Radio problems with remote site


Thread Starter

Jeff Cobb

Hey all, i am having trouble communicating with a remote site, and I'm about out of ideas for how to fix it. Hopefully yall can give me some suggestions.

A little background, I work for a system integrator in TX (what accent? i dont have an accent.). We do turn-key systems, and then a large part of my job is servicing those systems if/when they have problems down the road. In this case, we are dealing with a water plant my predecessors put in almost 10 years ago. They pump out of a river appx 2.3 miles away from the plant. The communication link worked fine for about 10 years, but about a month ago, they started having communication problems. It started out where they would only have problems a couple of hours a day. Then they'd have comm fails all day, but it would work at night. Now they have a comm fail all day and all night.

I'll try to just throw as much info at yall as i can think of and see if any of it sparks an idea.
The System:
-There are about 10 sites around town. All of them except the river station have 99%+ valid polls.
-The radios are MDS 9300s (and probably some 9310s)
-they use the 900 MHz unliscensed spread spectrum band. frequency hoppers and all that jazz.
-The polling is done from a computer running Lookout SCADA software
-The RTUs are TI435 (which was bought out by siemans, which was bought out by PLC direct, but basically its the koyo processor and we can get replacement parts from
-We are using modbus protocol and the radio is connected to a MB slave card in the PLC rack. (f4slv-mb is the new part#)

I have double checked the alignment of the antenna, it was off by about 10 degrees, but repointing it didn't help.

I am pretty familiar with MDS radios, esp the 9800s, but a lot of my favorite commands are not available with the 9300s. (like rssi, zone data, snr).

the lights on the front of the radio indicate it is seeing a good carrier signal from the master, and it is recieving and sending data. the MB slv card is also recieving and sending data on a regular basis, but back at the plant, they show no responses.

i have already tried replacing both the radio and the MB card.

I checked for reflected power (to see if the antenna installation was OK), and I had rather interesting results. First of all, it was only putting out .5 watts (BTW, the spare radio was set up the same way) and second, the needle did not even twitch when i switched it to reverse power. Granted, my meter's lowest setting is 5 watts (dont get me going on this one, i told my boss we wanted the MAX range to be 5 watts, cause we work on low power stuff, but instead i got one that goes up to like 500 watts! WTF, i aint touching a 500 watt system with someone elses 10 ft wooden pole!), but i've never seen one that didn't even make the needle twitch. Theoretically, i guess this means i have a perfectly tuned antenna, but i have trouble beleiving this to be the case on a 10 yr old installation, and frankly, my predecessors who put this up knew diddly about radio systems back then. I forgot to check, but I'd be surprised if the connectors outside the building were even wrapped. I went ahead and adjusted r69 on the radio until it was putting out a watt (actually i left it at about .9 watts just to be on the safe side) So, question 1 for any radio gurus out there, if you get zilch on the rev power, does that mean your antenna is ideal? Or are there other possibilities?

>From watching the lights, i think the RTU is seeing the poll and answering, but the master is just not seeing its response. I have honestly never had a communication link that worked one way and totally crapped out the other way. There were a few times when i *thought* this was happening, but it always turned out to be something else. Like a bad transmitter, or the setup was wrong. I was really hoping that turning the power up would do some good, but it did not. has anyone else seen something like this?

Right now, I am leaning towards saying the problem is interference, but without a service monitor of any kind, i dont know a good way to test this theory. And saying that its intereference does not fix it. There is another plant within sight of the River PS, and they have a big omni antenna, so i am planning on contacting them and at least finding out what kind of telemetry system they have and how long its been up. Maybe I'll get lucky and they will have a bonafide guru around with a service monitor and everything.

thanks for the help
-jeff (i hope there's no char limit on posts)

A few suggestions--

1. I would first verify there is good serial communications with the remote and master, even if it meant taking the remote plc and hooking it up directly to the polling master. You never know if a setting in software got changed that's preventing succeful comms.

2. Try polling the remote plc through the radio link with a polling terminal program for modbus on your laptop like Wintech's. You can see received data stream and check the response for exceptions, extra bytes sneaking in due to noise or bad rts-cts timing, or clipped bytes due to bad rts-cts timing or faulty radio keying.

