Create a New TARPN Network From Scratch
If you are interested in seeing a TARPN in your area, you might have to build one from scratch.
The cheapest first step ($160-ish) before you can see the software work is two Raspberry PIs and two NinoTNCs.
Using clipleads you can connect the NinoTNCs nose to nose and now you can join chat on TARPN-HOME from each station.
Note that you will need two different callsigns to run two chat stations.
While you are running without antennas, the callsigns can be made up.
K1QRM and W4QSY.
Know what it takes
This document discusses a measured/staged method of trying the TARPN recipe out, without building false hopes, and without investing huge money, before figuring out if it is possible for you and your community.
This document is about getting a TARPN established and having it grow.
Getting the chat-room system running at the start is a big part of this.
Promotion of your project to the other hams is another big part.
Your plan must include being successful in the long run or people won't put time or money into the project.
Packet radio, the basics
The communications scheme we're using puts digital data onto a frequency modulated radio transmission (FM) on VHF or UHF radios.
The devices that modulate the digital data are called TNCs, or Terminal Node Controllers.
The name means that the device is the end-point of a network and that's not exactly what we're using it for.
The TNC accepts data from a host computer and attaches details to the front and back of that data, and then transmits it all at once, activating the PTT to the radio, sending the packet
and then unkeying.
The receiving TNC will receive the entire packet, verify that it is fully intact, and then hand it off to its host computer.
The packet includes elements related to verification of the contents, and also telling the receiving TNC if the message is a procedural
, i.e. about the communications, verses data intended for the receiving program, and also some addressing information so the network stations can route the information appropriately.
The rate of data transmission while the radio is keyed up is specified (identically) by the TNCs on both ends, and is between 1200 bits per second and 9600 bits per second.
The amount of data sent from one application over the radio to another application can widely vary, but is usually maxes out at around 25 letters per second at 1200 baud or about 200 letters per second at 9600 baud.
There are architectural details which can make this considerably slower, depending on how many stations are sharing a channel and if the stations can adaquately hear each other.
Building a Network
If your goal is to establish connections to specific target locations, the TARPN architecture may provide you a means to that end.
However, there are topological limitations inherent in VHF/UHF radio, and the modems used for 1200 to 9600 baud are a bit needy in terms of signal to noise ratio.
If not all of the stations are within simplex range and there is no way to string together the volunteer participants, then the project can get stuck.
It's not obvious to hams how to put together a network node to provide reasonable performance.
On the TARPN web pages I'll try to provide some clarity in the issues.
FM transceiver links used for 1200 to 9600 baud data generally must have a path loss of under 150dB (success with reasonable antennas calls for 138dB path loss) and a fade margin of about 20dB after including the antenna performance.
Note that 6m link may require lower path loss to overcome atmospheric and man-made noise.
Note that 440 links can be supported with high gain yagis so higher path loss may be supportable.
150dB is probably a decent limit to look at assuming you have really good antenna mounting for yagis.
You can use tools like Radio Mobile Online to provide this information for a given site and antenna location.
Using FM transceivers, you'll want to have on the order of 20dB quieting (fade margin on RADIO MOBILE), on every transmission, for proper operation.
The range of a pair of 2meter 80 watt radios and large 20 foot high omnidirectional antennas on flat-land with dry or no trees is expected to be about 25 miles.
In the mountains with line of site, 60 to 100 miles is possible.
In rolling hills with trees, the range can be as little as 2 miles or as much as 20.
With indoor antennas and rolling hills, it could be as lttle as 1 mile.
With a TARPN architecture, you are usually not talking about house to repeater site, but house to house.
That's very different.
If you can't get a network strung together with the existing volunteers, you may have to try recruiting.
Depending on what is more important, having a TARPN, vs having a specific limited list of participants, you may have tough choices.
If your participant list can be strung together within the physical limits imposed by VHF/UHF radio and your topography, then you win.
The other option is to attract random-hams to the project and focus on building a local network with the available local hams, and then work on expanding to bring in the desired longer-distance hams.
Making it attractive enough to grow
The best application for attracting hams to a new TARPN is going to be the chat-room in TARPN-HOME.
Most VHF ham packet radio systems cannot handle chat with more than a few participants.
The collisions kill them.
