0
\$\begingroup\$

I am interfacing two RS422 devices but failing to do so. I started to monitor my device signals and fell into confusion. Please help me understand a few things. I am attaching snap 01 of my device (LTC2872) data on TX+ line measured with respect to GND.

oscilloscope 1 Snap02 of Standard RS422 to USB Converter data on TX+ measured with respect to ground.

oscilloscope 2

Question 01. From LTC2872 Snap01 I get data as per theoretical concept. With data appearing in Snap02, why are there extra bits appearing here? Data length should be 1.04ms at 9600 8N1.

Question 02. Why are the voltage levels of both devices different? How do I match this level?

P.S: Differential Signal (A-B)

enter image description here

Differential Signal measured A w.r.t B

enter image description here

\$\endgroup\$
6
  • 2
    \$\begingroup\$ Show the schematic of your circuit, especially how you handle the grounds, and any fail-safe resistors. \$\endgroup\$
    – CL.
    Commented May 26, 2018 at 6:57
  • \$\begingroup\$ I have attached Schematics please check. Plus I am using 120ohm resistor for termination. \$\endgroup\$
    – Sidk
    Commented May 26, 2018 at 7:12
  • 2
    \$\begingroup\$ First of all, your 232/485 select pins are not connected to anything, this is not indicated as an option in the datasheet and thus may give unexpected behaviour. One of your TE pins isn't connected, same story. Lastly measuring just one signal on an RS422/485 line wrt ground makes no sense at all. It's a differential system, you need to measure both signals and subtract them. \$\endgroup\$
    – Asmyldof
    Commented May 26, 2018 at 8:10
  • 1
    \$\begingroup\$ I think we would like to see the enable and data signals on the same scope snapshot. They could explain the changes between transmit and receive modes better. \$\endgroup\$
    – KalleMP
    Commented May 26, 2018 at 9:06
  • 1
    \$\begingroup\$ This schematic does not show the receiver, so nobody can tell you why the signals at the receiver behave that way. But it appears you forgot to connect the ground. \$\endgroup\$
    – CL.
    Commented May 26, 2018 at 9:18

3 Answers 3

1
\$\begingroup\$

Wait a second... shouldn't termination resistors go between Y and Z?

\$\endgroup\$
4
  • 3
    \$\begingroup\$ This does not provide an answer to the question. To critique or request clarification from an author, leave a comment below their post. - From Review \$\endgroup\$
    – pipe
    Commented May 26, 2018 at 9:57
  • \$\begingroup\$ Well, I think this IS an answer. It is simply phrased as a hint to the author to check their schematics. And it points exactly where to check. \$\endgroup\$
    – Maple
    Commented May 26, 2018 at 10:17
  • 2
    \$\begingroup\$ Instead of phrasing it as a hint, why don't you try phrasing it as an answer? \$\endgroup\$
    – pipe
    Commented May 26, 2018 at 10:48
  • \$\begingroup\$ This appears to be the correct answer and also forces the OP to verify application connections YZ are outputs and AB are inputs. Bravo \$\endgroup\$ Commented May 26, 2018 at 14:50
1
\$\begingroup\$

When using RS-485, a transmitting device will only drive TX+ and TX- when it is actively transmitting. When it's not transmitting, it will leave the bus floating so as to allow other devices to transmit. At the times the wires don't appear to be cleanly high or low, nothing is driving them. A start condition is defined as a transition from a an active marking state (TX+ high TX- low) to an active spacing state (TX+ low TX- high), and following each byte the line must be left in a marking state for at least one bit type.

A couple of things to note:

  1. RS-422 is generally identical to RS-485 except that there's typically only one transmitter, and thus no need to disable it when not in use. Your examples show that the transmitter is being gated.

  2. This particular transmitter only drives the line in a marking state for one bit time prior to transmission, but some others drive the line in a marking state for longer (e.g. 10 bit times) at the beginning of the transmission. Omitting the delay would reduce the time required to transmit data by nine bit times, but including it will prevent transmissions from getting garbled by noise that immediately precedes them. Which approach is better depends upon how the bus is terminated, and whether the protocol would be immune from phony start bits [e.g. if the transmission starts with a padding byte of 0xFF, but a receiver doesn't care about whether it is received correctly, that padding byte would serve the same purpose as a 10-bit marking prologue, and including both would be redundant and unnecessary redundancy].

\$\endgroup\$
2
  • \$\begingroup\$ Do I need to induce delays to match marking or spacing time? I have tried terminating with 120ohm but things do not work. I am stuck. \$\endgroup\$
    – Sidk
    Commented Jun 7, 2018 at 3:08
  • \$\begingroup\$ @Sidk: Preceding transmissions with a 0xFF byte the receiver won't care about will ensure that there's a nice sequence of marking bits before the real data. On the tail end, one should make sure one doesn't release the line before the stop bit is fully transmitted. When using half-duplex, it's necessary to ensure that one side releases the line before the other side tries to transmit usable data. Other than those issues, data can be sent with whatever timing is convenient. \$\endgroup\$
    – supercat
    Commented Jun 7, 2018 at 14:54
0
\$\begingroup\$

Tx1+=LTC_A1 = A1 , perhaps is gated by TE_1 with 485/_232="1" as a differential voltage from resting at midpoint Vcm to a start bit =1 which may be a tristate mode if not enabled.

while in "snap 1", Tx is binary idle =1 before the start bit if TE_1 = 1 is left on all the time.

added:

The new waveforms satisfy the IC specs on page 4. Keep in mind this just defines the lowest of 7 levels of the OSI model for the physical voltage levels and nothing higher such as protocol or application.

RS232 Receiver
Input Threshold Voltage  0.6min 1.5nom 2.5max V
Input Hysteresis         0.1    0.4    1.0 V
\$\endgroup\$
22
  • \$\begingroup\$ This is not USB; both traces show the RS-422 bus. \$\endgroup\$
    – CL.
    Commented May 26, 2018 at 6:57
  • \$\begingroup\$ That’s what it says for snap 2 \$\endgroup\$ Commented May 26, 2018 at 6:58
  • \$\begingroup\$ Nothing about USB. Its 422 data line. \$\endgroup\$
    – Sidk
    Commented May 26, 2018 at 7:13
  • \$\begingroup\$ Then it must be tristate . Identify your test pins \$\endgroup\$ Commented May 26, 2018 at 7:21
  • \$\begingroup\$ RS-422 is differential signalling. You don't measure it from the ground, you measure between two lines. \$\endgroup\$
    – Maple
    Commented May 26, 2018 at 7:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.