Long Distance Transcievers

These guys can make the link Microhard and if you need LOS with highly directional antennas you could use electronic compasses to determine direction of each radio and point the antennas correctly. Don't know if in your budget but they know what they talk about, they have made the links for a lot of UAVs at far greater distances at higher data rates.

Yes the higher the gain the greater the directionality, a bunch of ways to calculate it but the easiest way to understand is conservation of energy/power.

Build_and_Break:
Hello all!

I've been working on a new project lately that involves sending data from the rear of a train all the way to the front. Obviously there's a lot of metal and other materials between the two points, and some trains have been said to be up to 3.5 miles long. This poses a problem as there's a lot of interference for standard wireless transmission and it's just too far to transmit data by wire. Satellite/SIM communication isn't a great idea since there is virtually no carrier that covers all of the USA and I don't want to constantly pay for data. I'm not sending very much data, just a live readout of a few variables (can't say) to the front from the back.

Do you now, or have you ever worked for a railroad? Or been around railroads at all? Telemetery and radio control is transmitted from front engines to back engines and sometimes mid engines. Trains also go through tunnels and the front of the train may be on the other side of a mountain from the back.

Railroads have whole divisions charged with communications on board trains and from train to control centers. When communications may be questionable, they use repeaters. Railroads have many, many frequencies available to them for communications.

Also, you don't want to interfere with normal railroad communications. What will you do with multiple trains on parallel tracks, or one behind the other, or one on a siding while the other passes it?

I am concerned that you have not communicated with the railroad communications people.

Paul

I've been pondering ways to provide real time telemetry and video from the N&W 611 engine (recently refurbished after 20 years in a museum) to the passenger cars on excursions. I think one way might be bridged wireless routers atop the engine and passenger cars such that there would always be a LOS connection between adjacent routers along the length of the train.

35 tons of coal
22,000 gallons of water
872,600 pounds of engine and tender weight

zoomkat:
I've been pondering ways to provide real time telemetry and video from the N&W 611 engine (recently refurbished after 20 years in a museum) to the passenger cars on excursions. I think one way might be bridged wireless routers atop the engine and passenger cars such that there would always be a LOS connection between adjacent routers along the length of the train.

35 tons of coal
22,000 gallons of water
872,600 pounds of engine and tender weight

https://www.youtube.com/watch?v=US-TfA6e3F8

This should have been you first post! Now we know what you are working with and towards. Sounds like a neat idea. Since all the cars and the engine have electrical power, have you considered coax cable between each? Sure, one more thing to connect/disconnect, but they already have a bunch of stuff to deal with.

Each car would have it's own video monitor. The project then becomes how to use an Ardiono at the engine end to generate the picture and send it on a coax cable.

If you use any RF generation on the train, check with the people contracting for time on the commercial RR tracks for any statements about radio use.

Paul

@PaulS and @el_supremo Thanks for the clarification. Obviously there isn't much of a direct line of sight available in my case.

@Boardburner2 unfortunately there is seemingly no carrier that covers all railroads, at least none that are reasonably priced or are available for use still.

@SurferTim thanks for the link and tips. I'll definitely look into Linx.

@wwbrown Thanks! As mentioned earlier though, I will likely actually need an omni-directional antenna as the recieving end could move around (significantly).

@Paul_KD7HB The readings being transmitter must be read from the back anyway, so I cannot transmit from the front. The only time anything is being transmitted is while the train is at a station and is stopped completely. Repeaters don't make a lot of sense because then every time the train stops, someone has to place repeaters along the train. To answer your question, no, I haven't personally done any work with a railroad or train systems. I've already got a solution for multiple nearby trains, esp. because that'd be a problem when it's at a station.

Sorry for the delay in my reply, my desktop was offline for a while with Windows 10 problems :frowning:

I've considered using transceivers with wired antennas that can by mounted by magnet to the exterior of the train, which doesn't have to be too strong because it won't be mounted when the train is moving.

I'll also talk with a few friends who are in the railroad industry to see how plausible it is to wire it in to the existing communication system on the train instead.

Thanks for all the replies, I'm still open to suggestions and feedback.

Thanks again and regards,
Chris

I have access to the user manual for the Linx stuff now. No library is really needed. It is a direct replacement for wired TTL serial communication. Only a couple of settings need to be changed, and that can be done from a PC prior to use.

Build_and_Break:
@wwbrown Thanks! As mentioned earlier though, I will likely actually need an omni-directional antenna as the recieving end could move around (significantly).

