[FUN] One Terminator Performance

[Stephen Mitschke]

Please provide some information regarding one terminator fieldbusperformance.1) How often are fieldbus systems run with only one terminator?2) When only one terminator is present, are the bus devices/instruments onthetrunk/segment in service (i.e., a network of wired Function BlockApplicationsis running while only one terminator is present)?3) For a one terminator condition, how many problems are experienced withdevices on that trunk/segment? What type of problems are encountered?4) Since many end users have test labs to evaluate FF equipment beforeincorporation into their process control system, is a one terminator testconducted? If so, what are the test conditions?Sincerely,Stewart Thoeni, ABB

[Stephen Mitschke]

Stewart,This is a good question. I'll share the experience of Emerson.1. We find that segments with either one or three terminators are quitecommon. Very frequently it is because the user purchases a component suchas a power conditioner with an integral terminator, but does not switch theterminator into the circuit. Thus the user thinks there are two terminatorswhen in reality there is just one. Another case is where a user extends asegment, and removes the terminator from the "old" end of the segment, butdoes not reinstall it. A final situation is where the user 'makes theirown' terminator out of discrete components to save a bit of money, and acomponents falls out from under a terminal screw.We find three terminators when the user again purchases a component such asa power conditioner with a termiantor, but does not realize it, and addsanother terminator at the host end. Or a user extends a segment in thefield, adds a terminator at the new end, and does not remove the terminatorat the old end.Both of these conditions are easy to detect with fieldbus segment checkouttools. Both of these conditions cause the segment to be out ofspecification.2. This situation depends on the equipment on the segment. At Emerson wefound this problem of either one or three terminators on a segment to becommon enough that we make our PlantWeb architecture operate in thiscondition. We extensively stress test components such as power conditionersand recommend those that allow the segment to function with these conditionsactive. If Emerson devices, DeltaV, and Emerson recommended components areused, the segment will almost always work with either one or threeterminators. With non-Emerson components on the segment, or components notrecommended by Emerson on the segment the segment sometimes but not alwaysworks. There are many components supplied by vendors other than Emerson that meetall the fieldbus specifications, but do not operate, or cause the segment tonot operate in either one or three terminator conditions. Emerson usuallydoes not recommend these components even though these components meet thespecs and operate well with 2 terminators on the segment. This isfrequently a difficult decision to make, but incorrect segment terminationis so common that this has become our practice.3. With only one terminator the signal strength is higher than the specsallow. With three terminators it is lower than the specs allow. Frequentlythe condition is detected by an excessive number of message retries. Thesegment usually appears to run, but devices may drop out, and some devicesmay not work at all. The problems seen depend on how loaded the segment is,segment length, spur length, and a variety of other conditions.4. Emerson always stress tests with one and with three terminators todetermine how the segment operates. Almost 100% of foundation fieldbus communication problems are caused byinstallation problems with power, grounding, isolation between wires, andterminators. Checking these out properly is an almost foolproof guaranteethat your installation will work reliably and well.Regards,Tom WallaceEmerson Process Management952-828-7856

[Stephen Mitschke]

Stewart, At the fieldbus lab at SAIT, we have conducted several test situationsinvolving no terminators, one terminator, two terminators, and even threeterminators per segment. Since a terminator is essentially an AC load, ifone of the two terminators on a segment is removed, the peak-to-peakfieldbus signal voltage will increase and exceed the specification. We havenoticed that with only one terminator on a segment, the communication maystill work but becomes unreliable. Moreover, considering we performed thesetests in a lab facility it is safe to assume that the condition would beworse in a field installation.I have attached a couple of scopemeter captures. One is of a segment withtwo terminators and one is of the same segment with only one terminator.Notice the difference in peak-to-peak signal voltage.Hope this helps, Chris George Southern Alberta Institute of Technology

[Stephen Mitschke]

I second Chris' memo about one terminator. The problem is that the signalmay generally get through, but any device, host, etc, that rejects or clipsa saturated signal, will not recognize it. It is such an easy part of thestandard to comply with that we just make sure everyone is told to "don'ttry this". period Chuck CarterLee College

