MODBUS maximum wire length 2

[guilio2010]

I'm going through calculations for various equipment and the one item I haven't found much information on was calculation the maximum distance for MODBUS communication based on wire size. I seen general rule of thumbs and basic equations in regards to data rate and distance, but I would think there would be something out there that would calculate your distance based on wire size for resistance (expected Vdrop of the signal) and the capacitance. For voltage drop conditions, would this be just your 12V signal voltage drop across the wire and capacitance of the wire affects the signal?

Thanks,
guilio

 

[xnuke]

I suspect you mean Modbus over serial communications on a TIA-485 shielded twisted pair (STP) network, if you use 24 AWG or larger, the distance is 1000 m. If you are using CAT5 cable, the distance is 600 m. These distances apply up to 100 kbps. If you try to communicate faster, the allowable distance gets shorter. I don't know of a formula involving voltage drop and filtering due to cable capacitance. Manufacturers of line drivers for TIA-485 using STP should have application notes that detail the drop in distance as a function of bps rate, so I'd look for those.

xnuke
"Live and act within the limit of your knowledge and keep expanding it to the limit of your life." Ayn Rand, Atlas Shrugged.
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[IRstuff]

The standard 485/422 driver has a limited sourcing/sinking capacity, and the longer the line, the more capacitance, which cannot be mitigated by thicker conductors. If you want substantially longer runs, you probably should be looking at a fiberoptic interface.

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529

 

[JG2828]

Like IRstuff said, for long distances you might want to look at a pair of fiber converters. I have had good luck with the Dymec / Garrettcom / now Belden line. 5846HRT-H is a part number that comes to mind.

https://www.belden.com/products/ind...nications/dymeclinksserialmediaconverters.cfm

 

[guilio2010]

I don't think we are at the point in distance of needing fiber optics. From readings and what I have calculated in distances, we are at about 2000' with 12 devices connected. Our baud rate is 9600. Due to he layout, we run up to the device, back to a central marshaling cabinet, then back out to the next device where we terminate at our last device inside of this marshaling cabinet. Just from reading and researching, I feel we are fine, but I want present this as to why we can reduce our wire size with calculations that also is supported with recommended practices.

We currently run 16/4 wire, 80pf/ft between conductors, 4.35ohms per 1000'. We have no issues with to date. I want to reduce the wire size to 20/4, 60pf/ft, 12.35ohms/1000'.

From my standpoint, and I am no communication expert, I'm trying to break this down with what I know of 2 components in the system. Resistance and Capacitance. The resistance will help me know what my voltage drop will be (to determine that we are still within the 12 to -7 volt input range). I'm not sure how the current is calculated. If the driver will output all the current it needs per device attached, which I think it will, then current will be the driver current per device....I think. Thus 12V - Voltage drop = what Voltage the receiver will see.

The other part is capacitance, which when I see this, makes me believe that in order to ensure I see the difference between High and Low of the signal at the device, the higher the capacitance of wire, the slower the rate I need to allow so that the capacitance will discharge.



Thanks,
guilio

 

[xnuke]

I think you're putting too much work into this. Why don't you buy something like this (I'm not affiliated with the manufacturer in any way):
2-pair Modbus 485 Cable

xnuke
"Live and act within the limit of your knowledge and keep expanding it to the limit of your life." Ayn Rand, Atlas Shrugged.
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[danw2]

B&B sells lots of RS-485 stuff. Assuming your topology (out and back and out again) is not really long stubs, but is truly point-to-point-to-point (daisy chained), then their graph of data rate vs distance says you're well within limits:

 

[itsmoked]

guilio2010; Seems to me your pf/ft numbers are really lousy bad. I'd pick cable with something better than 60pf/ft if I could.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[guilio2010]

Thanks all. I wish I didn't have to put this much work into something like this but this is what is needed to make a change.

For the wire, we generally run the same wire for all the devices. One pull can contain 12 to 20 of the same wires that will land on the 485 device but also provide power, AI, DI, DO, etc...by using the same type of wire, construction doesn't need to worry about separating different spools for different runs, landing the wrong type of wire on the wrong device, etc...I'll definitely look into a 20AWG with lower capacitance. At first glance it seemed that the capacitance was related to the number of strands.

 

[itsmoked]

Understood. I'd go with larger since DOs and AI can all remain functional out farther. The capacitance is a function of the insulation thickness. The closer the conductors get to each other the more the capacitance. It's a fine line in the manufacturer's design because farther apart can mean more opportunity for noise to pass between the wires of a pair verses lower capacitance.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[xnuke]

Ah, the last OP post cleared things up.

You're trying to make it too easy on the contractors in my opinion. They should be able to pull at least two kinds of cable without screwing it up. I second itsmoked's post to go with larger wire, then - 18 AWG at least since you said you're distributing power.

xnuke
"Live and act within the limit of your knowledge and keep expanding it to the limit of your life." Ayn Rand, Atlas Shrugged.
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[dV8r]

Maybe those contracted undocumented workers do not follow instructions very well?

