RS-485 biasing

[MartinP]

Fellow LabVIEWers
I have some problems related to a series of units from a specific vendor,
and they suggest I have to do something with the biasing Now, the issue itself
seems easy enough, as it involves just adding
a couple of resistors, in addition to the terminal resistor at the last unit in the line.

My question is this; can it be correct that a problem related to not having the bias resistors present,
will cause the RS-485 comunication to fail, i.e. that I get a time-out from my units ? I get a vague impresion that
problems caused by not having bias handled should give me garbage data instead of no data at all.
But as you understand, this is unknown territory to me. Any coments and knowledge about
biasing in RS-485 is appreciated.

Martin

[K C]

Hi Martin,

Yes, you really should use terminators. Or even better a active terminator at one side (begin) and a passive at the other side (end)

But, if you have a testsetup with just a short line it could also work with it. If you can look at the data with an oscilloscope you can see if the other side responds or not and at what levels the lines are operating.

Beside the terminators you also should connect the 0V line (not ground, earth or shielding) especially with isolated interfaces.

Another strange thing that I came across was the naming of the signals. A or B, + or -, I spoke to a vendor of a device that was not responding (doing all the above) and he asked me how I connected the two devices. I said just A to A and B to B. That was wrong he told me that I should connect A to B and B to A. My though was this man is crazy but I tried because it was just the naming of that vendor. It worked. Now I think the whole vendor is crazy.

And there is also the protocol you are using that could fail. Are you using a protocol ?

Kees