BACnet Network Wiring
I am curious about terminating resistors in a BACnet network. I have been told several conflicting ways to use terminating resistors and am hoping someone can clear up how, when, and what to use.
I am currently under the impression you use a terminating resistor at the end of any network. I typically use the EOD resistor built into the equipment. Not quite sure if I should be using an external resistor, if it needs to be on both sides of the network, and how to chose what size to use.
If the equipment has a select-able terminating resistor, then you can use that. If the equipment does not, you have to add your own. You have to have each end terminated with a resistor.
Tech what system are you working with? I've been reading recently a lot of Bacnet wiring manuals for different manufacturers and it seems everyone says something different about something.
Originally Posted by MatrixTransform
My most current system has components from several different manufacturers including ACI, Honeywell, Carrier, Grundfos, and Lochinvar.
Generally speaking terminating resistors are placed at both 'ends' of the entire trunk. In my experience with working on Delta Controls, terminating resistors are left off unless there are comm problems that can't be remedied by any other means... then I add some terminating resistors to see if things get better.
Other systems I have worked with like Automated Logic pretty much require the use of their 'sponge bobs' (terminating resistors).
Originally Posted by TheControlsFreak
The man wins a prize.
If you don't know, and if it's a mixed bag of stuff ... leave the terminating resistors off. Unless you have a problem. Then try them.
Otherwise, if it is all one maker's kind of stuff, follow their rules.
Although ... fact is I've done that and the maker demanded terminating resistors in ALL cases, and the result was ... they were WRONG. Short runs and/or few controllers ... you got better results without the terminating resistors. LONG run, and/or LOTS of devices ... better with terminators engaged.
Just my experience with these things. I leave all terminating resistors and bias off ... unless I have problems.
The very reason it is important to know how the comm run goes. So yah know where EOL is, etc.
The RS-485 spec requires termination. The transceivers on the boards are all of differing brands/tolerances, and of course, noise levels vary, but I have never seen an RS-485 trunk work WORSE with termination. It is part of the design for a reason, without them, a high impedance network is far to prone to noise, even a differential one like RS-485. The terminating resisters, in essence, provide a path for current flow, meaning the signal's current flow down the wire is higher, so that the level of current induced by radio interference is less significant. Terminators therefore raise signal-to-noise levels. This may only be a problem with high noise environments - but I fail to see how anyone would make a conscious decision to forgo their use.
Originally Posted by osiyo
A lot of people don't understanding the different purposes of terminating and biasing of the network, they are two separate things serving separate functions.
Originally Posted by osiyo
"Profit is not the legitimate purpose of business. The legitimate purpose of business is to provide a product or service that people need and do it so well that it's profitable."
The best way is to use an oscilloscope on the network and check it. It will show the the effects of non termination vs termination,cable selection and noise on the line and let you check the wave form as well (ie, voltage, shape, etc) I check both lon and bacnet installs of any size to make sure that all is well.
Good day CraziFuzzy,
Originally Posted by CraziFuzzy
I do not have the RS-485 spec in front of me, but I believe it does not include termination. I say this, as "termination" and its value is solely dependent upon the characteristic impedance (ideally Termination resistor value should be the same as the characteristic impedance) of the wiring used... Secondly, the wiring to be used for RS-485 is not formally spec;d... other than it being twisted pair.
The RS-485 spec requires termination.
Indeed each manufacturer can use different RS-485 electrical devices... however, the devices will all have the same "specs" if they are RS-485 compliant. The issue is what circuitry the manufacturer places around the RS-485 device's Inputs...i.e. Do they include biasing (passive or active), input protection (transient, ESD, etc), etc circuitry? It is these circuits that create variances and/or potential issues to RS-485 networks... as these circuits are not formally defined within the RS-485 spec.
The transceivers on the boards are all of differing brands/tolerances, and of course, noise levels vary,
I have. The reason is that the termination is used to improve signal integrity of the communication signal and is based upon the wiring used (characterisitic impedance, Zo), and wiring topology. Effectively the termination resistors are meant to reduce the signal reflections that occur when an electrical signal encounters discontinuities in the wiring impedance. It is these reflections that cause distortions of the communication signal (i.e. reducing signal integrity) thus affecting/corrupting the content of the transmitted signal (Transmission Line Theory).
but I have never seen an RS-485 trunk work WORSE with termination.
Secondly, depending upon the overall impedance of the network (wiring and all of the RS-485 devices) adding the termination may cause too much current to be required to ensure that correct RS-485 voltage levels are met. Remember that the RS-485 spec dictates that a RS485 driver must supply at least enough current to drive a maximum of 32 unit loads in order to have a differential voltage of at least +/- 200mV. If one places more than 32 unit load devices on a RS485 segment one runs the risk that lower RS485 voltages are generated... and thus problems can result.
What part of the network are you referring to, as the wiring is relatively low impedance... usually anywhere from 100 to 200 or so ohms (the Zo value). That being said the RS485 electrical devices themselves have a much higher input impedance (anywhere from 12Kohm to 96Kohm), but given that the wiring is so much lower, it will dominate the system's impedance.
It is part of the design for a reason, without them, a high impedance network is <snip>
As for electrical noise, etc.... The termination resistor is not primarily used to deal with this problem (see above about signal integrity). If one uses the correct communication wire (i.e. twisted pair, etc) the differential signalling of RS485 will actually cause the induced electrical noise to cancel out. This is true for common mode noise, as this type of noise will be induced on both of the RS-485's signal wires and will be canceled out when the transceiver does its differential calculation (i.e. Voltage on A - Voltage on B).
With all that being said, adding a termination resistor at the beginning and end of a RS485 segment should not cause problems. However, given the reasons I stated above it can create more problems. In those cases one should really analyze the RS485 waveforms for their magnitude and wave shape as well as any obvious signs of AC coupled noise.
Good day Kevin,
Originally Posted by klrogers
Good day Deejay,
Originally Posted by deejay
Again, well said and very true.