Newer Controllers on N2 Bus Wiring
Hi, I'm an HVAC engineer working on putting together specs for a building automated controls overhaul for a 7 story office building. The goal is to switch out old N2 controllers with newer controllers but still use the N2 bus wiring ran throughout the whole building.
I have scoured the internet looking for technical specs about the actual, physical wiring that makes up an N2 bus and how that differs from other control systems wiring without much luck. I also fear I am misusing the term "N2 bus", my understanding of the term is that it refers to pretty much everything in the control system; the front end, the controllers and the wiring and all the gear in between. So my question is can other manufacturer's put their own controllers on the existing equipment and utilize the existing N2 bus wiring for the new BAS. And if that's not possible, what is it about the existing N2 wiring that prevents it?
Thanks so much!
JCI was pretty loose on the wire spec. N2 info is on their site.
Metasys Network Technical Manual 636
Network Communications Section
Issue Date 1199
Tip Search: N2
Suggest you look at what wire is in the building in question. N2 works over an RS485 serial connection. 9600 baud was pretty forgiving compared to newer faster networks.
Great, thanks for the info, quite helpful.
The first Controls class I went to back in 98 said to NEVER reuse trunk wiring. The potential problems with comm can create havoc on a job. Just a little tidbit I remember.
Don't worry zombies are looking for brains, you're safe...
Good day Bhollon,
Originally Posted by bhollon
BillControls is right... JCI's wiring specs were all over the place and branches would use pretty much anything they could... As a consequence the probability of being able to reuse existing older N2 wiring for newer systems (i.e. MS/TP, etc) is extremely low...
As for N2 terminology, you are correct. The term N2 was used to identify both the protocol (actually there are three that coexist on a "N2" bus) and the electrical interface (RS485/EIA485). As BillControls also stated the "N2" bus electrically is a RS485/EIA-485 bus... Which JCI violated as well in terms of the number of controllers that can be installed on a RS485 segment. The RS485 Spec is rated for 32 unit loads and JCI in numerous cases exceeded this value dramatically... The result was a reduction in noise immunity and transceiver drive ability... which periodically caused site communication problems (i.e. random offlines, etc).
If you are upgrading your site, spend the time and investigate the installed wiring to ensure that it is fully RS485 compliant (i.e. twisted pair with a reference lead, low capacitance, has a characteristic impedance of 100 to 200 ohms or so, etc). Secondly, ensure that it is daisy-chained wired (i.e. no stubs or star configurations). If your wiring fits the above conditions then you are extremely fortunate and there is a good probability that the wiring will not cause you grief. If not, then replace the wiring with correctly spec'd RS485 wire and you will be set for whatever RS485 based system you wish (i.e. BACNet MS/TP, etc).
RS-485 is a two wire protocol. JCI modified it to a 3 wire system to expand the devices. While its loosely RS-485 it really isn't and that's why they can use 64 (or 100) devices on the bus. I would say they hijacked a protocol .. It's been pretty solid despite a lot of abuses since 1990 so I think they must've had someone pretty sharp on the job...
The real reason I replied is I disagree with your statement about varied standards. FAN-410 outlines a very strict requirement for N2 cabling (and MS/TP)... Down to the color. I would agree that I see just as many sites using the wrong wiring but I would say its generally poor electrician management (who at JCI systems pulls wire over 20-50 ft?) but in most cases I see third party contractors that "know JCI" causing the issues. That includes incorrect wire size, length, too many devices, etc. I could care less who installs it.. Just do it right. Granted I've never really visited any sites on the east coast or down in the Deep South... I can say the west and mid-west in general the branches have excellent wiring practices ... Though everyone has that occasional bad-egg...
If you're looking for specifications look up FAN-410 or the N2 Troubleshooting Guide.. They'll specify a specific type if wire manufactured by Belden.