A TARPN, being collision free, can cope quite well with chat, especially as the chat applications are installed in every TARPN node.
As such, the first thing you should show any new potential participant is going to be the chat.
Chat is also attractive, useful, and good for retention in the hobby.
A key method for getting the attention of local hams such that they will be receptive to seeing a demo, is to talk about the project on the local airwaves.
The best conversation for a potential participant to hear isn't going to be bragging about how great it is.
What will perk their interest is your attempts to get it working.
Hams love to figure out how they can help a given situation.
They are much less interested in systems where they will always be a follower.
My point here is that you should use local communications methods (local repeater?) to discuss solving the problems.
You don't have to hide out on telephone lines, secretly getting your system ready for show.
Bring it out in the open.
You're building something cool and technical.
This is a good thing!
Old packet wasn't so good.
You'll run into an attitude of "we did this in the 90s and it was terrible".
There was an epoch of general-purpose packet radio
where hams used common channels for many stations attempting to do mixed use text delivery.
Many stations sharing a channel without a serious attempt to manage the available channel capacity vs agressiveness of the stations led to frequent communications failures.
The TARPN plan solves that problem, but there is much history here.
Calling what we are building "packet radio" is not going to be a win.
Call it an "off-the-grid chat-room using ham radio" will get better result.
We're USING packet radio to build it, but we're not building a 1990s mess.
Younger hams, and new hams, may be more receptive to what we're doing, but there are some older hams who have a clue and who will like what they see.
Just beware the nay-sayer.
See the articles in the FAQ section.
There is much good stuff there.
Beware trying to start a TARPN when there is an established general-purpose Internet connected packet network.
That's really hard.
Why Off The Grid?
Stress the ham-radio-exclusive system and the live-chat as a purpose.
You can have both the old-system and new in the same town without interference.
The way we move Internet data to and from the TARPN network is by manual cut and paste.
No automated Internet gateways permitted.
A good reason for this rule is that a ham-radio-exclusive project is very
attractive to a large enough group that ignoring the Internet lovers is worth it.
When you have a working chat-room system, it is easy to impress people by pointing out that this is really a ham radio project with no cheating.
This flies in the face of the DMR/DSTAR/C4FM/EchoLink/IRLP projects.
Those last-mile-to-the-Internet projects which masquerade as real-ham-radio really bother some hams, and those hams may be your advocates or participants.
My suggestion for starting out is to build a demo system with two or three portable nodes, and don't start
building a permanent, always-up, network until you are ready for the build phase.
The very first project will be two Raspberry PIs and two NinoTNCs sitting on a desk with clip-leads connecting the TNCs.
No radios until you get that working.
Before you enter the build phase, collect, with or without help, enough hardware: 8 radios, 8 antennas, 8 NinoTNCs, 5 Raspberry PIs, power supplies, batteries and whatnot, to build five nodes at people's houses with dedicated point to point links as per the TARPN plan.
You'll have some multiport nodes and some single-port.
Distance is not a startup goal.
People are a startup goal.
More than five builder/operators is great.
Start-up may still take another two or three node sites NOT at participant houses (beg space from other hams along the way?), but you need five regular (nightly) participants to start.
Less will result in a flash-in-the-pan project with no life.
People who say they'll host your links but not get on the air do not count in your participant count.
The demo phase would start with two AC-powered (probably) portable nodes, one radio each, as a demo.
It takes two nodes.
One is a non-starter.
Don't try to bounce one node off of existing hardware.
It won't be impressive.
You'll want to have TARPN-HOME fully working and configured on your two demo nodes before you show them.
I recommend using 1/4 wave ground plane antennas on your demos with 10' of coax between the antenna and the radios.
Try to get radios with a 3watt or less low-power mode for the demo.
Icom IC2AT and IC4AT radios are perfect but they don't look like the radios most people will use.
Maybe one of the radios should be a real 25w mobile radio set in low power.
Demo Two, maybe
Having three nodes, one with two ports, using 440 for one link, and 2m for the other, is useful for showing the advantages of having separate link frequencies for each hop.
It becomes really obvious how well the network scales.
A setup with three demo nodes really wants some cabinetry for the purpose.
Electro-mechanical failures are common when the individual elements of a node are in motion relative to each other during transport.
You'll need to really do a good job with getting the 2-port node to start up reliably.
