Industrial Ethernet connector woes


Thread Starter

Rob Hulsebos

The Profibus user group has announced their choice for RJ45 as the ProfiNet industrial Ethernet connector: "The reason for the use of the RJ45 connector is primarily the positive experience in the industrial environment. Numerous tests, including shock and vibra-tion, have produced very good results in industrial use. "(see This contradicts all sorts of stories by many vendors during the last several years, investigations by Rockwell, etc. about the <unsuitability> of the RJ45 connector, especially under vibration, and the different connectors now being developed by various others. Who am I to believe now? Rob Hulsebos

Alex van Dalen

I agree with you we use the 9-pins sub D connector for the industrial Ethernet. Siemens and Hirschmann also use this in there products. We also have bad experience with the RC45 connector.
Rob, if memory serves, Rockwell discovered that vibration removed the thin layer of gold that the RJ45 pins have. It seems, IMHO, that this may take some time if the connector has a thicker layer of gold. i have seen specs that stipulate the thickness of the gold layer. So a RJ45 connector to spec may last longer ? BTW, the specs are 1) a 8P8C ( ie RJ45 ) connector should have 1.3 um (50 microinches) of gold 2) a 8P8C ( ie RJ45 ) connector contacts should exert a minimum contact force of 100 grams 3) a 8P8C ( ie RJ45 ) connector should have a minimum durability requirement of 700 plug / unplug cycles > > Who am I to believe now? Rob, i agree with you - we need to get to the bottom of this. My concern with RJ45 are corrosion issues. Eg - Always use a spring loaded shutter to keep dirt out. Anyone out there got the Rockwell results on hand ? mike barker Durban, South Africa

Curt Wuollet

Hi All Remember that Bell Telephone probably spent huge amounts of time, money, and effort in developing the RJ connectors in the first place. When you have an installed base in the millions and people expect 100% reliability, you choose your connectors very, very, carefully. And they do perform very well as phone connectors and more recently as serial connectors and Ethernet connectors. The flaws they do have are more than compensated for by the fact that they are completely characterized and tested with demonstrated reliability, They are also inexpensive and ubiquitous. Before I would condemn them in any case, I would make sure that the ones in question are actually telco spec connectors. The $10.00/k taiwanese varieties, of which there are many do have extremely thin gold flash. It is so thin that it is pale with the base metal or underplating showing through. I think the incentive for the automation companies to use proprietary connectors is clear and makes their objectivity somewhat suspect IMHO. I heartily approve the effort of just one group, just one time, trying not to reinvent the wheel and use an accepted de facto standard. I wish the industrial ethernet folks would take heed and do the sensible thing rather than the most lucrative thing. We can indeed see farther if we stand on the shoulders of giants. There is also no point in reinventing TCP badly. Regards cww

Hullsiek, William

In addition to Curt's comments, there is another incentive of using RJ-45. If you run out of spares, then maintenance can quickly go the local computer store and pickup some generic components. Then install the good stuff later when supplies get replenished. - Bill Hullsiek "at least I have read the BICSI and EIA/TIA manuals"

Michael Griffin

I'm not an expert on this subject, but I suspect that good mechanical support and sealing for the connector may make a difference. Many of the people who have used RJ-45 and similar connectors in the past have just used bare unsupported and unprotected connectors. For example, Brad Harrison has an RJLinx (spelling?) connector which appears to be just a housing over a standard RJ-45 style connector. The housing has a ring on the plug which screws down onto the receptacle, similar to a large version of a proxy cable. I haven't tried this product out, but I have been investigating using something for connecting computerised test equipment to our network. I used D-shell connectors in the past (inside a heavy duty Weiland housing). Although they have worked well, our computer department doesn't like them because they don't have any official network rating. ********************** Michael Griffin London, Ont. Canada [email protected] **********************

Curt Wuollet

I have seen silicone rubber boots that would serve in corrosive environments, and a compatible design with an O-ring wouldn't be rocket science either. There are perhaps some environments that require something else, but it makes a lot of sense to use the commodity plug for the 98% of applications that it will handle well and use the expensive non-standard plugs where they are needed rather than make everybody buy $80.00 connectors and kill the market immediately. I know this flies in the face of the existing automation thinking but, I have flown all over jiggling very expensive mil-spec sealed onnectors for the very same reason. Very few connectors have enough contact pressure to be gas tight and those that do wear very quickly. Every engineering decision is a compromise, I think this is a good one. There is nothing preventing someone from marketing alternatives for truly hostile environments. Making them the exception rather than the rule is the most sensible thing I've seen in this market so far. The pressure to adopt Ethernet is for commodity networks, not simply another proprietary to the hilt fieldbus or interconnect. That simply makes no sense as there are plenty and more of those to choose from. Or you can even invent your own:^) Regards cww