Good day chas_md,
Originally Posted by chas_md
The RS-485 interface is actually a three-wire interface and not two-wire as you stated. Many people and companies make this mistake, as it is true that there are two signalling wires that determine the logic level of the signal (RS485 is a differential signaling system), but the third wire (the reference) is critical for correct long term operation of the bus as well as minimizing the effects of ground currents. Indeed, for short distances one could ignore the reference wire and all would probably work fine, but it is best to keep things consistent and use a total of three wires (1 twisted pair and a reference).
Secondly, JCI did not modify RS-485 to be 3-wire (my explanation is above) nor did they "modify" it to allow for more devices on the bus. If you do not believe me take apart any of the JCI devices and examine the N2 interface... They are a all industry standard RS-485 devices with enhanced electrical protection. There is no magic here nor any enhanced engineering to allow for for than 32 unit-load (UL) devices on the bus. What happens when you exceed 32 unit-loads? Well, you violate the specification and so your bus may or may not work as it is supposed to. The first item to be compromised is noise immunity (200mV noise floor)... the second will be drive ability... If you wish I can go into greater engineering detail on this and prove to you why this is the case. The only reasons why JCI was able to get away with violating the RS485 spec was simply do to the robustness of RS485 to begin with, the N2's low data rate, and the supervisory's relatively low polling latency.
As for newer devices... There now exists (since the late 90's) RS485 devices that are rated less than 1 unit load and so one can add more than 32 UL on a bus. You can buy 1/4 UL (48Kohm input impedance) or 1/8 (96Kohm input impedance) UL devices and so with these you can have 128 or 256 devices on a RS485 bus segment. And for the record JCI did not use these type of devices on their legacy N2 devices... they were all 1 unit load devices.
Thirdly... as for JCI's direction for wiring... I have never actually seen an offical JCI publication for N2 wiring that "must" be used. I have heard of the document, but I never found it. If you have a copy or can tell me the wire they specify I can check into its specs and will report back. That being said I have seen JCI's wiring recommendations for MS/TP and the wire they recommended is not even rated as communication cable... it is wire used for intercoms... Will it work? It can work, but you run the risk of having problems if you decide to go to faster baud rates, longer distances, etc.
Interesting reply; a good read. I don't know every aspect of the RS-485 protocol nor do I ever really plan to learn it so I can't argue on what it "is" and I am sure you could offer PDF references here and there but it'd be a bit heavy for a post asking about re-using comm wire perhaps? Not that it's the best source in the world but Wikipedia http://en.wikipedia.org/wiki/RS-485 states that it was a two wire network; "This allows RS-485 to implement linear topologies using only two wires." .. yes they follow on to add the note about the third wire reference wire. Many vendors use the 2 wire set up as I'm sure you know. I only reference this wikipedia document as it's the only immediate note about the subject and what I was taught in my original ASC class when I started with Johnson Controls and in fact I was just regurgitating what I was told (drink the blue kool-aid!); perhaps it was wrong? I can't say one way or another. I also have paper documentation on N2 stating it's recommended Belden number from that same class (Back when you got notebooks instead of CDs or a flash drive). If I can find the training binder in short order I'll see about scanning it. In the office we just call Anixter and say "Send me a box of N2 wire"... it shows up Blue, 18/3 and shielded.
On your note about MSTP wire: Agreed, it's not top secret that it is rated by Belden as a Commerical Audio system wire but it was chosen by the engineers because of it's capacitance and it worked best with 38,400 baud (again blue kool-aid manual). This was widely discussed in the FEC BACnet/MSTP roll out what 5 or 6 years ago now? You can talk to any of the FSMs if you have access to one. I really never had a reason to doubt a factory course because I've had no problems dropping 100 devices with the Belden B6501FE within length requirements.
Originally Posted by bhollon
You have quite a task ahead of you I'll say that much.
I consider s2sam one of the "experts" on the site when it comes to the hardware (technical) aspect of what you refer to. Some of the others may show up with comments in a while.
