Why is there so much noise on the discrete lines of my PXI-7831-R?

[thisisnotadream]

Good day! 

We are attempting to emulate encoders that communicate using SSI at 1MHz.  The traffic between our product under test and the encoder is over an RS-422 electrical bus.  Therefore, we have a MAX491 transceiver that converts the clock lines to TTL and our data lines to RS-422.  The problem is this: every time the TTL clock line pulses low or high, there is a great deal of ringing on the data line...so much so that the chip is interpreting it as a state change, causing bad data to be transmitted.  The addition of a needed second encoder setup to the mix (using more lines on the FPGA card and another MAX chip) is only exasperating the situation more. 

We are unsure why the line would be ringing so badly.  Is there some kind of additional circuitry we need in this situation?

Dan M.

[Coq_Rouge]

This is one thing you can try. Insert a small resistor say 4.7 to 10 ohm in both transitions lines(the + and - lines) at the transmitter side.

Message Edited by t06afre on 02-13-2009 06:56 PM

[Coq_Rouge]

You may perhaps go up to 50 ohm also. I do no know your system, so I can not say exact. Do some testing and find the best working value.

[Ben]

t06's suggestion may help.

What you are seeing is reflection of the transmitted signal due to a impeadance mismatch. At high frequencies data propgates like a wave and is subject reflection of magnitude defined by the ratoi of the input to the output impeadnce (please correct if I got this backwards!).

THat is where t06 suggestion comes into play. By inserting teh resistors he is changing the characteristics of the transmision line.

But if you look back aht the ratio of input to output you will see that you can reduce the magnitude of the reflestion by reducing the input impedance of your sink. So putting a resistance across your input that is the same as the device that is driving the line, the reflection is the same size as the incident wave and cancel.

Just trying to help,

Ben

[Coq_Rouge]

Excellent explanation Ben(kudos) . I often make this very short by saying that the signal will be confused. If it do not meet the same condition as it had then it left   This true for (almost) all phenomenon within physics like light as an example.

Message Edited by t06afre on 02-13-2009 07:33 PM

[thisisnotadream]

We've got the 120-ohm resistors on the RS-422 side of the MAX chips.  The noise is on the TTL side.  Is the theory the same?  Find the impedance of the driver side and put a resistor at the FPGA lines to ground?  Or am I totally missing that.  I'm pretty much more a software guy...

Dan

[Ben]

Thank you t06!

But since you brought it up let me reflect that was one my more mind blowing moments when I learned that the speed inside the other media was the casue but somehow light knew* this without going there.  (I can hear the twilight zone theme song just thinking about it).

"We now return you to your normally scheduled thread."

Ben

* Yes i use anthropomorphic models for all of my physics.

[Ben]

---

thisisnotadream wrote:

We've got the 120-ohm resistors on the RS-422 side of the MAX chips.  The noise is on the TTL side.  Is the theory the same?  Find the impedance of the driver side and put a resistor at the FPGA lines to ground?  Or am I totally missing that.  I'm pretty much more a software guy...

Dan

---

That is the reason the resistors are there! Yes I think you got it.

Ben

[Coq_Rouge]

Just be careful. If you insert a resistor in the transmission line at the RS422 circuit(TTL side) and a equal one to ground at the PXI-7831-R side. The signal voltage at the PXI-7831-R side will be reduced to the half. This may not be so good. I would start with the small serial resistor at the RS422 TTL side. See what this can do first. It is just a hint..

[ritch_by_nature]

That did not work for me, Im having similar issues with my software and a cdaq chassis with NI 9239 and NI9263 where I am generating and acquiring at the same time.

Here's the link to my thread:

https://forums.ni.com/t5/LabVIEW/Gibbs-phenomenon-issue-in-square-wave-from-labview-builtin/m-p/3104...