On Friday 01 January 2021 12:24:55 John Dammeyer wrote: > > -----Original Message----- > > From: [email protected] [mailto:[email protected]] > > Den Fredag, Januari 01, 2021 06:03 CET, skrev Chris Albertson > > <[email protected]>: �I don't know much about RS485 but I > > did read that the current return path is required. You should be > > able to send and about 100K up to 4000 feet. > > > > RS-485 use differential signals, currents flowing in differention > > directions so return path is in the other conductors. A termination > > is needed, should be easy to find how it should be made ib internet > > once you know it's needed. > > As Dan stated RS485 changes the polarity of the signal so current > flows from Pin A to Pin B for one logic level and from Pin B to Pin A > for the other. Unlike CAN bus, a ground is _not_ needed. > > But there are several rules to follow. One of them is the impedance > of the cable should be in the neighborhood of 100 Ohms. Using 50 Ohm > cable is pushing it. > > The cable, with twisted pair is a transmission line and as such needs > to be terminated across the A,B pins with a resistor that matches the > cable impedance. So a 50 ohm cable needs a 50 Ohm resistor at each > end. That's a parallel combination of 25 Ohms load plus the 2x 12K > receiver input impedance. For now forget the input. > > Differential output from the TI SN75176 data sheet at 54 Ohms is > Typically 2.4V. So 2.4/54=44mA. > > Now use two 50 Ohm resistors and 2.4V/25 Ohms = 96mA which exceeds the > maximum drive current. Assume the low output voltage of 1.7V/25 Ohms > and you get 68 mA which is still 8mA above max. > > If you don't use 50 ohm resistors on 50 Ohm impedance cable then you > get reflections from the edges of the data signals. Again, might not > be a big deal on a short cable or even a long one. But the right > combination of 1's and zero's might well create a travelling wave that > randomly distorts the signal with no repeatability or possibility of > testing. > > I suggest finding a piece of cable that is rated at 100 Ohm or 120 Ohm > impedance (twisted pair) and using that with proper resistors at each > end. > > Say you google "shielded twisted pair cables" and chose the link for > the Beldon 9451. Seems pretty good from the Digikey spec list except > they interestingly don't list the impedance. For that you need to > download the Beldon data sheet and there you find the characteristic > impedance is 45 Ohms. That' pretty typical. And TBT, shielding is not required for such a twisted pair, PARTICULARLY when the shielding over-wrap is calculated into the cable impedance because the added capacitance can easily cut the impedance by at least 50%. Commmon flat ribbon cable is a good example, with every alternate wire a ground, its impedance for the ire between those grounds is in the 130+ ohm range. Put the 9451, which essentially twists and shields the same 2 wire pair, and you get that 45 ohms. In a broadcast facility with audio buses extended from the DA output, looking at the usual 600 ohms load each devices audio input is terminated with, and you've just designed a heck of a good lowpass filter. We like to keep stuff up to 15 kilohertz in our tv audio. But due to the missmatch, 200 feet of 9451 cable is down over 12 db at 15 kilohertz at the other end of that cable. So I designed an audio DA with a high power output swing of the + and - 15 volt supplies, and source terminated it with 47 ohm resistors and had one of my more technical operators build it. Sounded noticeably better, but 22 such cards in a 4 unit high rack sure made a good control room heater. And since the control room was in the middle of the building, I had a hell of a time convincing the AC service techs that I could care less if it was 20F below outside, that air conditioner absolutely HAD to run 100% of the time. I found one tech that knew how to fix it, he ordered special fan motors that could be speed controlled, and found a modulator that turned the outside condenser fans down to about 30 rpm in really cold weather. That kept the condenser coils hot enough that it worked just fine. But he went to greener pastures, and the next guy was a ducking fummy and tore it all out when I wasn't looking. I taught my replacement about those finer details, and retired in 2002. Then came digital and a brand new tech room was built, sharing an outside wall with the building. And that simplified the air conditioning needs.
> Now add 100 Ohm to the search "shielded twisted pair 100 Ohm cables" > First hit is for 9501. You have to look carefully to see it's 75 Ohm. That 75 ohms means its still shielded, you did specify it. This stuff on the amazon usb-485 cable is not, no foil or drain wire in sight so I'd hazard a SWAG at 90+ ohms, maybe even 100+ ohms, not too bad IOW. In other words, twisted is good, but get rid of the shielded requirement in your cabling search. Take a lesson from ma Bell and let the signal balance do its job. > There are other brands. Clark Wire and Cable in Chicago might be worth a call. Coax's for video with 120 db of shielding is one of Susans products. I replaced about 5 miles of west penn's 50 db junk with her stuff during my nearly 19 years at WDTV. And amazingly, for that much shielding, it handles like well cooked spahgetti. Our sat truck spent 3 days at Shanksville after 9/11, with 3k feet of it laying in the mud. Rolled it back up when the "show" was over and I expect its still on that satellite truck right now. Have a better new year everybody. > John Dammeyer > > > > > _______________________________________________ > Emc-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/emc-users Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) If we desire respect for the law, we must first make the law respectable. - Louis D. Brandeis Genes Web page <http://geneslinuxbox.net:6309/gene> _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
