How do I store and forward data in MicroLogix 1400 PLC over radio telemetry

skandoll

Member
Join Date
Dec 2016
Location
Silverdale, WA
Posts
3
Hello

We have approximately 60 remote sites which have Remote Telemetry Units using MicroLogix 1400 PLCs and Viper SC 200 radios. We have a Master Telemetry Unit using a CompactLogix L32E which polls each RTU 1 at a time in order and collects about 8 binary values and 3 analog values.

Our system takes about 5 minutes to poll every site. Fixing the polling times is an ongoing project, however in the mean time we would like to be able to somehow store the flow rate every 30-60 seconds in the RTUs and have the MTU grab all of the stored data and then somehow have our Wonderware Historian save all of the data properly so that we get a nice pretty trend of flow (currently we get what looks like a square wave since the data only changes once every 5 minutes).

Any ideas?

Thanks,

Shawn
 
You would have to timestamp them in the PLC, then when they arrive use tome code to "push" them into the historian. I wouldn't think that would be a trivial thing to accomplish.
If I remember right, DNP3 protocol does exactly this so you could perhaps implement that instead.
It shouldn't take that long to poll all those radios, I do 30 every minute and that's mostly because I am running PLC5 as my polling master. Maybe addressing that first would be a better option.
 
Hello

We have approximately 60 remote sites which have Remote Telemetry Units using MicroLogix 1400 PLCs and Viper SC 200 radios. We have a Master Telemetry Unit using a CompactLogix L32E which polls each RTU 1 at a time in order and collects about 8 binary values and 3 analog values.

Our system takes about 5 minutes to poll every site. Fixing the polling times is an ongoing project, however in the mean time we would like to be able to somehow store the flow rate every 30-60 seconds in the RTUs and have the MTU grab all of the stored data and then somehow have our Wonderware Historian save all of the data properly so that we get a nice pretty trend of flow (currently we get what looks like a square wave since the data only changes once every 5 minutes).

Any ideas?

Thanks,

Shawn

what is the purpose of the compactlogix? Does it do anything with the data from the micrologix outside of acting as a data concentrator? If not I would remove it. There is nothing more annoying to me than using a plc as a data concentrator for no reason.

I cant speak for the allen bradley protocol but if you convert to modbus on the 1400's, there is no reason why you shouldnt be able to poll all of the sites in 1-2 minutes. I am guessing that the compactlogix is causing the bottleneck. Also, if you have 5-10 stations that are missing polls in the poll cycle, a high response timeout like 10-15 seconds could add a couple of minutes while it waits for the station to timeout. reducing that could speed up the cycle.

As far as storing the data then pushing it into wonderware, I cant speak to the way of doing that but I doubt it will be an easy task since it will be coming from more than 1 tag. You would need to look at an exception based protocol like dnp3 or mqtt, but I am not sure of wonderware's support for these protocols and the exception based feature.
 
If they are Rev B 1400s they support DNP3. Get a DNP3 engine (ClearSCADA or TOPserver DNP3) and do away with the MTU.

You'll need to do some reading on DNP3 to understand the Class 0, 1 and 2 points for remote logging.
 
what is the purpose of the compactlogix? Does it do anything with the data from the micrologix outside of acting as a data concentrator? If not I would remove it. There is nothing more annoying to me than using a plc as a data concentrator for no reason.

No it does not. It polls the RTUs and stores the data until the next poll cycle. Wonderware then grabs the data out of the CompactLogix.

...I am guessing that the compactlogix is causing the bottleneck. Also, if you have 5-10 stations that are missing polls in the poll cycle, a high response timeout like 10-15 seconds could add a couple of minutes while it waits for the station to timeout. reducing that could speed up the cycle.

I do believe the CompactLogix is causing the issue. If I understand this correctly, the CompactLogix has a max TCP connection inactivity timeout of 120 seconds. Since our entire poll cycle is above 120 seconds it causes each TCP connection to restart each time it polls an RTU. I also was reading that the CompactLogix can only have 32 simultaneous CIP connections. I am not sure if that is an issue as well though because we are using the SLC Read type message instead of the CIP Table Read type message.

When my partner and I were hired on 6 months ago our poll cycle time was about 10 minutes. We have managed to improve that to 5 minutes so far by improving radio paths, eliminating repeaters where they weren't necessary, etc. Our goal is to have a -85db or better signal strength from each radio. Once a radio consistently gets -85db signal or better we have been increasing the data speed from 4Kbps to 12Kbps (we only can chose from 4, 8, and 12Kbps with the selected bandwidth setting, which is set by the FCC). It used to take about 25-28 seconds for some of our really bad RTUs to communicate so we had our response timeout set to 30 seconds. We have improved slightly on this and can now safely have it at 20 seconds without missing too many polls.

I cant speak for the allen bradley protocol but if you convert to modbus on the 1400's, there is no reason why you shouldnt be able to poll all of the sites in 1-2 minutes...

If they are Rev B 1400s they support DNP3. Get a DNP3 engine (ClearSCADA or TOPserver DNP3) and do away with the MTU.

