I'm trying to decide on the best cabling standard to use to carry an LVDS signal over ~1m. The signal in question is a square pulse, however it is not digital, as the pulse width carries the actual measurment. Hence, I need high bandwidth (>600MHz) as the shape of the rising/falling edges are extremely important, along with minimal skew and jitter. Also, the capacitance needs to be minimised.
I've looked at a number of solutions CAT5/6/7, CX4 (for which I can't find any electrical specifications) and I'm aware it's possible to used twisted pair with LEMO connectors (again I can't find any information on these, beyond the socket spec).
So in short any help/advice on the best solution, and any cabling standards I've missed would be aprreciated.
Use a cable that has a characteristic impedance that matches the driver and also is terminated with that impedance. You will then have maximum signal integrity with least ringing or reflections. Basic AC transmission line theory.
But "since you are here," what is the "characteristic impedance" of Cat?
Back when I did "telephone word" the first order approximate for the "characteristic impedance" was a 600 ohm resistor in series of a cap. I don't remember the size of the cap.
Except that the ideal driver impedance is zero. As long as the termination is perfect the driver "mismatch" doesn't matter. "Close enough" works here.
Likely close to 100ohms, though I've never measured it.
That's a hold over (hangover?) from open-wire lines of 70-100 years back. There is no way twisted pair has a characteristic impedance that high. OTOH, the characteristic impedance doesn't matter all that much at audio frequencies unless your line runs across the country (they don't anymore). The cap is there to block the line voltage. Its size shouldn't matter much.
I'm not so sure that's necesarily a good idea. If you take your zero Z driver and put a resistance in series you haven't really lost anything that can't be compensated for by increasing the driver power. But you gain some short circuit protection and if the ternmination at the far end is less than perfect you will soak up a lot of the echo at the sending end rather that reflecting it back toward the receiver with zero attentuation.
I suppose it doesn't matter much for POTS but there are still plenty of metalic lines that run for MILES and MILES. I suppose that at the higher audio freqs (3.5 kHz) you might not be able to completely ignore the transmission characteristics.
Increasing driver power isn't normally a good thing. The object is exactly the opposite, which is why series termination is popular.
Short circuit protection is normally done in different ways. ECL used a collector resistor to put the output device into saturation. Since a shorted condition isn't in the normal operating area all that's needed is to keep the smoke inside.
As far as reflected energy goes, make the termination better. It doesn't have to be perfect.
Transmission line characteristics come into play when the rise time is on the order of twice the electrical length. Most POTS lines are in the low single digits of miles, usually less. Where do you find POTS lines that run "miles and miles"?
Into the 1990s, I could still see a couple of old open wire on glass insulator phone lines along New York State Route 8 east of Wells and Route 10 south of Piseco, inside the Adirondack Park. These were party lines running to camps several miles past the end of electric service. The longest might have been 15 miles from the local exchange building. By 2000, the local independent telco had replaced these with modern cable.
Yes, it's possible that some of that is still around in the boonies somewhere. However, even 15mi of open wire is too short at 4kHz (round trip something like 150uS). Most POTS lines today are less than 2mi. Most of the system is digital, today.
So a farm house 15 miles away from any 'city' doesn't have copper all the way back to the central office? Is there something along the way between farm house and CO that breaks this up?
Back some 32 years ago when I was part way "in the business" the use of "concentrators" was increasing. These had a digital link to the CO and, in some cases, did a little switching. With fiber being so cheap, I guess (I have no way of knowing) that there is some kind of "box" between my home and the CO which digitizes my line. Since my dial up modem speeds have NEVER approached the 56k limit, I suspect "they" have used a "bit saving" encoding scheme other than the T1 standard of 64kb/sec "u-law". We don't have ISDL or any of that "D" stuff where I live. It's some 9 road miles to the CO and there are folks who live a few miles further down. So far as I can tell, most of the telephone cables are along the road so there are few as the crow flies short cuts. In the next county over, the CO is definitely over 10 miles to the most distant customer.
But, again, I have no good way of finding out. The only telephone folks I run into are the guys who come out when lightning or a power cable failure fries my service (or once, when a "bridge" connection got noisy and they fixed the problem by disconnecting the offending pair.)
Since some of you seem to be 'current,' maybe you can tell be: 1) what's in a "shiny" round can that has 2 underground cables run into it? 2) Are those "loops" that seem to take up slack in cables an indicator of a fiber or at least a high speed digital link?
In there are concentrators or switches in the "outside plant," WTF do they look like? Are they powered like the old T1 repeaters (I think they used up to 135 volts DC.)
Just curious. I like to know how things work.
BTW: it's just over 2 miles from a major digital line that AT&T put in place back when AT&T wasn't owned by SouthWest Bell. The construction of that digital network was one of the reasons the old AT&T was taken over by a "Baby Bell."
Almost certainly there are T1 repeaters inside the cans. Every 6000 feet.
Virtually all systems today work over old copper cable, but the actual "line card" is in one sort or another of "remote" unit that will be within
4 miles of the customer. From the "remote" to the CO might be almost any distantce. (It is literally, width digital, possible to do the switching at a location hundreds or even thousands of miles distant. In fact, where I live a local cell phone call will actually be switched at a location that is 700+ miles distant, and the connection is via satellite so the actual circuit miles will be nearly 100,000 miles!)
It might be anything from a very small box to one large enough to have several racks of equipment.
I don't think that is true. The reason AT&T failed was an entrenched upper management that could not conceive of the PSTN being anything other than message traffic (which is to say, circuit switched voice calls).
Originally that was where the profit came from, and most who grew up in that environment could not change. But technology did change, and by roughly 1995 is was clear that very shortly more than half the revenue would be from "data" circuits rather than voice traffic. By 2000 it was well over half. Today there is virtually no money at all in switched message traffic as such.
AT&T failed because the Board of Directors could not find a CEO who could replace the entrenched "message traffic" mind set with management that understood an all digital world. They tried several, and eventually gave up and sold it.
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.