I strongly recommend mounting each node into a cabinet with an eye toward surviving being moved.
A howitzer ammo box is perfect. So might a go-kit rack box.
I'm partial to the IKEA KNAGGLIG wooden boxes
Make sure your radios come up with the same settings (volume, frequency) you left them at when you last demoed the node.
Three demo nodes is really cool if observers can listen to both channels on monitor receivers.
This is blazingly fast compared to channel sharing.
Even if you can't take it portable, doing three stations on the bench is a great learning experience.
You'll have confidence for what comes next.
I strongly recommend against deploying the demo hardware for more than overnight until you get five stations.
Three is NOT going to help you make a sustainable network. You need five.
Make sure you look for links which will be FULL QUIETING between the sites you are going to use for the network.
You may test with HTs over the path just to get some signal, knowing that you're going to add real antennas and 25watts to complete the link.
VHF and UHF yagis are excellent.
Even a small yagi is pretty terrific compared to an omni antenna.
By this point you've read everything on the TARPN site, especially about debugging poor links and about the various servers and features of the node.
Read the FAQs and build instructions about shopping and cabinetry and so-on.
It is quite possible to run this system with few or no experts in Linux but you'll definitely want experience starting and stopping the services, doing updates, configuring the links, testing audio levels with a scope, adjusting the transmit level using an HT held up to the ear.
It's not hard, but you will want to get familiar with the details.
Assemble as much of the hardware as you can at a single point, on semi-portable cabinets, or even screwed/tied‑down/Velcroed on wooden shelves.
Test as much as you can with the gear in one place, before deploying.
Use dummy loads or low power as needed.
Once you get five participants on-line and chatting nightly, expansion over distance will become important in order to attract new participants, but you don't need that to get started.
You need five on-the-network people to start.
That's a tough barrier to entry but you don't need to spend $$ until you attract and get commitments from those five builder/participants (four plus yourself?).
Once the builder/participants are signed on, move quickly so people don't get bored of the idea.
The more people you have the merrier, but get five up ASAP after you start this phase.
I don't know why, but three people on packet just doesn't create a growing network.
Part of this may be the way the several participants discuss the project on other ham channels.
What bands work for packet?
Each of the bands has their own interesting characteristics for point to point linking.
10m to a large extent, and 6m to a lesser extent, change performance during the day.
The range you actually get varies.
I don't think we're allowed to do greater than 1200baud on 10m.
You are trying to make full-time-up 24/7 radio links.
If you don't have that, then your network can't grow.
One of the cool things about TARPN is everybody is invited to expand the network because everybody is on the backbone.
Make sure you start out that way, and continue to treat your network that way.
In NC we use 6m successfully but the links aren't much longer than you'd get with 2m.
Two meters is the best and for point to point links but with caveats!
The two meter band
- is the most crowded band in the VHF spectrum.
- is susceptable to jamming by modern electronics and consumer integrated power supplies used in home electronics and light bulbs.
- is susceptible to temperature inversion range variation.
- Yagis are not particularly cheap and barely work at all indoors.
- When installed around or in foliage, the foliage needs to be 4 or 5 feet from a colinear-type antenna to preserve gain
It is possible to have two links from a house on 2m if you separate the frequencies by three megahertz, and separate the antennas, or use PASS-CAN cavity filters.
You can get four UHF links at the same house.
The same rules apply (as in the 2 meter band, but use six megahertz separation. 425, 433, 440, 446
UHF is much less susceptable to interference.
The antennas are smaller and can fit in attics.
The distance to foliage can be a foot or so and still have the antenna operate properly.
Chinese UHF yagis (Ailunce) are rational and inexpensive.
Indoor Antennas and Antennas on buildings
Antennas on 50 and 144mhz will work best when they are 20 feet or more from a house.
While 440 and 900 mhz antennas can be indoors, 50 and 144mhz antennas will pick up switching power supplies, LED lightbulbs, and small computers, causing the noise floor to be so high as to make only a very short network link operable.
If you are using indoor antennas or roof-mounted antennas, consider using only 440mhz and 900mhz systems.
All things being equal, 144mhz will go further, but that won't help you if your receiver can't hear the other end of the link.
Tree-hung antennas are the most likely solution for 144 and 50mhz where stealth or invisible mounting structures are required.
