Serial Tap not recording data?

I'm trying to get this configuration working:

  winA <-> 232/485 <-> tap <-> 485/232 <-> winB

winA and winB are Windows computers with their serial ports connected to a pair of RS232/485 convertors (I have two each of B&B Model 485SD9R and 485LP9N and tried tried every combination).
The SerialTap is connected to IO Ninja (also running on winA).

I've been using PuTTY (because it is trivially easy to set it up to test a serial connection) and I can send characters from winB and see them appear on winA's screen. Strangely, IO Ninja is only showing CTS changing on and off. No characters at all.

Nothing is going the opposite direction and I think I need something a bit more sophisticated than PuTTY for force the appropriate signals high on the convertors, but that's not why I'm here. I don't understand why IO Ninja w/Serial Tap isn't showing the characters that are being displayed on the receiving system. Oh, FWIW, they are the same characters I was entering on winB, so they aren't random noise or anything.

Any ideas?

Oh, I'm still using IO Ninja v3.15.1a

Just to be sure, I upgraded to v5.3.2 (amd64) of IO Ninja and tried again. I sent one character in each direction, one at +00:20.xxx in the log and the return at +00:27.xxx. The first one was displayed (a lower case 'z'), the second was not (a lower case 'a').

My only real question here is why don't I see the 'z' in the log or any characters at all in the information pane?

If this is expected, (and it very well may be) I just need to learn why.

For reference, here's a photo of the Serial Tap so you can see what's lit up (the connection is idle -- no traffic at all):

You are using DB9 connectors on the sides of your Serial Tap (which only works for RS232); at the same time, the white mode switch in the middle of your tap is set to RS485. Change the mode switch to RS232 -- that is, if your USB-to-Serial adapters are indeed RS232. For RS485, you must connect lines to the terminal block at the top of your Serial Tap.

Sorry, I should have remembered that from the last time I used it. Thanks for being so patient and helpful!

Sigh. I'm feeling foolishly stupid this morning. I have searched for a couple hours for examples of how to connect a Serial Tap to a two wire RS-485, but found nothing. If it exists, I can't guess the right keywords. So, I'm going to ask here...

Using the 485 header, I've managed to capture the data bytes being transmitted, but in between I see 40 or 50 FRAMING ERRORs. The test programs are opening the ports at 8,N,1/38400 baud and I've set IO Ninja to use the same settings.

The 485 pinout diagram for two wire the RS-232 to RS-485 converters I'm using show the following:

Pin 8 = TD(A)-
Pin 3 = TD(B)+
Pin 2 = RD(A)-  [tied to TD(A)-]
Pin 7 = RD(B)+  [tied to TD(B)+]
Pin 4 = GND

I am connecting Pin 2 to TX- on the Serial Tap and Pin 3 to RX+. I've tried with and without the Pin 4 connected to GND and seen no change in behavior.

So, in the absence of any documentation that is clear enough for a moron like me, I need to know if I have connected things correctly and if not, what is the correct way to connect a two wire RS-485 to the Serial Tap?

[EDIT] It just occurred to me to try connecting Pins 2&3 backwards from the labels, just to see what happens: It works much better! I still see a framing error between almost every message, and BREAKs scattered throughout, but the pages of framing errors are at least gone. Adding the GND connection appeared to make the capture quality lower, so I'm just using Pins 2&3. Am I doing this right or is there a better/more correct way? Thanks!

[EDIT #2] I just stopped and started both the program sending data and the IO Ninja. Now I'm getting continuous FRAMING ERROR messages ... even when I stop the data generator and close its serial port. When I also stopped the receiver and closed its serial port, IO Ninja started continuously recording zero bytes and BREAKs with a strange sort of heartbeat pattern. I will try to attach a short session showing the strange pattern of null characters and BREAKs. Nope. Can't figure out how to attach a file, but you can get it from IO Ninja Session Null-Break.zip if you want to look at what I see.

@steve-valliere said in Serial Tap not recording data?:

I am connecting Pin 2 to TX- on the Serial Tap and Pin 3 to RX+. I've tried with and without the Pin 4 connected to GND and7 seen no change in behavior.

This is where it goes wrong. For a two-wire RS485, you should use the RX+/RX- pair (or TX+/TX-, doesn't matter. But not TX-/RX+ or TX+/RX-). Connecting GND is not strictly essential, but is recommended. See more here: https://ioninja.com/forum/topic/14/rs485-modbus-rtu-comms-snooping/2

You may not have noticed the pinout I listed, but TX- & RX- are jumpered together in the RS-485 port, as are TX+ and RX+. Therefore, pins 2 & 8 may be considered as Data- and 3 & 7 may be considered Data+. I'm using pins 2 & 3, so I'm connected the way you've suggested, and was even before your suggestion.

HOWEVER...

After performing a "long power cycle" (i.e. leaving EVERYTHING unplugged overnight) the exact same connection is now working perfectly. The manufacturer of the RS-232/485 converters I'm using told me to "just use PuTTY" to test the devices. Apparently the port on one of the machines saw an XOFF and simply stopped transmitting. When I powered on I set both copies of PuTTY to no flow control before starting any testing.

Sorry if I wasted your time. I'd somehow gotten the ports into a bad state and had nothing that would show me that state. But everything is OK now.