An interesting article, video over CAT5

ARCNET wasn't even created yet, much less Ethernet over twisted pair.

On a sunny day (Tue, 6 Oct 2009 17:15:36 -0700 (PDT)) it happened Tim Williams wrote in :

Well, I was sending some composite PAL over 20 meter *audio* cable, of unknown impedance (should have measured it, but then OK it is already there) so I looked on the web for some compensation schemes, and found that article. I liked their RGB (3 channels) time compensation chip.

62 nS in 2 ns steps adjustable via digital control, with a lot (>100MHz) of bandwidth, can be used for many purposes.

I solved my video over audio cable compensation in spice last night, using discrete. Have to, as I lost some detail in the video with that cable. It is easy, as the amplitude of the color burst is a known at the tx side, you can measure the attenuation for 4.43 MHz. With all that capacitance of audio cable there is only a single pole I think. Then there are reflections.... it is not 75 Ohm cable I think... It is easy to tell between loss of detail because of bad focus or high frequency loss, high frequency loss does not affect vertical resolution, only horizontal resolution, bad focus affects both. Just fun old analog video stuff.

And yes, I have PAL over ethernet here too, with power, for some years,

VLC never worked for me (why? dunno), but mplayer and netcat does the same, more universal.

I didn't say I used it in 1970. Didn't even live in California at that time. I did use it for about half a year around 2001, for the reasons stated in another post. And it worked nicely.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Being a standard or not has no bearing on whether something works. Electrons don't need the incantations of some committee to flow correctly. That's the whole point :-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

bandwidth,

discrete.

frequency loss,

resolution,

And it works quite well. Haven't done much viode but signals of similar bandwidth and dynamic range.

more universal.

VLC works here. Where does it hick up?

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Well, the old UCSB video distribution system used balanced twinax in a custom cable, controlled by custom DTL logic. It was built around

1970. Ran through a 16x256 switching system, wth equaliztion available at both ends. It did great baseband transmission, BW was around 10MHz, and was finally being replaced back in 1994 because construction around the campus destroyed the cable plant. they replace it with fiber.

Charlie

Fiber makes sense. No lightning induced-issues, tons of bandwidth. When I did my first larger design for a fiber bundle a couple years ago it almost felt like the ferry flight of a Lockheed Tri-Star from England to Ireland with a dozen passengers in that huge airplane. Here I was, with my puny 20MHz worth of spectrum, all alone in several GHz of space.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Speed loss vs distance is stunning too, as compared to copper. Nanosecond stuff over a meter of single-mode fiber looks just the same over kilometers.

John

The only thing I don't like is connections. You always have to keep those green "revolving cloth" cleaner boxes handy. Forget that one too many times and the signal is hosed. And you also can't solder them thangs :-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

We deal with 15 racks and 100 fiber connections a day, of which half are likely moved from port to port, and we NEVER have had ANY interconnection interface issues. If a cable goes bad, replace it.

I know, but when doing R&D it's easy to slack off on the diligence in cleaning before mating a connection. Because you can often get away with it for x times. Just not for x+1 times ...

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Every mate/unmmate, or even wiggle, can cost a couple of dB. And lasers mode jump, drift with temperature, overshoot, do all sorts of nasty stuff. You have to work around all that. The enormous bandwidth helps.

John

It's the mode jumps that really got me, combined with the fact that you can't let cuss words slip your lips while at a client. Coming from the analog world that was like wearing ice skates for the first time. Then there was the layout. I had to have fibers on the boards and the minimum bend radius is huge. Plus they have to be secured really well and never wiggle at all. Oh, and no DPAKs and stuff anywhere close. At first I was a bit unhappy there in my 20MHz chunk of bandwidth but over time I got used to all that.

The other thing is that with high power RF your arm gets warm so you know something's amiss. With fiberoptics and IR you can really screw up your eye sight and I was told that when you feel the pain it may be too late. Next week I'll have to work around stuff running at around 4500 PSI, a whole new ballgame. If I quit posting out of the blew maybe something flew off. Wish me luck.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

EIA/TIA-568 defines CAT5, CAT5E, CAT6, and now CAT7. Read it and weep.

existing technologies include baluns and putting computers at either end

USD23 - comparable to the baluns.

Indeed. All you need is to go from 75 Ohm to differential 100 Ohm and back.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
                     "If it doesn\'t fit, use a bigger hammer!"
--------------------------------------------------------------

On a sunny day (Sat, 10 Oct 2009 13:11:26 GMT) it happened snipped-for-privacy@puntnl.niks (Nico Coesel) wrote in :

Maybe you did not notice, but that article I referred to is about SXGA over CAT5. R, G, B. It does indeed do the 75 Ohm to 100 Ohm differential, 3 times, but it also provides a way to correct delays between the R, G, B channels in 2ns steps.

Also the EL9111 strips of the sync and provides H and V sync, see for example:

So those all knowing yawners have probably not even read it, let alone understood the issues.

There is also the issue of output voltage swing and slew rate. But I guess all is simple for somebody who thinks in 100V audio systems.

I know. And I don't understand why he can't grasp that there was a 568A in the early 90's. Oh well. However, it wouldn't matter one bit to me whether there was such standard or not. If it works it works.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

CAT5.

provides a way

understood

Sync separators have been invented a very long time ago and they don't cost even close to that. In fact, one of the first ones I ever used in the mid-80's came from your country. Or maybe Philips Hamburg, don't remember exactly.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.