How bad is serial encapsulation over Radio, and is that the only concern?

TheWaterboy

Lifetime Supporting Member + Moderator
TheWaterboy
Join Date
May 2006
Location
-27.9679796,153.419016
Posts
1,924
I need solid reliable comms and because of Terrain and RF congestion I have to communicate my telemetry over licensed VHF (220 MHZ) That of course means low bandwidth.

The PLC's currently selected for the remote location are Ethernet only L30's . No serial (DF1) comms are even available for these. That part might be OK in this case because as I understand it, in addition to having smaller limitations on lengths of MSG elements, DF1 type comms must use N Registers and the desire is to use a single UDT containing Dints, Reals, Bits and a few Strings. This UDT would contain all the info that field site has to offer. Nice and clean bundling back at the concentrator end.

DNP3 was briefly discussed but I haven't read up on DNP3 enough yet to even think about what it would mean to this discussion. I welcome input on that too.

So, if all the above is true that leaves Ethernet comms as the only option for a data chunk this size (about 1k words).

So I set up a test in the shop using a MSG to transfer 500 reals (reading from the L32 to an L30) using a CalAmp Viper pair I have available.

When hardwired the MSG takes 45 ms to complete but when run over the radio it takes from 2.8 to 11 seconds to do the same task.

Can someone confirm for me this seems like normal numbers?

If I had someone in my area I could believe in (ahem, Ken Roach... :) ) I would buy a few hours of time to consult, but everyone seems to have their own angle depending on what they rep. I can get a concise RF consultation and did that, but the PLC to Radio / protocol selection is a different beast to get straight answers for, and I know enough to know I'm wading into deep water.

I welcome comments or pointers to useful materials.
 
Firejo is in Redmond and he does radio stuff, I believe.

Does both location gets good cell service? You can also try cell hotspot VPN.
 
We did consider cellular, and might use it for cameras at some point, but I don't want to depend on the cell company for my reliability numbers. If a station is going bad we need to know about it immediately and not wait in the dark till their tech can fix whatever went bad.
 
Yea, that I knew and I'm accepting of that speed in exchange for reliability. You should see the 900 spectrum on an analyzer, its like a square wave that starts at 900, Every traffic light in the city, and everybody else around seems to be using that spectrum.

It's now a choice of trying to use the ethernet channel or change plc's and schema to use only serial comms.
 
1000 words, 2000 bytes, 16,000 bits. depending on the baud rate 2-3 seconds might be in the right ballpark depending on how wide your channel width is set as the web page lists bitrates as low as 4000 bits per second: http://www.calamp.com/products/ip-modems-and-routers/viper-sc

If it takes 11 seconds sometimes that might indicate packet loss from radio interference. Sometimes shop environments are bad for testing since there are different obstacles and reflections than in the final location.

you can use radio mobile to see how good the radio path is but it sounds like you already have the RF side answered: http://www.ve2dbe.com/rmonline.html

One nice thing about DNP is it can do report by exception, so only values that have changed by more than the deadband are sent immediately, and otherwise all values are sent at some regular interval.
 
V0N_hydro said:
One nice thing about DNP is it can do report by exception, so only values that have changed by more than the deadband are sent immediately, and otherwise all values are sent at some regular interval.
Click to expand...

The field sites would not currently initiate communication, they would be polled, If I can get 35 sites polled quickly enough I wont need exception based comms. But I will file that away.

I recall that DNP3 timestamps the data and that could be useful.
 
TheWaterboy said:
I need solid reliable comms and because of Terrain and RF congestion I have to communicate my telemetry over licensed VHF (220 MHZ) That of course means low bandwidth.

The PLC's currently selected for the remote location are Ethernet only L30's . No serial (DF1) comms are even available for these. That part might be OK in this case because as I understand it, in addition to having smaller limitations on lengths of MSG elements, DF1 type comms must use N Registers and the desire is to use a single UDT containing Dints, Reals, Bits and a few Strings. This UDT would contain all the info that field site has to offer. Nice and clean bundling back at the concentrator end.

DNP3 was briefly discussed but I haven't read up on DNP3 enough yet to even think about what it would mean to this discussion. I welcome input on that too.

So, if all the above is true that leaves Ethernet comms as the only option for a data chunk this size (about 1k words).

