1

I've never implemented RS422 before (crazy right?). Since the impedance of the line is controlled to 120Ω do I need an impedance controlled connector for that? If anyone has implemented RS422 which connector did you use?

I was thinking I'd use a standard 0.1" connector or a molex connector of simmilar size, would this change the impedance of the line to prevent me from getting the full bandwidth of the RS422 line?

If I remember right on industrial equipment people use screw in terminal connectors.

Voltage Spike
  • 75,799
  • 36
  • 80
  • 208
  • 1
    I have never faced any issue using just any connectors for RS-422 with standard rates up to 192k. – Eugene Sh. Aug 02 '18 at 18:01
  • kbaud or kbits? – Voltage Spike Aug 02 '18 at 18:02
  • 1
    kbits/s.. maybe even up to 250kbits/s if I recall correctly. – Eugene Sh. Aug 02 '18 at 18:03
  • 1
    You might find this topic about impedance of 0.100" headers useful, https://electronics.stackexchange.com/q/387223/117785 – Ale..chenski Aug 02 '18 at 18:09
  • 1
    The RS-422 rise time is too large thus the connector spacing is insignificant and will not affect overall signal because there is insufficient bandwidth to detect it. – Tony Stewart EE75 Aug 02 '18 at 18:15
  • @EugeneSh. 192kbits should be good enough – Voltage Spike Aug 02 '18 at 18:23
  • @AliChen yeah, thats the question that got me thinking about this – Voltage Spike Aug 02 '18 at 18:25
  • 3
    expected rise time >10ns at source with a wave velocity of >20 cm/ns will NOT affect impedance even at 20Mbps. – Tony Stewart EE75 Aug 02 '18 at 18:25
  • 1
    @TonyEErocketscientist, "will NOT affect impedance" - I would say for the sake of accuracy, "will NOT affect signal waveform"... – Ale..chenski Aug 02 '18 at 18:42
  • Both are measurable with a TDR and correct since we can interchange time and frequency with constraints just as we can interchange voltage and impedance from controlled current switches. – Tony Stewart EE75 Aug 02 '18 at 18:50
  • @TonyEErocketscientist, for the purpose of engineeringly-reasonable characterization, all TDR test usually limit the signal rise time. USB 2.0 limit is set to 400 ps, USB 3.0 limit is set to 50 ps. This is done exactly to exclude concerns and confusions about higher frequencies that are not there in real signal. – Ale..chenski Aug 02 '18 at 19:43
  • Exactly my point . The impedance within the confined risetime is not affected with any reflections even if mismatched because the time duration of any such mismatch is far too small compared to the channel rise time or BW. Just an equivalent perspective. – Tony Stewart EE75 Aug 02 '18 at 22:05

1 Answers1

1

I've never seen 0.1 or Molex used for RS-422, although I did see screw terminals. I don't think you'd achieve "full bandwidth of the RS422 line", which is 10Mbps over 50 feet, using these connectors.

AFAIK most people use EIA-449/EIA-530 as guidelines for 37, 25 or 9-pin D-SUB.

Maple
  • 11,755
  • 2
  • 20
  • 56
  • 3
    0.1" 2x5 headers are successfully used in all USB 2.0 (480 Mbps) connections to front panels on desktop PC and towers. – Ale..chenski Aug 02 '18 at 19:38
  • Hmm.. true that. Well, maybe it will work. FWIW CAT6 + DB9/25 combination works despite cable impedance mismatch, although I never tried this above 1Mbps or over 25m – Maple Aug 02 '18 at 20:06