Ideal RJ-45 Connectors ?

[Birken Vogt]

That looks like one my installs
I have a terrible time terminating those dangable connectors,
Obviously I'm not as fastidious as LittleBill (although a lot more handsome)

You know what, whether you color completely inside the lines or the wires are a little mussed, the data still flows.

 

[LarryFine]

You know what, whether you color completely inside the lines or the wires are a little mussed, the data still flows.

You have to maintain the twisting to minimize data slow-down due to poor noise rejection.

 

[Little Bill]

Obviously I'm not as fastidious as LittleBill (although a lot more handsome)

Beauty is in the eye of the beholder. Obviously I haven't beheld you!

 

[tallgirl]

A couple of years ago, a friend asked me to troubleshoot the phone-over-ethernet jacks he added to an existing system at a veterinary clinic. Here are the before and after pics:

View attachment 2572872 View attachment 2572873

Termination porn for the win!

 

[tallgirl]

You have to maintain the twisting to minimize data slow-down due to poor noise rejection.

This is the correct answer. Especially if the Cat 5 is being used for speeds over 100Mb/s, which it isn't supposed to be anyway.

There are some weird things about Ethernet, which is why I'm bothered to respond (hi, everyone! long time, no me!) some equipment may require 568-A on one end and 568-B on the other. It depends on how old it is, but this likely isn't the issue.

Some switches can get very picky about trying to upgrade the connection from 100Base-T to 1000Base-T whether you like it or not. And here's the kicker -- it will do it even if the control equipment end is limited to 100Base-T but the edge device will do 1000Base-T. Straight-up Cat5 doesn't reliably support 1000Base-T.

That said, it's most likely a lack of fastidious terminations.

 

[LarryFine]

(hi, everyone! long time, no me!)

 

[LarryFine]

some equipment may require 568-A on one end and 568-B on the other

Only crossover cables that I'm aware of.

 

[tallgirl]

Only crossover cables that I'm aware of.

Requiring a crossover cable was the norm before auto-sensing became the norm.

Remember that TX and RX don't change on the interface. This means straight-through results in the wrong signal being connected at the edge. MDI-X fixes that, but it fixed it ages ago.

Medium-dependent interface - Wikipedia

en.wikipedia.org

 

[tallgirl]

View attachment 2572900

It's been a pretty wild ride since my last time here. My employer is moving the entire engineering team to a new building and the next 6 weeks should be very chill until we put the new lab together and go insane.

I'm trying to spend some time above the ceiling before we get our conditional COO and if I do, I may post pictures since everything up there is already production. Which will be really sweet since normally I work on pre-production equipment.

 

[Birken Vogt]

You have to maintain the twisting to minimize data slow-down due to poor noise rejection.

One or two twists missing out of thousands in a 50 or 100 foot run is not going to make any difference.

 

[LarryFine]

One or two twists missing out of thousands in a 50 or 100 foot run is not going to make any difference.

Maybe not no difference; depends on the environment. It doesn't take much noise to slow down data.

From two different older threads:


The purpose of twisting is to assure that both conductors will pick up any nearby interference (common mode), so the receiving end (differential amplifier - common-mode noise rejection) can ignore it, and pass only the difference signals. The greater the number of twists per inch, the higher the frequency (shorter wavelength) both conductors will pick up equally.


Twisted-pair works by assuring that both wires pick up the interfering signal equally (common-mode noise), so it can be ignored by the differential amplifier at the receiving end (common-mode noise-rejection).

The difference signal is passed through, while any signal common to both wires is rejected, or filtered out. The greater the number of twists per inch, the higher the frequency of interference you can assure will be ignored.

 

[Birken Vogt]

Yes, but if they are untwisted for a very short distance but still in close proximity to each other, common mode noise rejection will still be very high. Proximity and equal length are far more important than maintaining twist.

 

[LarryFine]

Yes, but if they are untwisted for a very short distance but still in close proximity to each other, common mode noise rejection will still be very high. Proximity and equal length are far more important than maintaining twist.

To get nit-picky, it depends on which way the interfering signal is coming from, broad-side vs edge-ways, and the orientation of that signal, parallel vs perpendicular.

"Nature's rules, Daniel-san, not mine."

 

[GoldDigger]

Yes, but if they are untwisted for a very short distance but still in close proximity to each other, common mode noise rejection will still be very high. Proximity and equal length are far more important than maintaining twist.