You'll need to do some reading on DNP3 to understand the Class 0, 1 and 2 points for remote logging.

I checked and our MicroLogix 1400s do have series B processors and do have the option in the Ethernet channel configuration (Channel 1) to enable Modbus TCP.

I am not very familiar with Modbus/OPC server. Is it possible to use the TOPserver to poll each RTU one at a time similar to how our CompactLogix is polling? If that would poll faster than the CompactLogix and result in less than a 3 minute poll cycle we would be very happy! 🤞🏻
 
I cant speak for your system, but in my system, pulling that amount of tags using modbus tcp over a cellular modem would take about 1 second, and thats with the 3-600 ms lag time inherent to cellular. Then if one fails you have to look at things like the timeout, and then the retries. Topserver is essentially kepware so I would have to look at kepware to see what options they give but I seem to remember being able to set the timeout and then retries, there is also a demotion mode, but I dont remember much about that.

You could do a comparison on how long it takes to poll one using ab protocol vs modbus tcp to get a gauge on how long it would take, but I doubt it would make a huge difference. I believe there should be a tag that reports the last scan time, but you may also have to use use their diagnostics to look at the start and finish.

*** also, if you swap a micrologix to modbus tcp, you have to save the program and then power cycle the plc. you wouldnt imagine how many times people get tripped up on that due to the power cycle..
 
Most slow SCADA systems suffer from more than one problem. If you were able to improve performance at all by tuning radios and repeaters, then reliability of the radio network is part of the problem.

You are correct that the 1769-L32E is limited to just 32 CIP connections and 32 TCP/IP connections. The messaging to a MicroLogix 1400 is a CIP connection even though it's not a ControlLogix family controller; the MicroLogix 1400 uses CIP but uses it to transport SLC-style messages.

You're having to wait for the TCP/IP connections to time out before it can make more. The Wonderware computer and any online sessions of RSLinx are also taking up TCP/IP and CIP connections. The CompactLogix was simply the wrong design choice.

If you have the opportunity to test this system out with a 1756-L6x or 1756-L7x controller and a 1756-EN2T (128 TCP connections), I predict that your performance would increase significantly. Even if you had to move I/O off the CompactLogix to do this, it would require the minimum degree of effort.

There are a handful of programmers and consultants in your area who I recommend for this sort of thing; send me a private message if you want a referral.
 
The way I see it is that there really is no reason to bring this data into a PLC at all. Ive seen enough of these disasters to know that I avoid using a PLC as a data concentrator at all cost.
 
That may be the only way to poll at the interval needed. Kepware's Scheduler wont do it very well unless the pipe is pretty fast.

I have a hard time believing that. Kepware wuldnt be in business anymore if they couldnt poll 60 rtu's in 2 minutes.
 
I service a network of 65 sites across (4) old MDS MTUs. (1) MTU is local and (3) are scattered throughout the county.

We used (4) Pro-Soft modules and an L35E and can poll all of these sites in around 30 seconds. I believe the ProSoft module buffers the comms from these but can't remember exactly. Might be worth looking at.
 
Am I correct that you are using the licensed radios because line of site isn’t achievable? If that’s not the case, I.E you can achieve line of site, then a 900MHz frequency hopping spread spectrum radio is a much better option. You will be able to get a much higher baud rate and a much better signal to noise ratio. If that is the case (no LOS) then if you want to get better throughput you need to change they system design. One master with 60 remotes simply isn’t going t get the job done and I’m surprised you’re able to get the results you’re getting now. Depending on the topology you’re either going to need to update your license to reflect multiple masters on different frequencies or if you can use sector antennas pointed in different directions and divide the network with three or four masters you can reduce the number of remotes off of each master allowing the poll rate to drop from 60 to somewhere around 15 to 20 (sector antennas should prevent you from having to update your license but I don’t know that for a fact). The other option would be using cell modems, but they are far from reliable in many cases and using 60 of them would be very complicated.
The problem with “store and forward” is you really aren’t saving any time. All you’re really doing is stacking data up at one location so when you get to that location it takes longer to download the data (because there’s more of it). Increasing the throughput or reducing the size of the network are the only two ways (that I see) to increasing efficiently in most wireless networks.
 

Similar Topics

Hi - I have an application where we are planning using an Edge device to connect some equipment, likely an S7-1500 and WinCC Unified Comfort...
Replies
0
Views
919
I am drawing a complete blank this morning looking for a unicorn. I would like to install a device at a remote site that will gather the data...
Replies
15
Views
2,240
I'm running into a problem when trying to use the RSLinx Backup Restore Utility where it throws an error when it tries to restore a backup. The...
Replies
6
Views
453
Hi y'all I have a KTP400 connected to a s7-1200. Now i have on my screen the encoder value which comes from my SEW movitrac. under there I have...
Replies
2
Views
368
This was asked in another forum and taken down due to being a 'backdoor', but this has useful info that might be useful to someone in a pinch. I'm...
Replies
6
Views
565
Back
Top Bottom