RS-485 Communications with standing waves

Thread Starter

BillDolan

Joined Jun 24, 2014
6
I have an RS-485 network that was installed without adhering to the proper cabling spec. such that it's not a twisted pair with the associated characteristic impedance of 120 ohms. As a result, I'm guessing the SWR is not equal to 1. This may sound like an elementary question, but does the existance of a standing wave on the transmission line result in data sent or received being corrupted?
 

Papabravo

Joined Feb 24, 2006
21,159
I have an RS-485 network that was installed without adhering to the proper cabling spec. such that it's not a twisted pair with the associated characteristic impedance of 120 ohms. As a result, I'm guessing the SWR is not equal to 1. This may sound like an elementary question, but does the existance of a standing wave on the transmission line result in data sent or received being corrupted?
It depends on the waveform, the baudrate, and what is on the other side of the RS-485 transceivers. It also depends on how many transceivers there are in parallel and how long the cable run is.

SWR applies to resonant antennas. It is unlikely that your cable run would be long enough to be resonant at a typical data rate. It might be subject to reflections and reflections at the sampling point of a bit are generally a bad thing.

BTW - what does the cabling look like exactly?
 

Thread Starter

BillDolan

Joined Jun 24, 2014
6
The existing cabling is very short, less than 1 foot. I really should not call it cabling because it consists of two insulated single conductors. Each end has a DB9 connector. Baud rate is 38400. There is only one slave on the network.
 

Papabravo

Joined Feb 24, 2006
21,159
For something that short, at such a relatively low baudrate, you hardly need to worry. In such a case, even with a twisted pair, we would put a single 61 Ohm resistor, across the lines, at one end to simulate two 120 Ohm terminators at each end of a long cable, coming closer and closer together until they finally merged. Using the single terminator has the advantage of preventing the cable from holding a charge due to long strings of 1's or 0's. The reflections if any will die out long before they become a problem and the capacitve load on the transceiver is minimal.

Are the transceivers being driven with UARTs or some other synchronous protocol?
 

Thread Starter

BillDolan

Joined Jun 24, 2014
6
Yes they are UARTs. I have had many comms failures (CRC errors) with this setup to-date that present intermittently. I have 32 installations like this and all present with intermittent CRC errors at times. I have another 32 identical setups that use a Belden RS-485 approved cable that is approx 50 feet in length and have no issues.
 

Papabravo

Joined Feb 24, 2006
21,159
Yes they are UARTs. I have had many comms failures (CRC errors) with this setup to-date that present intermittently. I have 32 installations like this and all present with intermittent CRC errors at times. I have another 32 identical setups that use a Belden RS-485 approved cable that is approx 50 feet in length and have no issues.
Do the 50' Belden cable runs have terminating resistors?
Do the nodes that are 1' apart have separate or common power supplies?
 

Papabravo

Joined Feb 24, 2006
21,159
No terminating resistors in either case.
Power supplies are the same.
So no terminating resistors anywhere. Not too surprising.

I was not asking if the power supplies were the same kind, but rather if they were physically the same supply powering the nodes separated by 1' and two nodes separated by 50', or are separate supplies with a common ground or are they separate supplies with a ground potential difference between them.

Unfortuneately there is so much going on here that we can't see and words fail to convey the nature of the situation. It is really really hard to help you in these conditions because you don't know what to look for or measure and we can't see what you see. Schematic diagrams and pictures are you only recourse.
 

Thread Starter

BillDolan

Joined Jun 24, 2014
6
I do appreciate your weighing in on this to-date. Sorry, when I said the "same" I meant there is one supply and it powers both nodes. I do realize this is a challenge to assess with words only. I'm happy to send a drawing if you still want to continue the discussion. If not, that's fine too. I'm new to this forum and so I'm not sure how to attach a drawing just yet but I will figure it out.
 

Papabravo

Joined Feb 24, 2006
21,159
So we are left with a counter intuitive situation where two nodes using UARTS at 38400 baud and RS-485 transceivers are subjected to a higher error rate than the same setup with 50' of Belden cable. I suppose you could try 1' of Belden cable but I doubt that would provide anything except an anecdotal result. There is still something going on that is not obvious from the information presented so far.

BTW with respect to your earlier question about standing waves you would need a signal in the 480 to 500 MHz. range to produce one on 1' of cable.
 

THE_RB

Joined Feb 11, 2008
5,438
... I have another 32 identical setups that use a Belden RS-485 approved cable that is approx 50 feet in length and have no issues.
Maybe on the longer cable the cable shield is acting like a ground plane and improving the performance of your antenna? Maybe its just a signal strength issue, with better signal strength using the longer cable?
 

vk6zgo

Joined Jul 21, 2012
677
Can you temporarily replace the 1ft link with a 50ft one,terminated like the other 50ft ones?
Or try the Electronics in one of the other setups.

We had a bunch of Transmitters controlled via a RS-485 link,& had a lot of "no-starts' & "wrong frequencies"---not good!

We were sure it was a RS-485 problem,& tried all sorts of stuff to no avail.

One trick we learnt from "BB Electronics" was the use of a bias network to stop the conductors taking up random voltages.(Can't quite remember how it was done--sorry!)
They had a lot of good stuff on their website,but seem to have taken it down.

In the end,we found the problem was faulty connectors where the RS-485 signal entered the PCB with the MAX485 & PIC on it!

Nothing to do with the external cabling,at all!:eek:
 

PRFGADGET

Joined Aug 8, 2011
53
Just for the sake of trying something cheap and simple.
You might try snapping a far-rite around the line.
I had a similar situation several years ago on a short instrument run (lots of data corruption) , a $3 "Radio-Shaft" far-rite cured the problem.
Best I can figure is that there was an external signal causing interference.
May help,,may not ,,but if it works it's a cheap fix.
 

KL7AJ

Joined Nov 4, 2008
2,229
I have an RS-485 network that was installed without adhering to the proper cabling spec. such that it's not a twisted pair with the associated characteristic impedance of 120 ohms. As a result, I'm guessing the SWR is not equal to 1. This may sound like an elementary question, but does the existance of a standing wave on the transmission line result in data sent or received being corrupted?
It CAN....what happens is you start getting dispersion...the square waves aren't as square as they should be. It depends on many factors, such as total length, data rate, etc.

Eric
 
Top