In most cases what typically exisits for "N2-Bus" wiring, if it was installed after about 1995, is an 18 ga. 3 conductor cable with stranded conductors, a shield and a drain wire. It is almost always a CL3P rated cable. Just based on my experience I would recommend aginst trying to re-use the existing cabling. JMO .
The basic info you provided is that you are going to "...switch out old N2 controllers with newer controllers...". With that in mind the "newer" controllers will probably NOT be "N2 Devices". The "newer controllers" could be any of 3 major communication protocols (BACnet, LON, MODBUS) or a number of other minor protocols which I won't even attempt to list. MODBUS is the only one that would probably be able to use the existing communication cabling but I am not real familiar with the details.
I will try to post a link to specific JCI documents that relate to your quest for "Tech Specs" later.
Good Luck !
Controls Cause Insanity
Last edited by Cagey57; 03-26-2013 at 10:51 PM.
Reason: eye kint spill two gud
If sense were so common everyone would have it !
All opinions expressed are my own. Any advice provided is based on personal experience, generally accepted fact or publicly available information. As such, it is worth exactly what you paid for it, not a penny more not a penny less !!
Back to the original question, many types of wire may have been installed as "N2 Bus". If you want to reuse it you should inspect the site to see what wire is installed. And just as important How was it installed. Good twisted pair wire 22 to 18 gauge installed in a neat logical manner both in the panels and in between could possibly be reused for Bacnet. You would likely need to break up the Bacnet into shorter segments than the N2 used (floor by floor with MSTP to IP Bacnet routers?). BTW I would spot check on every floor during inspection. Joe may have done a great job on the first floor but Jack on the 4th floor may not have been the best man off the bench.
Also consider that one or two problems in the existing wiring will eat up any savings over running new. Last year one of my service techs looked for a problem for three days in some N2 wiring in a school. He finally decided to replace the wiring between the units. 4 hours later he was done and the segment works. Sure wish he made that decision on Day 1.
If the new equipment is LON install new LON wire.
The Old Proprietary networks like N2 and FLN ran slower than the modern networks (9600 baud). They also transmitted a lot less info with the data. The N2 mostly just transmitted a point # and the value. Bacnet transmits all the object data all the time(object name,description,units, alarm values, object state, value... . The head end of the old systems filled in a bunch of the data that Bacnet transmits from data files that were made when the system is built. New systems let you discover all this on the fly if you want to. What this means to converting old wire to new is the new controller needs to put a lot more data through the wire. So a faster speed is selected and you can not put as many controllers on the same segment as on the old networks.
As others have said whether you go LON ("Friends Don't Let Friends Do LON") or BACnet you will want to use either 22 or 24 gauge wire for the faster communication speeds.
Also don't make little service loops in the wire at the controller, it just confuses all those little electrons and drags down the com.
When you decide upon a product let us know and we can recommend a cable type and some tips on making terminations. I installed N2 products when they first came out and was @ big blue for almost a quarter of a century so I have some experience with it.
I don't care what system you have be it N2, BACnet etc just install it by the book and don't cut corners. Then if a problem arises there will be a lot fewer factors to troubleshoot or sort through.
If you can't fix it with JB Weld, Duct Tape, and Ty Wire it has to be replaced.
No good deed goes unpunished.
If you want to take off friday to go fishing then make sure you train your helper right.
Good day Charles,
Originally Posted by chas_md
Indeed, one does not want to get into this too deep, but I feel it is important to go into some finer details for the benefit of others who use RS485... whether it be N2, Modbus, MS/TP, etc.
That being said... You are quite correct that RS485 only requires two wires to signal digital data and this digital data is "encoded" by the voltage difference between these two wires (The voltage difference has to be either greater than 200mV or less than 200mV in order to be a valid signal level and anything in between is considered indeterminate and thus can be incorrectly encoded). The benefit of differential signaling is that any electrical noise that will appear on both wires (common mode noise) which will cancel out when the RS485 device does its voltage difference. How does one get the electrical noise to appear on both wires? Well, if one uses a consistent twisted pair wire then the electrical noise will be induced on both wires... So rule one if you want noise immunity then use twisted pair wire...