3. Investigate your antenna system at the remote and master site, coax can deteriorate and just absorb the rf, not even showing reflected power. Also check all connections for corrosion/water. (This is especially suspicious due to the way you gradually lost comms).

4. If all this fails, you may need to find someone with a spectrum analyzer to see if you are being clobbered from another radio system in the same vicinity.

I hope some of this is useful, and you find your solution soon.

Any electric fences in that direction? Is it worse when it is wet out? Get yourself a field strength meter and see if there is a pulsing of broad bandwidth of signal at about one second intervals. If so you can get a small directional antenna, of any kind (like an old TV yaggi) and use the meter to find a fence problem. It's usually a weed on the wire or a broken insulator causing a ground loop. Sometimes it's hard to get the owner to fix the problem or allow you to do it. You did say Texas, right?


Peter Whalley

Hi Jeff,

A very long, very lossy cable will make even the worst antenna look good and water in the cable could be increasing the losses substantially. Try
checking the VSWR at the antenna end. It'll tell you both the cable loss as well as the antenna VSWR without the cable effect.

Given the time effect on system performance, interference sounds like a very likely cause.

From what I've heard, unlicensed spread sectrum systems like yours have already become unusable in many urban areas because of the levels of
interference between them.

Either change the radio systems to a less conjested band (maybe licensed) or try a different antenna. Sometimes it's better to align the antenna so that a null in it's pattern points at the interference source and not worry so much about pointing the antenna directly at your own remote site. Remember that nulls are very sharp but the actual beam will be fairly broad for most antennas. To make this work you may need to get hold of the polar patterns for a number of antennas and select one with a suitable pattern.


Peter Whalley
Magenta Communications Pty Ltd
Melbourne, VIC, Australia
e-mail: peter*no-spam*
delete *no-spam* before sending

Alan Rimmington

We had problems with a licence free system between two pumping stations. Installed fine but would occasionally crap out for hours at a time (usually around set times of day). We had the experienced radio people back loads of time and everytime it was all working as sweet as a nut!

In the end it was put down to other band users (in uk radio hams, personnel walkie talkies etc use the same frequencies), we then had a choice licenced radios, or leased line, client decided on leased line.

Fred Townsend

It is probable the remote site antenna or feed line is bad. This can happen when some bird hunter shotguns the feed line or with old age. Old age happens a lot faster if you are in the Gulf Coast area of Texas ( I didn't hear an accent either). Most radio men can do a standing wave test (VSWR) test. If this doesn't reveal the problem, come back and we can talk about more esoteric suspects.

Fred Townsend
I had a similiar problem before and the Base Station antenna was the problem.

What kind of surge protection do you have? Did you change it out?

Do you have a jumper cable from the surge protection to the radio? Did you change it out?

Usually when this happens- it is process of elimination.

MDS has a command called SEND (1-10) I usually use SEND 4. This will send several packets of data at once. Put this in all of the radios.

These are some suggestions- hopefully it might fix your problem.
Had one of those problems myself recently, even took a protocal analyzer out to the site to verify that, yes the signal AND data get here and even leave, but they never got back to the master station. In my case it was a paging system encroaching on my recieve freq (licenced)and it took a spectrum analyzer to find that. It sounds like you've got interference, but it's hard to look for signal interference on a spread. Chasing the neigbor for info is a good start. Did you try skipping zones to see if that helped? If someones not overpowering you maybe you could change your network ID number. Good luck.
Are you saying that the whole systems down or just the one site?
Just a few thoughts, if its the whole system, start with the master, after ten years (I'm guessing regular maintenance wasn't performed) its time to recallibrate. If its just the one site I think its your antenna. You should get a 5:1 ratio on your reflected power. Either way, its been ten years.... sell them a new system, and a maintenance contract.
Ok. had a lot of suggestions, and a few questions, So i'll try to address them all here. thanks for the suggestions, even the thug, who knows me and emailed me something that didn't make sense.