Masts and towers are ideal, but not everybody can have those.
6m antennas work in trees, but be careful about funny tech antennas on 6m and 2m around trees.
1/4 wave groundplanes and dipoles are pretty reliable, but collinear antennas and antennas without ground-planes or antennas without a balanced counter-poise are messed up by foliage within a wavelength.
6m beams are mighty, but somewhat expensive.
We have 3 MOXON antennas, 2 of which are made by PAR, in the network on 6m with good results.
Beware of unbalanced antennas.
J-poles are inexpensive but they can have unbalanced behavior including sensitivity to coax length, common-mode-current which causes RFI in your hamshack.
VHF RF getting back into your hamshack can cause unpredicable behavior on small digital circuits.
It can also make it difficult to use multiple frequencies simultaneously.
See also TARPN builder: Antennas
Some sites may be able to get away with low antennas and low power, especially if the participants are in the same neighborhood!
I can help you get the radios! It will take maybe $400 total investment for 8 radios on VHF and UHF to get your first system going.
But, to save $ I'm sure you can scavenge 25watt single-band ham rigs too.
Mobile stations don't have much use for monoband 2m rigs without control head support and modern FM operators like multi-band radios.
This leaves many decent monoband radios unused and available for packet operations.
Icom IC28, 228, Kenwood TM7850, Tm7950, TM221, 231, 241, 261, 281. Yeasu FT211, FT2500.
I don't know Yeasu model #s as well as Icom and Kenwood.
OLD HTs (without CPU control) work, Icom IC2AT, IC3AT, IC4AT, Kenwood TH21, 31, 41.
Prior to 1983 ham radios with CPUs are not great for packet. Yeasu Memorizer, Icom IC25, IC27.
Not so good.
These were CPU controlled but the CPUs used 5 V at high current for data switching.
The Icom 22S, on the other hand, was pretty good and is first generation synthesized, but has no CPU.
Crystal ham radios work but the crystals are way expensive.
HTs after microprocessors came out are not so good.
They tend to be too slow at going in and out of receive and transmit.
Some real modern ones have audio processing that make packets unreliable.
Modern HTs tend to be a cross section of over-feature-rich, too expensive, or very badly made.
We can get commercial surplus 2-way mobile radios, often for under $50, so don't bother with Chinese brand stuff and especially don't bother with expensive modern "packet ready" ham radio equipment.
"packet ready" meant there was a connector or menu option.
It doesn't mean the manufacturer paid attention to everything.
Actually-has-packet is much better, but too expensive for a permanent link.
A packet equipped ham radio could make great test equipment because of their front panel VFO control, power control, S-meter and simultaneous packet reception and speaker operation.
People will learn from your demos and make choices based on what you have working.
They'll want to see how-to-build a network and you are the expert.
Using reproducable examples may be the best thing, or at least something that looks like what you'd use for the next node.
Learning how to use Commercial Surplus Radios
I can help you with tools and procedures for many Kenwood and a few Vertex commercial computer-programmed radios.
But what's best is you find somebody who knows some surplus radio well and will take care of getting them and setting them up for you.
Don't try to get everybody on the project to learn all of the aspects.
It's just too much to manage.
Spread the work out.
Make sure everybody knows your goal to make an Internet-Free network.
All chat, ham to ham traffic, and any real-time delivery of the network traffic MUST be via ham radio.
If you give in to temporary (or permanent) Internet links you are kicking down anybody who did the real radio work to make this go.
It lots of fun to get on a live chatroom built entirely from ham gear and everybody owning the infrastructure.
TARPN-HOME is fabulous too.
If somebody wants to see your network from a distance, use a smart-phone and make them a movie.
Apple Mac computers know how to make a movie from their own screen.
See if one of your cadre has a modern Mac available.
Learn the tools
You'll want to master CHAT, the 7777 web pages, QT-term for monitoring traffic, the R R command from both ends of a link, linktest, TINFO, TRR, flash, and the log files.
Mess around with the TXLEVEL controls and make sure you can quickly tune up a new setup.
Figure out how to tell the TXDELAY is set optimally.
If you made it this far
in this dry and boring page, I applaud you. Please make contact on firstname.lastname@example.org or look me up in QRZ.
-- Tadd KA2DEW