RS-485 Passive Tap

My basic question is, what is a passive tap on an RS-232 line, and what is a derivation cable? It sounds like a passive tap would just be wires connected together, but the only definitions I'm finding online talk about it being 1 way communication on each line, and it sounds like they're talking only about fiber.

So the situation is that I have a Modbus RS-485 network that is limping along. Another engineer and I were discussing it. I think one of their problems is that at every device, they have a wire about 2 feet long tied into the main line going from a JB to the meter. According to this paper written by someone at TI a while back, that's a stub and they need to be pretty short (generally less than 6") to prevent reflections on the line. Check out pages 4 and 7 of this document to see what I'm talking about.

https://www.ti.com/lit/pdf/snla049

My colleague says that's bubkis and that he's seen lots of Modbus RS-485 networks working with stubs that were significantly longer. Up to 20 meters, and he prefers that topology. He showed me this Modbus over serial specification document. It talks about passive taps and derivation lines. That may be what we have, but it doesn't define the terms anywhere. See page 21 for an example.

https://modbus.org/docs/Modbus_over_serial_line_V1_02.pdf

So, help me square these 2 documents please. Can anyone tell me what a passive tap is? Right now the wires are wirenutted together in a junction box.

I apologize for the direct links to the PDF documents. I included them just in case the attachments don't work.

I just realized I had a typo in the first sentence. RS-485. I'm not asking about RS-232.

I think that the term "stub" or "dropline" is a better way to describe a connection to an RS-485 bus that consists of nothing but conductors and terminals. The documents from companies that make RS-485 transceiver chips tend to use the term "stub".

The term "tap" is more proper for things with at least some passive electronics in them (like diodes or transformers), though the Modbus serial line document calls a nothing-but-conductors connection a "passive tap".

I think your colleague is incorrect, and the documents from experts tend to disagree with his preference for very long stubs, or "spoke" networks. While the ISO-8485 specification allows for stubs of up to 15 meters at low data rates, every expert agrees that stub length should be minimized.

At a higher level than those chip-maker documents are stuff from companies like Robust DataComm and R.E. Smith, Jan Axelson, and B&B Electronics (now owned by Advantech).

Back to your system: 24-inch stubs aren't excessive for a typical low-speed RS-485 network, but ones made with wire-nuts might have some intermittent, corroded, or otherwise unreliable connections.

Do you have any ability to capture diagnostic information from the Modbus master or any of the Modbus slaves ?

Do you have access to a digital storage oscillosope, especially one that can interpret serial data waveforms ?

Do you have any serial capture/analysis equipment, like the excellent Stratus Engineering Versa-Tap ?

The serial tapping Guru (Ken Roach) turned me on to Stratus Engineering.

I have 4 ot 5 different taos but it's always a Stratus that I reach for. I've done days and days of DF1 decoding by hand. I had 1 customer that had ASCII on everything. We slowly started changing them to Ethernet/IP. These tools are golden when it comes to serial networking

What baudrate are you operating at ?

How long distance is between cable ends?
Is there termination resistors (120-150ohm) on both farthest ends of cabling nad only two terminating resistors?

Clarification?

Is this system normal serial Modbus or is it Modbus Plus? Modbus Plus is the old proprietary protocol at a high baud rate. It was probably a 485 standard electrically.