UBG v.2025 – Let’s Go North

Posted on by Jim Belford

What’s New and What We’re Learning….Slowly

In 2023 and 2024 we learned some things. After getting absolutely pummeled by a thunderstorm in ’23 we learned that Honda generators can get swamped and choke out if you pour enough rain water on them. In all of the years we’d never had one fail but Mother Nature is relentless and she won that round. The result was two of our crossbands went off-the-air until we could get out and get them restarted. That turned into a full-on four-wheel-drive event we didn’t want to repeat.

We also learned that with over 6,000 riders this race is too big for a single radio talk path. In the early years, the single net made sense; today there is too much radio traffic for a single frequency to handle. As the race has grown, so has the latency in getting traffic through the network. This had to change.

Race organizers had a new challenge for us beginning in 2024. They had created a new race, the Unbound XL, a 350 Mile unsupported bike ride that begins at 3P on Friday. Contestants bike throughout the night and throughout the following day before finishing back in Emporia Saturday evening (in most cases). This additional 15 hours of support meant several more refueling trips and a lot less sleep for our team.

Multiple Frequencies

We talk mostly about the 200 and now the 350 mile courses. These long courses made this race famous. However, we also support the 100, 50 and 25 mile races. As the popularity of the event has increased so has the amount of radio traffic. So much so, that significant delays were becoming common and Jeeps with inquiries were having to wait (sometimes) minutes for a clear radio channel.

In 2024 we broke the net control operation duties in two. We had one net for the 350 and 200 mile courses and a second net for the 100, 50 and 25 courses. Our technical team decided to continue using our local 146.985 repeater as the backbone for the longer courses and use our local UHF repeater for the second second net supporting the shorter courses. Having these nets on different bands allowed us to use a dual band antenna at net control and not have to worry about desense issues.

We then added a second Kenwood V71 rig, and second net control operator and operated the two nets simultaneously. The results were outstanding. With traffic rolling on two separate talk-paths, radio wait time was reduced significantly and response time improved.

Mike and Codey at AlmaW.

After studying coverage maps and some driving to determine soft coverage spots, we had to implement a drop crossband repeater at the far reaches of the 100 mile course to provide the necessary coverage across the Northern section of the course. We’ve learned how to do this and had spare equipment we could easily put into service.

Re-thinking Our Power Setup

Historically, we have used the Hondas as a 110VAC power source to power standard 12V power supplies. When this gig started years ago, we had the power supplies on hand and the solution was simple. The obvious problem with that setup was when the generator quit, so did the radio. We had to take the crossband nodes down to fuel the generators and with the longer race times, pulling crossband nodes offline 4 times throughout the race was not acceptable. The Honda EU2000 generators have a DC battery charger output which we had never used. So, we abandoned the 110VAC output in favor of new DC cables used to float charge small 12V 7Ah batteries. The battery was then used to provide stable 12V power for the radios. With this arrangement, we could take the generators offline to fuel, check oil etc. without interrupting radio service. The batteries would also give us a buffer if we had a generator die unexpectedly, run out of fuel or get swamped by a thunderstorm.

Remote Health Alerts

The addition of a battery backup had hardened our service considerably. We were no longer concerned with fueling outages or a generator failure causing the sudden and unexpected loss of a node. However, we needed a way to determine remotely if the generator was actually running so we could intervene before batteries died if we had a generator failure. We considered the addition of an audible alert tone at the end of radio transmissions. This ‘alert tone’ would only be active if the generator was offline and would give our technical team an audible ‘heads up’ of a problem long before it impacted the talk path. This option was ultimately vetoed in favor of a more robust solution.

Telemetry

Every year, during the event our technical teams ask the same questions:

  • “I wonder how much fuel is in the genny at crossband X?”
  • “I wonder what the air temperature is inside of our enclosures?”
  • “Are we moving enough air through our enclosures to keep our gear cool enough?”
  • etc.

We decided that if we were going to build a notification for the generator power, we would build it in such a way we could gather all types of data from our nodes to answer our questions and really be able to determine node health and status in near real time.

Voltage and current sensors are readibly available and inexpensive. Since we were redesigning our power setup anyway, adding voltage and current monitoring was included in the new design. The INA3221 breakout board has provisions to monitor 3 current paths. We used two of these boards at each node, one for monitoring battery and generator vitals and the second to monitor respective radio and fan data. This data was aggregated using a Raspberry Pi PICO and an I2C bus. Having I2C available, it was an easy task to add temperature monitoring inside of our enclosures using the AHT10 boards. Our team thought it was important to directly monitor the heatsink temperature of our rigs and that was done using an 18B20 IC epoxied to a paper binder we clipped onto the heatsink fins.

Early prototype of our RF monitoring hardware and temperature sensors.

Determining fuel level in the generators was a primary goal of our team. We were not interested in putting anything in the fuel tank and didn’t want to modify the generators extensively. We found a capacitively coupled liquid sensor we could stick on the side of the fuel tank using ‘Alien Tape.’ This sensor provided us a status voltage we could monitor with our PICO.

SWEET LORA

All of the information we could collect was great but pointless unless we could get that data to a central location where it could be monitored. Enter LORA.

LORA is a radio technology that provides reliable, low bandwidth data communications between nodes. In the US, this operates at 915Mhz with a max power level of 30 dBm and provides a throughput of around 300 baud and respective distances across a line of site flightpath. This was perfect for our needs as we were only going to squirt around 100 bytes of data 20 miles every 15 minutes or so. After more splat! point to point studies, We found a Heltec v3 LORA board and began testing.

