framing error between pcmcia and rocketport

[kaufman]

I've seen other posts on similar issues, but they haven't helped me out. Using a windows NT machine with a rockeport with standard serial D-9 connectors, I've been sending commands and reading from an instrument (with labVIEW 5.1). I'm transmitting on one line and receiving on another. Now, I'm trying to move the setup to a laptop running windows XP and a pcmcia output card with 2 D-9 connectors (with labVIEW 7.1). I've had no luck with the instrument, so I've tried to just send messages between the laptop and the old desktop. I've been testing the setup with the serial communication examples. When I send data from the desktop to the laptop I get garbage and a framing error, when I send data from the laptop to the desktop I just get error x4010, a problem with serial break. All of the settings are the same, baud rate:2400 data bits:8 parity:none stop bits:1 flow control:none. Anyone have suggestions?

Also, sending a command via the serial com examples from the desktop can make the instrument perform actions, sending the same command via the laptop results in nothing.

Any advice is appreciated.

[Ebele_Obijiofor]

Hello,

You didn't mention what version of NI-VISA you are using, but I would suggest upgrading to the lates version (3.6) if you haven't already done so. You could also try adding small delays before your reads and writes to see if this helps. Good luck.

Regards,

Message Edited by _Belle on 07-26-2006 11:07 AM

[Ben]

At the risk of sounding like a broken record...

Laptops often float relative to ground.

With no common ground reference between the machines, the hardware can not properly resolve the start and stop bits. NOt being able to determine where in time a byte starts and ends results in a "framing Error".

So check your grounds.

Ben

[kaufman]

Thanks Belle and Ben,

Belle: I've upgraded to VISA 3.6 Also, right now I've just been trying to write or read single commands/data, not both simultaneously. Like sending just the command to open a door on the instrument.

Ben: it seems from the experimenting I've done that the grounding might be the problem. How does one fix a problem like this?

P.S. I'm using RS-422/RS-485 protocal

[Ben]

"P.S. I'm using RS-422/RS-485 protocal "

Now that is a horse of a different color! (Wizard of Oz).

This still indicates an issue with the transport layer (wires, transcievers, etc).

You will have to break out the manf specs for both ends of the"wire" and make sure you are following their requirements.

THings to watch for are;

Are the termination resistors were they are supposed to be?

Is the wire type correct ? (shielded twisted pair)

Is there any source of external noise?

Handshaking control (some widgets use RTS or CTS to control the X-mit/Recv modes)?

Ben

Ben

[kaufman]

Is grounding still a problem on RS-422/RS-485? The cables from the laptop seem to react to being placed near each other or the metal table. For instance, I touch the serial connectors together, and the instrument responds as if a command to open the door was sent. How do you fix the ground signal of the laptop? Or do I just need to shield the cables very well?

Message Edited by kaufman on 08-01-2006 02:26 PM

[Ben]

Those are questions where I would expect the mnfg info to help clarify.

Technically, RS-485 is supposed be implemented as a transmision line when the capacitance of the cable, length and other electricall characteristics are controlled.

I believe (check mfg) that ground is optional in Rs-485.

What concerns me though is the common-mode rejection capabilities of the 485 converter. As a quick check you may want to meter (AC) the voltage at the gnd pin of the 485 converter and the gnd of your widget. Anything more than (hmmmm) a volt could cause problems (check mfg).

THe wire used should also be twisted pair, preferably indvidually shielded twisted pair.

I hope this helps,

Ben