WA8LMF Home Page | WA8LMF Resume | Main APRS Page | HF APRS Operation | Updated  08 January 2023

 

HF APRS Over VARA Modem

Jump To TEST SETUP (Farther Down This Page)

Jump To RESULTS (Farther Down This Page)

 

Introduction

The overwhelming majority of APRS activity has been sent over VHF FM radio networks using 1200-baud AFSK packet radio modems a.k.a. "TNCs".  For over 30 years a small percentage of APRS transmissions have been made over long-range HF radio from locations lacking 2-meter VHF activity.   HF-APRS has traditionally been done using AX.25 300-baud 200-hz shift FSK packet, mainly on the 30-meter band.

Classic packet works well in the relatively benign environment of VHF FM. It is a terrible mode for HF. It has very poor immunity to noise, selective fading and interference from other signals. It is an even worse performer in the one-way  "one-to-everyone"  broadcast beaconing of APRS (rather than the two-way hand-shaking ACK/NAKs error-correction of traditional "connected" packet). 
    If an APRS packet is lost at the receiver, due to a noise burst or interference, there is no provision to request a retransmission (ARQ).  The receiver just has to wait until the sending station decides to send another beacon minutes later.

For over two decades, a number of people have experimented with alternatives to AX.25 FSK packet for HF APRS activity.  G4HYG's "APRS Messenger" combined an APRS split-screen send/receive terminal, a serial port handler for an attached GPS device, and a soundcard modem using PSK63 or MFSK16 into a single application described here:  HF APRS over PSK63.
    The application was spectacularly more reliable on HF than classic AX.25 packet.  I would routinely carry on typed 2-way conversations with stations in the mid-west from my 100-watt HF mobile in Los Angeles or Las Vegas, even when the 30M band sounded dead.  I.e. literally nothing audible in the radio speaker.  When the developer abruptly decided to drop support for it, I experimented with using FLdigi on MFSK16 to do the same thing - see HF APRS over FLdigi modes.

 

A few years ago, yet another HF radio data transmission format burst upon the amateur radio community.

Developed in Spain by EA5HVK, "VARA" is a very sophisticated DSP data modem that runs on sound cards connected to radios with the typical "sound card interface". It comes in two versions: One adapted for HF over SSB transceivers, and a separate version adapted for the greater audio bandwidth of FM transceivers on VHF/UHF.  Both versions are adaptive - they will automatically adjust modulation formats & transmission speeds upward or downward depending on the quality of the radio path and radio bandwidths in use.

VARA is most-widely-used as an alternative to the expensive SCS Pactor hardware modems  for the Winlink radio email system on HF.  VARA is also used for the VarAC keyboard-to-keyboard chat program, as an alternative to FT-8.

 

Both versions achieve net data transmission rates comparable to expensive dedicated hardware modem devices such as the SCS Pactor modems, but entirely in software when properly interfaced to radios.
    For FM, this means using the dedicated data port connection of a radio normally used for "9600 Baud Packet" to achieve the highest speeds.  These charts of modulations vs speeds are taken from the VARA manuals.

Note that the free version of the VARA software only does the lower speeds. The "high speeds" are unlocked by registering the VARA modem for about USD $70. Once registered, you can use as many copies of both the HF and FM modem as you want, as long as they use the same callsign.

For HF APRS use, even the free unregistered version will work.

HF Speeds available in SSB bandwidth ("Standard") and 500 Hz CW bandwidth (Narrow).
Note that VARA can best a 1200 baud packet transmission on VHF-FM, using a 500 Hz CW filter on HF!

 

FM (VHF/UHF) Speeds
FM Wide requires data port input in 9600-baud packet mode.
FM Narrow
will work through mic/speaker jack like classic 1200 baud packet.

Note that APRS beacons contain very few characters.  Speed is not the important attribute of a modem for APRS.  Robustness in noise and at very low signal levels is!

 

 

Over The Air Tests on 60 Meters

Test Setup

I recently tested VARA as a replacement for AX.25 packet on 60 meters HF.  I wanted to take advantage of the NVIS (Near-Vertical-Incidence-Skywave) 0-300 mile / 0-500 Km propagation range on 60 meters.  This kind of close-in coverage doesn't occur on the much more widely used (for HF APRS) 30 meter band.  (30 meters normally has a 250-300 mile / 400-500 Km skip zone where nothing will be heard.)