[Stephen Mitschke]

My opinion is:Yes, this is an easy part of the standard to comply with - at least duringinitial commissioning of the segment. The real world fact is, that therewill be times when you will justifiably need/want to run an active segmenton one terminator, e.g. when you want to extend a segment or for otherreasons. There is no way to make your initial design cover all possiblefuture scenarios. And, it would be prohibitively costly to have to shutdownpart of the operation or to put in new segments all the time because youcan't disconnect a terminator. You should obviously plan for the shortestperiods of "one terminator" operation possible, but they will beunavoidable. Just make sure your vendor's FF products and your designapplications are as robust as necessary to meet whatever "risk management"guidelines you need to adhere to.Thanks,Gary AddisonEquistar Chemicals LP---- Marcos Peluso, Emerson Process Management writes in regards Chris'snote -----Agreed!Nobody should use one or three or more terminators in a segment. Busefficiency goes down due to a lot of retries and some loops may go to a safeposition caused by bad communication. Things get worse in a noisyenvironment or with long cable runs.It is a fact of life that people will remove a terminator inadvertently oradd more terminators than necessary (they may own Recon's stocks). A goodinspection during installation pays itself off. If that happens afterstartup ( it happened in many sites), the Host system should providediagnostics telling plant personnel that things are wrong. A survey with asimple and practical device like the FAT-3 would also help a lot.Marcos

[Stephen Mitschke]

Dear Mr. Thoeni,The terminator has two functions: one is as a shunt and the other is toprevent signal reflection and other distortion.Shunt: Devices communicate by transmitting a current that when shuntedthrough the terminator becomes a voltage picked up by all receivers. Whenone terminator is removed the signal amplitude is doubled. I think mostdevices can handle this. Conversely, adding a third terminator would reducethe signal strength by one third, but per the standard the device must pickup a signal attenuated by a factor of eight (of course including attenuationalong the wire) so this will not really be a problem either.Reflection: When one terminator is removed the network becomes one giantspur subject to signal reflection. As the signal travels down the long trunkand reaches the non-terminated end it will be reflected back in a weakerform, but which interferes with the original signal. This way thecommunication is disturbed. The non-terminated network is "open" and alsomore prone to pick up noise, i.e. the device may work well and good withjust one or two terminators in a test lab but not necessarily in a plant.Jonas Berge, Smar

[Stephen Mitschke]

Ian, I believe we should remind our list subscribers of the effect oftransmission line length. The effect of terminators is dependent on thelength of the H1 segment.Short segments will work very well with a missing or extra terminator givengood design of the transmitting and receiving circuits in Fieldbus devicesand hosts.However, the effect of missing terminators is greatly magnified by phaseshift in longer segments. The maximum segment length recommended in theoriginal H1 wiring guide document is 1900 Meters. This length approaches1/4 wavelength (around 2200 meters) for H1 signals. The reflected impedanceof a 1/4 wavelength unterminated transmission line is a dead short. Long segments that are not terminated at both ends will not work. Devicesthat are close together on such a segment will be able to communicate, butdevices that are on opposite ends will not. You could have some interestingfailure modes in such a case, and I don't generally like to explaininteresting failure modes to operating people.On another note, transmission lines that are properly terminated can runmuch further than 1/4 wavelength. The 1900 meter limit is a number withoutmuch theoretical or practical significance other than the fact that it hadbeen published and used as a guide.So, if your segments are short (less than 300 meters) you can be sloppy withterminators and probably get away with it. Relcom has done tests that show1000 ft unterminated spurs work fairly well, and theory would predict thisresult. You may still have trouble with particular line drivers andreceivers even though the transmission line is okay. If you have longersegments, you can expect trouble if the terminators aren't within 300 metersof each end of the segment.It does take some work to make sure that your terminators are okay, but itis well worth the effort. If you use good terminators, they should not failvery often. We should use good terminators and good work processes toassure properly terminated transmission lines for our H1 segments.Herman StoreyShell Global Solutions, Automation Engineering (OGUS OGLA)