That is why I had the following in my original reply
"and if you need LOS with highly directional antennas you could use electronic compasses to determine direction of each radio and point the antennas correctly."

Higher gain on the antennae is almost always a better choice than higher power on the transmitter, the antenna pointer could be setup very easily and each unit could send what direction they are pointing say every 15 seconds and then a quick calculation on each end would tell you what angle your antenna should point. A small servo could do the hard work, no problem these things are done in the UAV world all the time and we have to do it in 3D here you are in 2D pretty much.

Still don't want directional antennae just crank the power up to its max, the guys at microhard will be able to tell what is needed.

Might be application of A wireless mesh network?

A wireless mesh network (WMN) is a communications network made up of radio nodes organized in a mesh topology.

Put a group of nodes between front and end, let them self form and self heal.

http://forum.arduino.cc/index.php?topic=217078.0

@sonnyyu I've considered that, but these radio systems have to be removed and placed quickly, which makes having a lot of them to pickup and place and turn on and off inconvenient. Thanks for the suggestion though!

Key issue:self form and self heal. any given time few node turn off cause no problem, neighbours will be cover up.

One more thing It take place much higher layer than physics layer and could use lower frequency for longer distance.

wwbrown:
...

Still don't want directional antennae just crank the power up to its max, the guys at microhard will be able to tell what is needed.

If I recall it right, Max power will be 1W/1000mW, 30dBm at North American ( FCC rule)

@sonnyyu But in my situation, it isn't as cost effective as I can get this (https://www.sparkfun.com/products/9099) for about $60. Although I will definitely offer repeaters to purchasing railroads for better stability and signal strength like you suggested.

Build_and_Break:
@sonnyyu But in my situation, it isn't as cost effective as I can get this (https://www.sparkfun.com/products/9099) for about $60. Although I will definitely offer repeaters to purchasing railroads for better stability and signal strength like you suggested.

keyword is: Up to 6 miles (10 km) RF LOS with high gain antennas

high gain antennas is directional , and required front and end at fixed position (directional).

Indoor/Urban range (w/ 2.1 dB dipole antenna): Up to 3000 feet (900 m)
Outdoor RF line-of-sight range (w/high gain antenna): Up to 40 miles (64 km)

2.1 dB dipole antenna is omni directional antenna, will give double distance at line-of-sight ~1800m.

@sonnyyu True, it is only directional, but I only need directional. The conductor starts at the front of the train and walks to the end on the side that the transmitter on the end is. He has a handheld receiver, and since he is only getting closer and walking straight to the transmitter at the end, he is not changing directions and the transmitter at the end is still pointing at him.

Trains also don't usually exceed 3 miles, so I'm willing to sacrifice a few miles in range for getting a lower gain omni-directional if needed, and the one you linked to is great for that, but it is nearly 4 times as expensive, and in my opinion, it's not worth the extra money. Thanks anyway.

If cost is not a problem you could try these units:

https://www.taitradio.com/__data/assets/pdf_file/0020/59510/Tait_Specifications_TM8105_v2.pdf

They are used by taxi companies, mines and railways here in Australia.

They cost about $500 each and transmit using a 3.3v serial RS232 connection from your micro. They need a good omnidirectional whip antenna that will set you back another $50-100. They come in standard form at a baud rate of 2400 but you can pay extra for an unlock key to push them up to 19200.

I have 7 of them out in the field and the oldest has been running continuously for eight years over distances up to 10 kilometres at a speed of 19200 kbps (It is broadcasting RTK GPS corrections). I have tested them out to 20 Kms LOS.

I use them at 150 Mhz which has good propagation through metal structures and over/around hills as line of sight is not necessary like higher frequencies. I don't know what frequencies you are allowed in your area or whether you have to pay for a licence at these power levels. The datasheet shows a range of optional frequency bands and power outputs.

The unit shown in Post #22 looks like the one that Sparkfun used in their high altitude balloon experiment. They indicated that they would not use that module next time.

If the cost and licensing of the Tait is too much you could have a look at the Freewave offerings:
http://www.freewave.com/products/frequency/unlicensed.aspx

But a lot of these data modems use the 900Mhz frequency. You should try to get lower frequencies that propagate better through steel structures and trees as well as around hills. (things you may encounter going around a bend).