I took advantage of a recent road trip from my home QTH near East Lansing, Michigan to a town west of Chicago, Illinois.  This is a maximum straight-line distance of about 200 miles / 360 KM.  On the road, it is about 250 miles / 400 Km.  [On 30 meters, I just barely begin to hear my own home-station beacon that transmits 24/7 when I arrive in the Chicago area.]  

The test setup consisted of two Yaesu FT-891 transceivers - one at home and one in the car.

The home station used a simple horizontal dipole made from two 60-meter Ham Stick mobile whips mounted about 12 feet / 3.6 meters above the ground - an excellent setup for high-angle NVIS receiving. The FT-891's DSP variable IF was set to a bandwidth of about 900 Hz to accommodate both the VARA narrow transmissions centered on 1500 Hz and the classic HF packet transmissions centered on 1700 Hz (1600/1800-Hz tones). 
    The receive audio from the radio's 6-pin MiniDIN "data port" was split into two Acer "netbook" mini laptops, each with a Behringer UCA-202 high-performance USB sound card.  On one laptop, the "UZ7HO Soundmodem" soft TNC decoded classic 300-baud HF packet. On the other, a copy of VARA-HF was running.
    On both machines, copies of UIview functioned as an APRS mapping display and Internet gateway, passing received data to the APRS Internet system for display on http://APRS.fi or http://findu.com .

The mobile setup consisted of an FT-891, a Panasonic CF-51 "Toughbook" laptop and a 60-meter Ham Stick mobile whip ball-mounted on the left-rear corner of the body of my VW Jetta TDI. 
    The AX.25 beacons were originated by a Byonics TinyTrack 3 connected to the mic jack of the FT-891. (The TT3 can be programmed to beacon using 300 baud HF tones instead of the more commonly-used 1200 baud VHF tones.)
    The VARA transmissions were generated by VARA and a copy of UIview running on the Toughbook, connected to the 6-pin mini-DIN data port of the FT-891 with a home-brew tone-activated soundcard interface.  Thus, both modes were transmitting alternately with the same radio and antenna.  Both modes were set to produce about 30-40 watts output from the FT-891 on USB.
    To distinguish between the two modes on the APRS mapping websites, the AX.25 beacons used the callsign WA8LMF-2, while the VARA beacons used WA8LMF-3 .

Results

The results were overwhelming!   VARA was at least an order-of-magnitude better than AX.25 in reliably transmitting data on the rather noisy 60-meter band. The maps below, captured from http://APRS.fi, tell the story.

 

The first pair were of tracks starting about 9:30AM EDT (local time) on Friday January 6th, as I drove from East Lansing, MI to Kalamazoo, MI. Kalamazoo is about 70 miles / 110 Km from the home receiving station.  For this segment of the trip, the entire drive was in full daylight propagation mode.

Note that the density of VARA position dots suddenly drops southwest of the Charlotte, MI area. UIview has an unconventional beaconing mode that is neither fixed-rate or "smart beconing" based on change of direction. Instead, it fires off beacons, based on change of position.
    Initially, I had UIview set to beacon every time my position changed by 1 Km, regardless of change of direction. At highway speeds this was resulting in nearly two beacons every minute.  I realized this was far too high for a shared channel like 60 meters.  I reduced the beaconing to once every 5 KM, resulting in the wider-spaced dots on the remainder of the trip.

AX.25 Beacons Heard - East Lansing-to-Kalamazoo - 9:30-11:30 AM EDT
The destination is not downtown Kalamazoo, but rather my sister's farm east of Kalamazoo.

 

VARA Beacons Heard - East Lansing-to-Kalamazoo - 9:30-11:30 AM EDT

 

 

The second set were captured from APRS.fi between 3:30 PM and 10:00 PM EST (local time) Saturday January 7th as I drove back from Westchester, Illinois to East Lansing, MI, with a couple of stops enroute. This time period covers the transition from daylight (relatively close-in NVIS) to nighttime (longer-range skip and loss of close-in NVIS)  propagation in this area.

AX.25 Beacons Heard  3:30 - 10:00 PM EDT Chicago-to-East Lansing
Note how the beacons fail to be heard for extended periods as night falls, and close-in (NVIS) signal strengths drop.

 

VARA Beacons Heard 3:30 - 10:00 PM EDT Chicago-to-East Lansing
As night-time propagation started weakening the NVIS signal, VARA continued to receive reliably long after AX.25 started failing. The dropout between Kalamazoo and Charlotte is when nighttime propagation made the band "go long" (i.e. a skip zone - loss of NVIS). The coverage resumes when I got into the local ground-wave range of the home station.