Adrian Chesney

If you are not already on our mailing you should perhaps think about it as we are covering industrial ethernet cabling and connectors in the sixth edition of The Industrial Ethernet Book. We will have articles from Lapp Kabel, Leoni cables, Woodhead Connectivity and Lumberg. These should help anser your questions. Regards Adrian Chesney The Industrial Ethernet Book GGH Marketing Communications 1 West St., Titchfield, Hants, PO14 4DH, UK Tel: +44 (0)1329 846166, FAx: +44 (0)1329 512063

Bob Lounsbury

First of all I would like to clear the air regarding our position on the RJ 45 connector and it's use on the factory floor. Our statement regarding suitability did not say that RJ 45 connectors are not suitable for the Industrial Environment. What was said, " is not all RJ 45 connectors will survive the industrial environment". We at Rockwell Automation do recognize and recommend the use of RJ 45 connectors with specific performance attributes. In fact if you review the Industrial EtherNet specifications available on the web site, you will see that the RJ 45 is part of the specification. The issues that need to be addressed are: plating requirements, vibration performance, capacitance un-balance and cross talk. In addition some level of sealing of the connector is required. We all know that the RJ 45 technology presents some challenges in sealing. However there are sealed RJ 45 connectors on the market, that will become part of Industrial Ethernet standards. Our testing has clearly shown that some Off the Shelf RJ45 connectors will not survive high vibration testing. That is not to say that all connectors cannot survive the high vibration common to the factory floor. There is a big difference in an RJ 11 on your $10. telephone than on your $1million machine operating in a environmentally harsh environment. The main difference is the contaminants in the air that will attack the base metals once the gold and nickel has been pierced A failure on your telephone only means static and worse case the phone doesn't work. A failure on a machine means a process stops which can cost 100's of thousands of dollars / hour in profits. Would you risk your company profits on a connector designed for a telephone without some assurance that the connector was designed to industrial standards? Our job is to address the issues surrounding the components, make changes to the designs where necessary and advantageous and educate the users on potential problems with the use of Off The Shelf Components. I agree that there is no place for proprietary solutions in this customer driven "open" market. This does not mean that we can dismiss the noise, environmental issues and use commercially off the shelf components "as is". We have found adding a shell around the RJ 45 greatly enhances the stability of the RJ 45 connectors. Regards Bob Lounsbury Rockwell Automation
Hi Bob

A reasoned approach to be sure. And hopefully not incompatible with bringing the salient advantages of Ethernet to automation. My only point is that the benefits dissappear if the hardware is no longer commoditized and the protocols are not the ubiquitous and standardized Internet suite. These are the cornerstones of the vast acceptance of Ethernet in the general computing world. Lest we forget, there were many equally functional proprietary competitors to Ethernet. It was the adoption by the DARPA community and widespread standardization that made Ethernet the transport of choice. It was still just a head in the crowd until the UNIX world, at that time quite open and mostly academic made TCP/IP the UNIX networking standard. The rest as they say is history until now when the two are now referred to as Ethernet. When you say Ethernet, it is understood that you mean the cheap, ubiquitous, highly effective
networking that combination provides. The automation crowd seems determined to split them up. The fantastic utility of the standardized protocols is utterly destroyed by replacement with proprietary models, often containing the word open. And it seems they are determined to decommoditize even the hardware replacing it with your typical small volume expensive proprietary
black boxes.

My problem is that, if they are successful in doing this, and I see little organized resistance, what part of the goodness of Ethernet
remains? What differentiates this "Ethernet" from the dozens of other proprietary FB products that have failed miserably in comparison to the universal standard Ethernet?

That brings us the the question of: why?
Why instead of taking a standardized universal network and using it to do automation work, must it become another instance of a failed model that has actually severely inhibited the use of networks in automation, another brick in the Tower of Babel? Why can't anyone understand that standardization is the reason there is any demand for Ethernet? And what is being proposed will fall on the floor and rot because it seeks to tell people what they want and is contrary to the whole idea.

Even the mighty Microsoft finally "got it" when many years of trying _and_ even a monopoly failed to convince people that they didn't really want commodity networks and standard protocols.

What is so bad about simply doing what is succesful already?
This is the last bastion of proprietary computing. Only here would anyone even attempt to slide this by. But, unless things change for
the better, towards standardization and open protocols, I predict you are wasting your time. People know what they want when they say
Ethernet. I suggest the "Industrial Ethernet" folks work much harder on trying to deliver what people want and forget trying to decommoditize and privatize the idea for profit.

Please tell me when I stop making sense.


What are yoy saying ??
The Hirschmann Indusrial Ethernet switches, hubs, transceivers all equip RJ45.
Besides I think everybody should consider that the assembly of the plugs and sockets may lead to bad experience.
Every ISP and service provider today is using RJ45,there is big money involved. If the quality would be low it would have been replaced a long time ago !