1. the suggestion about taking the RTU back to the master and using a direct cable ... yeah i know its the right thing to do, but I'll put that on the list as a last resort.

2. Electric fences, its hard to say, there is a lot of livestock in the area, i dont think electric fences are common on ranches are they? And i cant say if it is any worse in the rain (I mean its dead now, and it didn't rain at all while it was going out). However, i think it was raining a day or 2 before the initial problems were reported.

3. A couple of people mentioned the cable. Am i understanding you right, a really bad cable could absorb the power and show no reflected power? As i said, the needle on my wattmeter did not even twitch. Would that affect communications both ways or just transmitting? BTW, I am not positive, but the cable looked like RG8. My best guess is that its 30-50 ft long. i will try checking the power from the antenna next time im out there. I also thought trying it using a dummy load might tell me something. But it sounds like it would be worth it to try throwing a new antenna and cable at it anyway.

4. Surge Protection. yeah, i am almost positive we have a lightning arrestor, and no i did not change it out. Next time I'm there, i will try bypassing it. and i did not try changing out any cable, not even the jumper to the LA.

5. MDS commands. I am familiar with MDS radios, these do not have the zone skipping options (at least not that I've been able to find). I've never used the send cmd, but i did not see it in the manual, so it may not be available. What does send4 do? is it a diagnostics tool, or is it a permanent setting?

6. Its just 1 site thats down. And no, there's been no regular maintenance.

thanks again. Let me know if yall think of anything else.


Peter Whalley

Hi Jeff,

Yes lossy cable lowers your reflected power level because it reduces the power going out to the antenna and then further reduces the reflected power on the way back. A cable with say 10 dB loss will show a 20 dB return loss even with an open circuit on the end (that's VSWR of about 1.2:1 if memory serves) which is more than good enough for most people.

Peter Whalley
Magenta Communications Pty Ltd
Melbourne, VIC, Australia
e-mail: peter*no-spam*
delete *no-spam* before sending
first, thanks for all the help.

This is just a follow-up for anyone still checking this thread. We figured out the problem, it turned out to be a bad cable.

It kinda hacks me off though, because I was going to check the cable first thing next time i went out there. But someone else with a service monitor got out there first and within 5 minutes saw it wasn't transmitting any power. So they checked the connector at the antenna, and it pulled right off. The customer told me they jammed it back in real hard and the station started working again. I mean, its not fair, he had the right equipment. Thats like cheating.

But, its a piece of equipment I've been asking my company to get me for a long time. So far, I've only been able to get them to invest in the aforementoned watt-meter. So anyway, after taking a black eye like i did on this problem, I've decided its time to stop whining and take some pride in my work. I'm going to have to bite the bullet and pony up the cash for a service monitor myself. I'll post another question here in a little while looking for advice on what to look for and where to look for it.

thaks again

I am an expert at radio communications, I have done a lot of setup and testing of the MDS radio line, do you need good help or just some one to give you pointers. If you need real help, contact me.

Things you shold know,

the MDS9300 is a spread spectrum radio at 1watt of power, so first of all check your feed line and connectors. INterfernece is not likely unless it is high powered paging system in the 900 MHz range. You can talk to the MDS radios witha laptop using hyperterminal, contact MDS directly to get some lit on the radios as far as commands. Buy a wattmeter w/ slugs for UHF up to 1000Mhz at 5 watt and 25 watt, get ome slugs in the VHF range 100 to 250Mhz 5 watt and 25watt as well. you can rent a decent servcie monitor from test equity, if you buy one, plan on spending 12K to 20K. IFR1220S is a great unit, rugged, does it all, except the 2.4Ghz stuff and most SCADA systems that I know of are not operating above 900MHZ bands. (so don't spend the money for a unit with features up to 5 GhZ or higher).

YOu can key the radio up by sending KEY command and unkey by sending DKEY command. the 9300's should all be setup for 9600baud N81 (better at lower rates like 4800 buad). So issues like connectors, feed line and anteaas are good to look at.

If you need help let me know.

Matt hyatt
[email protected]
technical consultants