 

[guilio2010]

I think in summary of this, one action item I will go is specifying the wire to match the device rather than trying for a one size fits all. We have a standard, stick with the standard. Realistically, we have all low powered devices and the only device I see that doesn't meet our standard is a small controller that would result in a 4.58% VD if we used 20AWG. All other devices are fine. I could twist these together with 2 of the 20AWG conductors to make 17AWG, but, like you all said, just pick the right wire in the first place. If the scope of work is outlined initially, then construction shouldn't have an issue. Right now, it's looking like 3 different sets of wires tentatively. 1 for 485, 1 for controller, 1 for all other devices.

itsmoked mentioned the insulation thickness and I got some reading last night. From what I am interpreting, of all this, the capacitance of the wire experiences a charge that effect the voltage, similar to a basic RC circuit. The Voltage Source can goes HIGH/LOW when transmitting data, but the wire will experience a delay due to this capacitance charging effect. (IMG14) When I see this, I go back and say if we are experiencing issues on older installation, I could slow the data rate, which will allow for more time for the charging discharging of the wire so that receiver can detect the voltage levels. Is this a correct analogy?

Also, I never understood the drive voltage range on RS485. I don't see why the output high isn't always 12V but rather a 12V to -7V. What effects the drive output voltage and is there a constant current that it is sending?

Last, why is the max distance always 4000'? It seems that if you had a low impedance/capacitance wire and did a low data rate (if the above capacitance analogy is correct), you could go further. Is this just a recommended standards? Hence the RS of RS-485?

 

[robertjo24]

Serial communications were never really meant to be run over controls cabling (salt & pepper). The main issues isn’t voltage drop, its line capacitance. A larger diameter cable will shorten your maximum effective communications distance. Comms links should be made via 22 or 24 AWG twisted pair communications grade cable, not 16 AWG instrument cable.

Here is a link to a simple primer if you’re interested.

http://www.lammertbies.nl/comm/info/RS-485.html

 

[guilio2010]

Based on itsmoked, so long the larger diameter cable has the same thickness insulation and dielectric as the smaller diameter cable, then it will experience more capacitance. Else, if they are separated further, then capacitance could be decreased. I totally agree that smaller cable for communication is the way to go, but from looking at this circuit equivalence, resistance should play some role into the distance as if your output from the driver X Volts and there is a Current, and you need Y Volts at the receiver, it should have a limitation. That's why I am looking at why is 4000 feet for lower baud baud rate the max for everything up to 100k?

 

[IRstuff]

You need to get rid of this hangup about the resistance. The termination resistance is typically 120 ohm.

http://www.ti.com/lit/an/slla272b/slla272b.pdf

at 11 pF/ft, 4000ft results in a time constant of 10.6us. Assuming 2 time constants per edge, we're up to 42.4 us, which is 25 kbps, irrespective of the line resistance.

TTFN
I can do absolutely anything. I'm an expert!
homework forum: //

http://www.engineering.com/AskForum/aff/32.aspx

faq731-376 forum1529

 

[itsmoked]

guilio2010; IRstuff stated it nicely. That's why it's always around 4000 ft.

As for "larger diameter cable" robertjo24 has used badly chosen wording (perhaps a language issue). He is using the term "larger diameter" to mean non-twisted power transmission wires which are indeed not communications wire. My capacitance description to you stands.

And, YES, your charge/discharge description fairly well describes the capacitance issue.

BTW: Consider dishing out out some stars here a lot of posts deserve them.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[guilio2010]

Thanks all. I'll start recognizing some stars.

Irstuff. Thanks for the document. I actually found in the datarate vs cable length that explained a question about can I go further distant then with slower data rate. I reworked out your example with 9600 bps and came up with 9800' but from the document: Section 3 presents the lower frequency range where the line resistance, and not the switching, limits the cable length. Here, the cable resistance approaches the value fo the termination resistor. This voltage divider diminishes the signal by -6 dB. For a 22 AWG cable, 120 Ω, UTP, this occurs at approximately 1200 m. Thanks for the link.

The only thing I need to find out is how the 485 has a driver voltage between 12 and -7 and how it's determined. Not that it matters but curious. Thanks.

 

[itsmoked]

The only thing I need to find out is how the 485 has a driver voltage between 12 and -7 and how it's determined. Not that it matters but curious. Thanks.

I'm not real clear on your question here... The device will have to have a +12V and a -7V source in it to provide those needed values. Now days there are integrated circuits that will provide those voltages without some extraneous extra power supplies involved.

Why those particular voltages is that 485 is a differential signalling method where the information is conveyed as the DIFFERENCE in voltage between the two wires and not the voltage between a ground and one of the wires. That means the signaling can be anywhere in a "region". Why specifically -7V and +12V is, I believe, because that's what the IC designers needed to provide the chosen differential signal robustly while allowing some unequal ground potentials "common mode" between ends of a long cable.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[guilio2010]

From my end, the specs states -7 TO +12. I see the TO as it can vary the output of the driver between these values and so long as your difference in voltage is 200mV to the receiver, then it detects the HIGH or LOW. I'm wondering what controls that output if it's +12 or -7 or +1, etc...is what I am trying to comprehend. I just can't see why it can vary for no reason. I'll research on my own and see if I can answer my own question first. I found more TI documents similar to what IRstuff provided and they have good information and will probably answer my own question.

Thanks,
guilio