So I set up a test in the shop using a MSG to transfer 500 reals (reading from the L32 to an L30) using a CalAmp Viper pair I have available.

When hardwired the MSG takes 45 ms to complete but when run over the radio it takes from 2.8 to 11 seconds to do the same task.

Can someone confirm for me this seems like normal numbers?

If I had someone in my area I could believe in (ahem, Ken Roach... :) ) I would buy a few hours of time to consult, but everyone seems to have their own angle depending on what they rep. I can get a concise RF consultation and did that, but the PLC to Radio / protocol selection is a different beast to get straight answers for, and I know enough to know I'm wading into deep water.

I welcome comments or pointers to useful materials.
Click to expand...

could you make the shot with 450 mhz instead of 220? I know a few companies make 450 mhz ethernet, trio being one of them.

I wouldnt worry about reliability with cellular. Only thing you have to worry about is a monthly bill. depending on how fast you need to poll, you ae looking at $10-$100 per month.
 
I don't want to mis-lead anyone, but we use cellular and have for several years. We have better UP time with cellular then with any of our Radio equipment.

Multitech cellular modems are used to monitor heart patients out in the real world, what could be more important??

Now with that being said "with a cellular connection and a VPN Tunnel" your speeds would compare to a 10 mbs network.

We have several remote monitoring unit in very remote location all setup with Cellular/VPN and this allows us to not only monitor but we can 100% completely re-program the units and even do firmware updates.

All of our new installs that require some type of wireless communication will all be cellular, unless your talking something that is a few thousand feet apart then we just use a wireless bridge.
 
TheWaterboy said:
Can you get a service level agreement from the cellular vendors that specifies damages for downtime?
Click to expand...

I seriously doubt that they would agree to something like that. plus you would probably have a very difficult time talking to someone that would have the authority to agree to that.

If you go cellular just be prepared because you are completely at the mercy of the phone company and their support is pretty terrible. dont expect to call and talk to someone that knows thatthey are talking about.

somehow it just works though.
 
As I said, I don't want to mis-lead anyone, but does your Radio Vendors provide one for you now?

Because if they do, I'm sure the cellular vendors will provide the same agreement.

All I'm saying is things have changed, you owe it to yourself and your client to sit down with some vendor such as Verizon to see what they have to offer.

We did, and we was surprised to find out everything they had to help us move forward in the IIOT world.

And all at very affordable prices.
 
No, of course not. but radio vendors are not service providers. Different requirements. But I will investigate again. It would be great to have that kind of connectivity... Last time I poked at Verizon they sent me to a third party vendor instead of dealing direct. I thought that was odd so I haven't tried since.
 
I love DNP3. Strongly recommend it instead of cyclic polling. But if you already have the compactlogix processors then I'm not sure that's an option?

I believe Schneider have some VHF Ethernet radios in the pipeline. Their Trio Q UHF radios are very good. Not sure when they will be available though.

Take a look at the SCADApack 575 if you haven't bought hardware yet.
 

Similar Topics

ndzied1
USB to Serial (Good, Bad, Ugly?)
Hi All: I've taken a new job where I need to purchase my own laptop. The company will help finanace an IBM T40 which has reviewed well but has...
2 3 4
Replies
54
Views
41,653
JesperMP
JesperMP
R
AB ultra3000i drives going bad
Hello all, I am brand new here :cool:. I'm limited in my controls knowledge but always trying to learn more. I recently started working at a...
Replies
14
Views
332
OkiePC
OkiePC
R
Ladder Creation Bad in Mitsubishi PLC
Dear All, I have an Mitsubishi PLC (FX1N-40MR) and PLC backup was taken and i compiled the program and its showing 0 error. then I tried to...
Replies
8
Views
266
Riyas Mohamed
R
J
PTC Kepware/ Linkmaster bad vartype
Hi, Have anyone came across issue with Kepware or Linkmaster not returning correct values (Unknown Quality) for some tags From OPC DA Client...
Replies
7
Views
650
Dale87
D
E
Micro820 analog output module 2080-OF2 bad?
I have run into a problem with my first time installing a 2080-OF2 module on a Micro820 and have a hunch that the module itself - brand new out of...
Replies
9
Views
913
willms
W
Back
Top Bottom