cRIO & NI 9871 RS485

[vikas.sethi]

Hi,

I am using cRIO and 9871 (RS485 module) for communicating with a device using MODBUS protocol. I have converted the Full Duplex interface of 9871 into a Half Duplex interface using a small custom proto cable (see attached schematic). Also I want to check the communication with the device over a distance of 1000 feet. As I do not have 1000 feet of cable so I am using a resistior to simulate the 1000 feet lenght for the following cable specs

AWG - 22, Cable resistance / 1000 feet = 16 Ohms

When I remove this resistor from the line I am able to talk with the device successfully, but when I insert this simulator in the line even though I get a 2V peak-peak differential signal at the input of 9871 I am unable to establish communication between 9871 & my device.

Please let me know what could be a possible solution to this problem.

Vikas

[Hueter]

Vikas

Since you are still getting a 2V Peak to peak value the signal is not being attenuated linearly, but the resister will make changes in the high frequency latency. I would recommend taking a high frequency oscope capture of the signal with and without the resister, we should be able to see a clear difference that is making the signal harder to read. My guess is we may see longer rise time which maybe skewing the data. Let me know what happens

Hunter 

[vikas.sethi]

Thanks Hunter. I tried the loopback test on the same channel with 16 Ohm resistors in the line and it worked. I also tried full duplex communication between two channels of NI9871 with 16 Ohm resistors in line and that too worked.

Also when I use a RS232-RS485 converter and test communciation between a PC & other device, it works without any problems even with 16 Ohm resistors in line.

But this thing does not work between NI9871 & the other device.

Can you suggest what could be possibly wrong.

Regards

Vikas

[DougFerguson]

I am using a 9871 on a sbRIO board and I will use it to communicate to a temperature controller which uses half-duplex mode. I have not been able to get communication with the instrument and I think it is related to this. Is there a good Knowledge Base on using the 9871 in Half-Duplex mode?

Wait a minute, is half-duplex another way of saying two wire? If so, I have already jumpered TxD+ and RxD+, etc. This is what I have done with all my NI serial devices. I thought that maybe my problem was with the TC instrument requiring <stx> and <etx> characters wrapped around all ASCII command strings. I wonder how this works when its converted to a byte array and sent through the FPGA.

[DougFerguson]

Never mind, I figured it out. The wires weren't even connected to the temp controller.