Hi everyone,
I hope that I will find an answer for a problem that I have with one of my networks.
So here it is:
I have two research sites that are operated by loggerNET on each site I have 6 CR1000s that are connected to the server via MD485 multidrop interface. Both networks are exact duplicates of each other with only one variation that is different location and one of the networks works perfectly fine without single problem while the other one pretty much doesn’t work at all. The network connects all 6 dataloggers where one end is connected to the server via RS-232 port on MD485 and the other end of communication line it ends on last CR1000. Also, both networks are about the same length of 1000ft. For both networks I use special EMI protected 2 twisted pair cable connected via MD485s with the reference ground. Each network runs with the speed of 38.4k and has set up 120ohms termination resistors (it seems that it helps a bit with the communication on the network with which I have problems. However, it doesn’t do any significant change to the other well working network). The only one difference between these two sites is EMI propagation and gradient of ground current flow. I think that there is possible ground loop on one of the grounds. I tried few different scenarios with insolating grounds, and then creating chassis ground, also I tried to bias the communication line as well, and no permanent improvement. I can see temporary improvements, but that is pretty much it. Also, I tried to use optical insulators but it doesn’t solve the problem either. I have some idea where my network my pick up excessive voltage. At this point I think that I tried all possible scenarios that I can think of and this is why I’m looking for a help here.
Thanks in advance for help.
artur
Artur,
I see this post has gone a couple of days with no response. I wish I could say that I had a "oh-duh" solution or comment for you.
However, I did query our design engineer. He made the following comments and suggestions.
---
The main issue for RS-485 communications and subsequently the MD485 product is ground potential differences between sites. The MD485 will only tolerate ~ ± 6V ground potential difference; the reference ground wire is used to bring these potential differences within this range. The standard way to use RS-485 in an environment with ground differences is to optically isolate the RS-485 lines. Or switch to fiber optic instead of RS-485.
It looks like the customer (artur) has done the wiring right, and is probably correct in suspecting a ground problem between the sites. If there is a ground problem, the only solution is isolation (other than switching to radios). He (artur) said he tried optical isolators, with no success. I wonder how they were used and where?
To troubleshoot these networks, it is best to configure the network for the simplest configuration that works, then scale up. He has 6 dataloggers and one server; does it work with a single logger and the server? Perhaps the ground currents are being injected at one site, and it is affecting the whole network.
Also, it should be verified that the MD485s themselves are not damaged; one damaged MD485 on the network will affect the entire network. I mention this as a possibility because it seems like optical isolation would solve a ground potential problem, and it apparently did not.
---
I hope this helps. We would love to hear your comments and final solution.
Sam
HI Sam,
Thank you for your response. Yes I think that this helps a bit since at least it will help me to make up my mind and possibly try fiber optics and maybe completely switch to radios.
However, before I will do any major step I would like to fully explore the topic. So, your design engineer asked how I used optical insulators. I used B&B electronics 485FPP single stage insulator. However, as I mentioned earlier there was not much improvement…
Also, I tested all MD485s and they seem to be okay…
Finally, I think that I have narrowed down the spot where the current might be injected to the network and I tried to insulate those spots, but again without success. However, I think that if I’m picking up the ground on MD485 it self not on the communication line (which at this moment seem to be the most likely when I think about it) then by using optical insulators such as B&B electronics 485fpp might not help at all, can it?
On the other hand I like the idea of using fiber optics instead twisted pair. However, I would need more info on that matter. Can I just get some standard fiber optic cable and connect straight to the MD485 or do I need some sort of converter?
Once again big thanks
artur
Was there a final solution to this problem. We are having similar communication problems with our MD485's