 2.048M bits/sec over a single CAT5 pair here at work over 500 metres, although I am using a simple first-order C-R equaliser at the receiver).
Schmitt triggers might add to the jitter, but why should it alter the reflections?
If the line is resistively terminated with the appropriate resistor (75, 100 or 110 Ohms), then this will be MUCH lower than the Z-in of the chip, so any change in the chip's Z-in as the input switches will have minimal effect.
If the line is resistively terminated with the appropriate resistor (75, 100 or 110 Ohms), then this will be MUCH lower than the Z-in of the chip, so any change in the chip's Z-in as the input switches will have minimal effect.
It affects the jitter because it affects reflections......
Schmitt triggers use regenerative action to create hysterisis. The energy that is coupled back to the input is effectively energy that is coupled back to the source.
It has the same affect as any other reflection.
Your basic CMOS gate will do th same thing, but at a much lower level. At one time, I made a list of what types and brands reflected the least, and most, energy back. But who knows where it is now..............
Jocko
Schmitt triggers use regenerative action to create hysterisis. The energy that is coupled back to the input is effectively energy that is coupled back to the source.
It has the same affect as any other reflection.
Your basic CMOS gate will do th same thing, but at a much lower level. At one time, I made a list of what types and brands reflected the least, and most, energy back. But who knows where it is now..............
Jocko
Ouroboros said:
I've been thinking for a while now of using CAT5 to send, not SPDIF, but IIS. Since CAT5 has four twisted pairs, you could send bit clock, word clock, data, and still have room for either power or metadata on the last pair of lines. Sending bit clock means you completely avoid the clock recovery problems inherent to SPDIF, and having an explicit bit clock and word clock means the receiver can figure out, without any flags, the bit depth/number of channels for the data. More details if anyone's interested.
Francois.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.