A low cost one that would certainly be worth trying would be the RFM23BP:
http://www.hoperf.com/rf/fsk_module/RFM23BP.htm
You can buy in quantity from Hope but for small orders get from:
http://www.anarduino.com/details.jsp?pid=130
(Another you may want to look at is:Anarduino and HopeRF Products)

At -120dBm sensitivity and 1 watt output at 433 Mhz through a decent antenna you would be getting many miles range out of this unit and at a more obstacle friendly frequency.
(Arduino library for module: RadioHead: RadioHead Packet Radio library for embedded microprocessors)

Update: So I've done some more research on SparkFun's database based on this (click) and I've narrowed it down to 3 products:

XBee Pro 63mW RPSMA - Series 2B (ZigBee Mesh)
XBee Pro 60mW PCB Antenna - Series 1 (802.15.4)
XBee Pro 900 RPSMA

Before you point out that only the third link meets the range I wanted, I talked with the people who requested this and said that the lowest range they'll work with is half a mile, so I'll begin testing with a less expensive mile one, perhaps, like the first and second links.

The first and third products are RPSMA type antennas, which would be externally mounted as I understand. However, the second link is a PCB Antenna, aka Trace Antenna. Apparently they perform about the same as wire antennas, but looks like a chip antenna. Can I still have a ABS or PLA plastic enclosure with a thickness of no more than 1/2" and still have a signal reaching at least 3/4 of a mile? Or can there be no obstruction on this type of antenna? Since it's built in, is it essentially omnidirectional? I understand that means it also outputs less power, and second link seems to hold true to that.

I also see that the first 2 links are not 900 series and are a little different to program, I assume I transmit and recieve with RX and TX and need VCC and GND hooked up as a minimal circuit, right? The SF website says you need to configure them, what do they mean by that and how do I do that?

@lemming Thanks for the link! $500 sure is a lot but 10-20kms in an urban area is pretty good! I suppose I won't be able to be paid for this and say I just wanted to test it, but I'll offer it to railroads as an option. Especially if they're still running strong after 10 years. Keep in mind that SparkFun has a 40 mile omni-directional for about $200, sonnyyu linked to it a bit ago.

Thanks and regards,
Chris

Thanks!

Chris Kuhn

$500 sure is a lot ...

When you said:

I've got virtually no price limit here,...

I offered the best, not cheapest option.

The SF website says you need to configure them, what do they mean by that and how do I do that?

Read the manual. There are a few of ways to set up these units. You can change the many settings through AT commands or use the X-CTU software.

Can I still have a ABS or PLA plastic enclosure ..

This will have little effect on the transmission. The built in antenna are loosely omnidirectional but are affected by their surroundings. Signal is better in some directions than others. Have a look at the hardware design specs to see how to set them up. There is usually a minimum distance suggested between the onboard antenna and PCB tracks and ground planes on your motherboard. You will really be pushing it to get these working in that environment so I would go for units with the SMA connector so that you can put a higher gain antenna on the units enclosure and not be affected by vagaries such as PCB traces. I haven't worked with the Digi Xbees despite their excellent software/firmware. I have used the Atmel (ex Meshnetics) Zigbee units because of their much better link budget specs. They might be worth a look with this testing application.

I re-iterate that you should be looking at lower frequencies and get your units on the roof of the train.

Keep in mind that SparkFun has a 40 mile omni-directional...

This would be with LOS with a large/pricey 11 dB high gain antenna.
I have worked with the X-stream units (essentially this unit that Digi package in a metal enclosure) on centre pivot and linear irrigators that are metal lattice structures several hundred metres long. While they are fine when on top, transmitting across the top of the ironwork, (with a $400, 1.7 metre 6dB antenna) they have trouble going through the lattice of metalwork.

I get the impression you would be using one of those 2dB antennas so you will have issues transmitting through a line of carriages.

@lemming By having no price limit, I don't mean I'm going to be likely to buy the most expensive possible option, but I said I'd still consider it and I have.

lemming:
... so I would go for units with the SMA connector so that you can put a higher gain antenna on the units enclosure...

As much as I'd like to do that, as soon as I found out that on-board antennas could reach up to a mile, I'd like to think that a product without any antenna visible or sticking out is more sellable, but nonetheless I'm sure PCB/Trace and chip antennas won't reach nearly as far as RPSMA, or even wire or u.FL. I'll look into the minimum space needed around the PCB antenna, as that could be my golden egg. As for the RPSMA ones I linked, I found a few antennas that look like they'll get the signal just about where ever I need it, provided there's a clear LOS. Thanks for the help!