Now, why do I and the Industry consider RS485 to be three wire? It is true that only two wires are needed to represent the binary data, but now one must consider the real World and this is where the third wire (reference) comes in... The reference wire ensures that the various RS485 attached to one another have a common voltage reference. Why is this important? Well the RS485 electrical components have limits as to how high the signal voltages can be before the RS485 electrical component becomes damaged. The RS485 spec dictates that the voltage can be anywhere from -7V to +12V and so the RS485 electrical components must be able to withstand these voltages. In most cases RS485 component manufacturers exceed these values, but the spec is what it is. Remember that these voltages are "referenced" to the RS485 electrical component's local ground pin... Thus, there is no guarantee that each and every RS485 device connected to one another will have exactly the same ground voltage level....unless , of course, there is a common reference point... which is where the third wire comes in. By having a common "reference" point each device can maintain adequate absolute voltage levels to each device's internal RS485 transceiver and thus minimize any potential damaging over-voltage conditions to the various RS485 devices. So, the reference lead is not used directly for signaling, but for maintaining a consistent reference level for the benefit of all of the RS485 components.
What happens if you ignore and/or not use the third wire (reference)? Well, you run the risk of damaging the various devices on your RS485 network and possibly have inconsistent, erroneous, or a problematic RS485 network. Will this be the case all the time? No, but it will be site, RS485 product, and wiring topology dependent. If you have a network where the devices are located close to one another, then the third wire is probably not really required. However, if you have your network spread throughout a large building, then problems can occur. If your network is spread across buildings, etc then you are almost guaranteed to have issues.
As for the wire... Here is where I will keep this brief, as we would have to enter the World of Transmission line theory which is not for the faint of heart... Given the robustness of RS485 you could pretty much throw any twisted pair and get it to work (most times) with N2... The two main reasons for this is because of the low N2 data rate (9600 baud) and the relatively low supervisory's poll latency. However, as soon as you start to increase the data rate (i.e. frequency) and/or the poll rate problems will start to occur. The main reason is that the wire's frequency dependent parasitic elements (capacitance and inductance) start to distort the digital signals on the wires which affects the interpretation (encoding) of the voltage levels. Thus, the wire that is working with an N2 installation may or may not work for other RS485 protocols, increased data rates, or increased distances unless it was properly specified to begin with. This is where I get crotchety because if the proper wire was installed to begin with it could have been used for pretty much whatever RS485 system you would want.
As for JCI's MS/TP recommendation... I do not know if you realize this, but the Belden B6501FE is not even twisted-pair wiring... which means you are giving up the maximum amount of noise immunity. Secondly this wire has a capacitance of 55pF/Foot.. which is 4 times that of properly spec'd RS485 wire. What this means is you will be out of luck should you wish to move to higher data rates in the future. Thirdly, there is no spec given for the characteristic impedance (Zo)... Which means this is not a communication wire and so transmission line problems (reduced signal integrity, reflections, etc) will almost certainly occur and thus data integrity problems will increase... i.e. a much much higher probability of a flaky bus.
With all that being said, I have tried to provide some engineering and insight into the low level details of N2 and RS485 in general. Hopefully those that have read my ramblings will leave with some better understanding of these technologies and in doing so help them to understand, diagnose, and install a more robust RS485 network.
N2, Lon, Bacnet or what ever comm you are going to use, IT IS NOT WORTH THE HEADACHE TO REUSE THE COM TRUNK. I have worked on jobs that were bid to re use the comms and no matter the protocol it has been a mess, improper cable routing, splices, termination points, the list can go on and on. Just replace it, in every job it is the unknowns that hurts the most.
"It's always controls"