RS 485 diagnostic

[Yevgeny]

At the moment I'm solving such a task! We need some diagnostic equipment for our RS 485 line. The system includes an A2SHCPU-S1 with communication module A1SJ71UC24-R4 and 17 Unidrives... Everything works fine, but sometimes connection breaks, and I have to define what is the reason of such malfunction... Can you give me some advice how to do that?

[Duffanator]

How many unidrives do you have daisy chained together? I've had a problem with 485 when too many things were connected in the line and the signal was getting dragged down. Also, are you using a terminating ressistor? I'm not sure but I think you have to use different value ressistors depending on how long the run is or how many slaves are in the line. Something to look into!

[funfrok]

I agree about terminating resistors. But i don't think its value depends on number of slaves, it depends on the cable. If not used, adding pull-up, pull-down resistors can help reduce effects of noise. If you have an oscilloscope you should examine line for noise. There can be burst of noise because of some activity that takes place not very often. Result of short googling: http://www.maxim-ic.com/appnotes.cfm/appnote_number/763/

[Yevgeny]

At the moment 17 Unidrives are chained together. When it was 19 everything worked fine too. Terminating resistor is 130 Om. I'm not sure about what to do with pull-up and pull-down resistors. In the module A1SJ71UC24-R4 manual (IB(NA)-66491-C(9810)ROD ) there is a function block diagram for the C24. If I'm not misstaken it has internal pull-up and pull-down resistors. So should I add my own resistors? Or may be I need adding pull-resistors to one or several Unidrives? And one more question: what twistedpair cable do you use in such cases? I have now thernet 5 category twisted pair.

[funfrok]

In the function diagram, resistors are in the external device block. There isn't resistors installed in the C24 module, which could cause problems. One pull-up, pull-down resistor pair should be used, i think.

[Yevgeny]

that drawing is on page 11, but on the page 9 there is "Function block diagram" with something, that looks like pull-resistors... I'll try to upload it now...

[funfrok]

Interesting. But on the page 11, it clearly states that : "Installation of pull-up and pull-down resistors will prevent reception errors". And resistors are added to external device.

[funfrok]

In fact we have a similar problem. There are two identical displays in a paint shop of a factory. One of the displays stops working sometimes and start to work some time later. The display is located in the center of the communication line. If we change the places of the displays, nothing changes. We attached the display which looks faulty with short cable and it works. When we examined the line with oscilloscope, we see significant noise . But one of the displays work with no problems at all The solution we proposed is to upgrade the software of the displays for better noise handling (no waiting for silence), and install pull-up,pull-down resistors to them (they are connected to the same ac source). I'll let you know if it works.

[Yevgeny]

Just have found picture :) I see you have interesting situation too :) What kind of display do you use? For me I want to try changing twisted pair for new one and adding pull-up and pull-down resistors. The only question is where should I place pull-resistors? On the last Uni in chain? or may be on several Unidrives? And even should I add pull resistors on the PLC side?

[funfrok]

It is a custom built display for displaying some temperature and humidity levels. It is built by a company that has a lot of experience in the field. I'd put them after C24 module. It would be easiest place i guess. It wouldn't make any difference where you put them. They will prevent line from floating when not driven.

[Yevgeny]

Ok! I see! Thank you for help, I'll try all that things and tell you about results! But that problem situation is rather seldom, so getting results will be a subject to think of!