Converting RS422 to RS232

[mhead]

I have been using the 422/485 four port cards for a year to talk RS232 without a problem. I tie the low sides together with a 10 ohm resistor to ground. I have a brand new card, that this doesn't work for.

[JoshuaP]

Hey mhead,

The RS-232 bus uses approximately -12 volts for High Logic, and +10 volts for Low Logic.

The RS-485 bus is a differential bus. It uses an approximate voltage difference of +3 volts when loaded, and +5 volts when it is not loaded (no bias resistors or instruments). The TX+ line is at +3 volts and TX- is at +1 volts for High Logic. The signals are inverted for Low Logic, so TX+ is at +1 volts and TX- is at +3 volts.

The problem with this is that you are only referencing the ground of the RS485. So you are forcing the TX+ to 2V for High and TX+ to -2 for Low. It may have worked in the past with other boards, because the differential for RS485 needs to be a minimum of 500mV, but it can be as high as 5V in some cases. It all depends on the UART
that was being used. The 5V difference would be enough to trigger the RS232, but 500mV wouldn�t do anything. In any case you are using the board incorrectly and I would suggest just purchasing a RS232 board if you want to talk RS232. Besides they are cheaper.

-Josh

[Chris_Rake]

To expand a little RS-232. The spec'ed min input voltage for RS-232 is +/- 2V. Many transceivers actually have a much smaller window (e.g. GND and +1.5V), which is why RS-485 transceivers can sometimes talk to RS-232 devices.

I agree though - using the correct transceivers is the way to go!

[mhead]

Thanks, Joshua. But the problem is that I have to use 485 cards. I have two 485 channels and two 232 channels required. No, I don't have additional slots in the computer. The devices (as most 232 stuff today) just use 5 volt single ended. We have run more than 10,000 devices across our test stands that work correctly. Only the newly purchased NI card does not work.

[Chris_Rake]

Did you switch over to the 843x cards and start having this problem? These newer cards use different transceivers than the legacy cards.

[mhead]

I am using the PCI 485/4 cards 777641-04. The problem may be the isolation resistors I am using to tie the low sides to ground. I have found a range of resistance from 500 to 10 ohms works across our product lines. Two of our stands work fine with the 10 ohms--these are the experimental stands. The one with the trouble had 10 ohms as well. I am upping it to 100 ohms and watching to see if the problem goes away. Methodology sucks, but it is apparently a statistical anomaly.