RS-485 via RJ-45 ???

Learn the NEC with Mike Holt now!
I don't use modbus rtu on anything that speed matters on so I always use 9600 baud. Reliability matters a lot more to me than speed.
If you have a small number of devices, or very long runs, 9600 baud is fine. I think most of what I work on runs at 38.4, but we also have shorter runs and more devices that most, probably.

My introduction to Modbus ages ago was with people using it for very large manufacturing and extraction industries where runs up to a kilometer weren't unheard of. When I started working on power conversion equipment, the number of devices - inverters, meters, weather equipment - made 9600 baud impractical.
 
If you have a small number of devices, or very long runs, 9600 baud is fine. I think most of what I work on runs at 38.4, but we also have shorter runs and more devices that most, probably.

My introduction to Modbus ages ago was with people using it for very large manufacturing and extraction industries where runs up to a kilometer weren't unheard of. When I started working on power conversion equipment, the number of devices - inverters, meters, weather equipment - made 9600 baud impractical.
I would likely not use modbus if speed was an issue. I have run it faster but if there is no reason to do so, why bother. If I can interrogate a device every ten seconds and that is more than adequate why do I need to crank up the baud rate.
 
9600 baud is for wimps. I've not worked on microcontrollers that can go over 430kBaud, but I have written software, and tested it out with actual cables which can go that fast!

Terminating resistors aren't as important at 9600 baud, because the 3.5 character timer is so god-awful long (3.6 milliseconds), but if you are running it at more reasonable speeds like 38.4 or 115.2 kBaud you absolutely must terminate and absolute must use low capacitance
 
These lower baud rates aren't about the capabilities of the cable, it is the speed that is typically used by the protocol. Serial modbus almost exclusively operates at either 19,200 and 9,600. We have had issues with 19,200 but never with 9,600. Modbus transmits minimal data so 9,600 doesn't have issues at the lower speed.

As far as the need for terminating resistors, it is a function of data density and distance. If one or the other becomes two great you would need an EOL resistor. Since reflected waves will dampen in 3-4 cycles, if the time for this to occur is less than one data bit width (or one half the bit width if sampling in the middle), the reflected waves will not interfere and terminating resistors are not required. The formula I use to calculate the maximum cable length in feet is as follows:

Cable length = (Vp x C)÷(12 x baud rate)

Vp = velocity factor of cable (typically about 0.75)
C = 983,571,056 (speed of light in feet/second)

9,600 = 6,400 ft. (This would exceed the maximum length for RS485 transmission lines.)
19200 = 3,200 ft.
38,400 = 1,600 ft.
115.2kB = 530 ft.
 
In case anyone is interested, here is how I dealt with the situation. I opted for a breakout with a pigtail because I doubted a solid adapter would fit within the VDF cover. As it turned out, even the plug wouldn't fit within the cover. So I drilled a hole .... I do still like AB operators, their VFDs, maybe not so much.

Sure there's gonna be a lot of Coulda, Woulda, Shoulda. But I was there and you guys weren't. :)
 

Attachments

  • KIMG0541.JPG
    KIMG0541.JPG
    761.1 KB · Views: 22
  • KIMG0540.JPG
    KIMG0540.JPG
    1 MB · Views: 22
In case anyone is interested, here is how I dealt with the situation. I opted for a breakout with a pigtail because I doubted a solid adapter would fit within the VDF cover. As it turned out, even the plug wouldn't fit within the cover. So I drilled a hole .... I do still like AB operators, their VFDs, maybe not so much.

Sure there's gonna be a lot of Coulda, Woulda, Shoulda. But I was there and you guys weren't. :)
I like their VFDs because they work really well with their PLCs. Just get one that has the built in ethernet port or buy an ethernet adaptor for those that don't have it built in and plug it into your ethernet network. They have AOPs for all different drives, and it is so much simpler than using hardwired I/O.
 
I like their VFDs because they work really well with their PLCs. Just get one that has the built in ethernet port or buy an ethernet adaptor for those that don't have it built in and plug it into your ethernet network. They have AOPs for all different drives, and it is so much simpler than using hardwired I/O.
And thousands of dollars for the software license.
 
And thousands of dollars for the software license.
If you have AB PLCs you already have the software.

In any case, the labor savings makes up for the software license fee pretty quick unless your company believes labor is free, which admittedly a lot of places seem to believe.

You might be surprised at how little a yearly lease of a license might cost for just CompactLogix stuff. Even the perpetual license for CompactLogix is not that big of a deal.
 
If you have AB PLCs you already have the software.

In any case, the labor savings makes up for the software license fee pretty quick unless your company believes labor is free, which admittedly a lot of places seem to believe.

You might be surprised at how little a yearly lease of a license might cost for just CompactLogix stuff. Even the perpetual license for CompactLogix is not that big of a deal.
I am with a service contractor. We don't have PLCs of our own. and there are customers with AB PLC that do not have a license for the software.
 
Okay, that helps. In my opinion, CAT-5 is the wrong type of cable for this purpose.

I would use a 2-conductor-plus-shield cable, with a foil shield with a drain wire.

View attachment 2573709
CAT5 is just fine for RS485; it has (4) low-capacitance twisted pairs and can be shielded if the right cable is spec-ed. The Entertainment industry uses CAT5 and RJ-45 connectors all the time for DMX lighting control (which is just a protocol on top of the RS485 physical layer).

It's cheaper than "Belden" RS485 cable, is more available, and more techs know how to install and terminate it. The DIN rail mount breakouts are perfect for breaking out the RJ-45 into individual terminals if needs be.


SceneryDriver
 
As far as the need for terminating resistors, it is a function of data density and distance.
It's also a function of the differential driver electronics and the slew rate- better devices use purpose-designed driver chips which limit the rate, some devices simply use electronic switches and heap on the harmonics; Maxim has a discussion of this at https://www.analog.com/en/resources/technical-articles/explanation-of-maxim-rs485-features.html. and *Hobbit* shows some scope photos at http://techno-fandom.org/~hobbit/lighting/dmxwave/
 
Rockwell have a RJ45 plug with 2 terminals, part number AK-00-RJ45-TB2P.
You can use this to daisy chain PF4x and PF52x drives as shown in their user manuals.
There is a diagram in TechNote QA18998 (access everyone, you just need to create and free account) that explains how to wire using the above plugs.

If I could work out how to attach an image I would have shown the network drawing in the TechNote.
 
Rockwell have a RJ45 plug with 2 terminals, part number AK-00-RJ45-TB2P.
You can use this to daisy chain PF4x and PF52x drives as shown in their user manuals.
There is a diagram in TechNote QA18998 (access everyone, you just need to create and free account) that explains how to wire using the above plugs.

If I could work out how to attach an image I would have shown the network drawing in the TechNote.
Would have been nice if they mentioned this in the manual. Sure glad YOU did!
 
Seems I finally figured out my problem. After inserting the wires through the connector, I had been pulling them too snuggly in order to get the sheathing well into the plug for make a sturdy assembly. Just pulling far enough to engage the sheathing seems to work better.
 
Top