At terminations, maintaining constant impedance to avoid reflections is likely to be a much more important factor that noise rejection.
The mechanics of twisting provides a much more consistent separation between wires and hence the correct characteristic impedance than "flying wires."

BTW, in the A versus B diagram, in the B section on the third line the printed colors are right, but the white/green wire is labelled white/orange.

The single error which will totally disable the cable, especially a long one, is terminating unpaired wires where a pair should go. I have seen this cause complete failure with a long wire DIY installation even for analog telephone wires at audio frequencies. The installer used multi-circuit phone cable, and decided that red and green should be the wire pair for a single circuit. Probably as a hold over from older 4 wire telco cabling.

 

[Birken Vogt]

In my experience, one additional twist removed because the installer was having trouble getting things to lay right is not going to mess up connectivity of the cable.

Of course the wires have to be in the right place.

But normal Cat cable does not have a red wire anyway. Did they think orange was red?

 

[Todd0x1]

This is the correct answer. Especially if the Cat 5 is being used for speeds over 100Mb/s, which it isn't supposed to be anyway.

There are some weird things about Ethernet, which is why I'm bothered to respond (hi, everyone! long time, no me!) some equipment may require 568-A on one end and 568-B on the other. It depends on how old it is, but this likely isn't the issue.

Some switches can get very picky about trying to upgrade the connection from 100Base-T to 1000Base-T whether you like it or not. And here's the kicker -- it will do it even if the control equipment end is limited to 100Base-T but the edge device will do 1000Base-T. Straight-up Cat5 doesn't reliably support 1000Base-T.

That said, it's most likely a lack of fastidious terminations.

802.3ab (the 1000 base-t spec) was designed around CAT5. Properly installed CAT5 should be fine. I think there was something about not having more than 4 plug/jack pairs including equipment, but it's been a while. 1000 base t isn't any higher of a frequency than 100, it uses all 4 pairs and different encoding to achieve the speed.

Usually installer error that causes problems, split pair and the wrong RJ45 plugs (where plugs for stranded wire are used on solid wire) are a couple I run into frequently, wrong RJ45 more so. Occasionally a 110 punchdown jack installed on stranded wire.

 

[tallgirl]

802.3ab (the 1000 base-t spec) was designed around CAT5. Properly installed CAT5 should be fine. I think there was something about not having more than 4 plug/jack pairs including equipment, but it's been a while. 1000 base t isn't any higher of a frequency than 100, it uses all 4 pairs and different encoding to achieve the speed.

Usually installer error that causes problems, split pair and the wrong RJ45 plugs (where plugs for stranded wire are used on solid wire) are a couple I run into frequently, wrong RJ45 more so. Occasionally a 110 punchdown jack installed on stranded wire.

Different encoding can make noise immunity even more important since more bits are transmitted per clock. Distortion of the wave form caused by wrong impedance, which can including wrong twisting, can cause data loss.

 

[Todd0x1]

Different encoding can make noise immunity even more important since more bits are transmitted per clock. Distortion of the wave form caused by wrong impedance, which can including wrong twisting, can cause data loss.

All valid points. People often disregard how important maintaining the twist is when terminating to a patch panel or jack.

 

[tallgirl]

This slide deck (See slide 19) will show how utterly insane the signals have become.

PPT - Gigabit Ethernet and 10 Gigabit Ethernet signaling PowerPoint Presentation - ID:3271674

Gigabit Ethernet and 10 Gigabit Ethernet signaling. Eric Lynskey ECE 734: Sept. 29 and Oct. 2, 2003. BS in EE from UNH in 2000 Currently MS student in EE Employed at IOL since Fall 1997 Have been active in FEC, GEC, 10GEC, EFM Interest in physical layer and optical communications.

www.slideserve.com

 

[HarryJames201]

Got a real surprise trying to get some Cat 5 ethernet cables I made with Ideal RJ-45 connectors. They tested fine with my Klein Scout Jr 2. Yet they would not work with some devices and not others. Anybody have had problems like this? Can you recommend a higher quality RJ-45 connector?

Hi i believe You are experiencing compatibility issues with your Ideal RJ-45 connectors despite them testing fine with your Klein Scout Jr 2.This can be frustrating especially when some devices don't work.Consider trying connectors from Leviton or Panduit which are highly regarded for their quality and reliability.