Serendipitously we discovered Meshtastic about the same time we were determining options and testing LORA. The Meshtastic platform had already done all of the work needed to send our data payload between our nodes using the LORA protocol. Our Heltec boards were flashed with Meshtastic and we began ‘playing’ with the integration between our Heltec and PICO. Thanks to the Meshtastic community, we were soon able to get data from our PICO and send data between nodes on the test bench using LORA. Things were coming together.

LORA DISTANCE

The Heltec boards came with small, rubber low-gain antennas. Given the distances involved, we knew we wanted a more robust antenna solution. High gain vertical antenna can be purchased but are big and relatively expensive. We chose to build two element collinear antennas out of tig welding rod.

PREPERATIONS

Prior to race day, several of us went out on course to do some hands on testing. We already knew the VHF/UHF talkpaths were solid from years prior. So we took some handheld LORA ‘talkies’ with us and zip-tied them to fenceposts at our node sites and began testing. The signal levels between nodes were nearly spot-on with the projected levels from splat!. Several of the point to point paths were not true line-of-site and going into test day, we were prepared to deploy a LORA ‘repeater’ if necessary to facilitate connectivity between nodes as needed. Needless to say, we were thrilled when direct communication was established without the need for an intermediate device.

RACE WEEKEND

We were on course about 10A Friday setting up equipment. The 350 mile race started at 3:30P and the riders were not expected up in our remote area until about 5P so we had time to get our equipment deployed ahead of the fray. We had mild challenges installing the equipment at each of the sites but nothing that could not be overcome. Well, except for our XBand2 LORA node where we had a PICO failure. Part of our overall design was to ensure that any failure on the LORA side would NOT impact our talk-paths at all. So, the PICO failure was not going to impact our ability to service riders. We considered a field replacement but decided to simply wait until after the race, to get it back on the bench to determine what happened.

One of four trucks ready to roll

Overall, things worked well. Software in our PICO had voltage, current, temperature and fuel thresholds. When all of the thresholds were within specified parameters, the node was considered healthy. If any status fell outside of expected norms, the node went into a troubled state. When the respective nodes were ‘healthy’ we received a status report every 15 minutes. If one was ‘troubled’ we received a nag status from that node every 5 minutes. These updates came to our phones via the Meshtastic app. and was fully independent of an Internet connection. That allowed us to monitor the overall system health from a central location and dispatch as necessary. Fortunately, we did not have any issues that required an unscheduled dispatch. Our biggest challenge was the fuel level sensors. They had a sensitivity pot and getting them calibrated properly was a challenge. By the end of the race, we had two of the three working reasonably well. After we got them home we think we learned the trick for a perfect calibration. We’ll see.

Screenshot of the data sent to our phones via LORA/Meshtastic

Dirty Kanza/Unbound Gravel Communication Background

Posted on by Jim Belford

A Little History

Unbound Gravel began as Dirty Kanza in 2006 when 34 bicyclists met in Emporia, Ks for a 200 mile bicycle race across the gravel roads in the Flint Hills of Kansas. It has grown in popularity through the years and now there are more than 4000 riders who toe the start line annually.

Given the size and remote nature of this course, mobile telephone coverage is spotty at best and as the ride grew, it became difficult for race directors to provide rider health and welfare support to participants on the course. With the growing number of riders, each passing year became increasing difficult to provide rider support and extraction.

Kansas City Jeep Club began providing assistance in 2008. They brought upwards of 20 jeeps to provide the “wheels on the ground” along the course. In the early years, KCJC used mobile phones and CB radios to communicate along the course. However, the extreme size and remoteness of the course demanded a more robust solution.

In 2015 the Jeepers began passing their amateur radio exams and earning radio licenses. To their credit, they had the foresight to set a radio standard in their club and many of the Jeeps were outfitted with Yaesu 7900 dual band mobile radios and nice antenna systems. They realized early that it would be much easier to manage a fleet of radios and operators if the hardware was the same.

In 2015, EARS (Emporia Amateur Radio Society) began talking to DK/KCJC about options to work together. We determined our objective was to provide a communications medium across the course which would allow the Jeeps to communicate with a dispatch center. This would provide the infrastructure to significantly increase the level of support provided to the riders. Meetings were scheduled and concerns voiced. Radio coverage maps were created over the area for our local 2 meter repeater and a goal was set to provide S7 or better radio coverage to over 90 percent of the course for the 2016 event; a goal EARS thought was attainable. When the course was released , we learned that race course creators are a sinister bunch and they find every hill, every crevice, every water crossing and wicked piece of terrain they can find to challenge the riders. A consequence of these rider challenges is the difficulty faced by the communications team to deliver quality radio coverage into these difficult locations. Even with very nice 2 meter repeater coverage across much of the course, it became increasing clear enhancements would need to be done to achieve our goal.

EARS set out to install temporary, linked repeater setups along the spotty sections of the course and had things in place when KCJC arrived in town the Thursday prior to the race. As they took off Friday morning to scout the course the only radio communications the Jeep drivers had ever experienced was CB. They continued to be skeptical as we told them we were expecting to have coverage for a large portion of the course. Armed with their new amateur licenses and quality radio hardware, they were amazed at the clarity and distances now available to them through the repeater network. Although the coverage was not perfect it was a significant improvement from previous years. They were stoked; we all were.

On race day, EARS put the repeater into net control operation and all communications were handled through net control. Even though we had talked about this format and the protocol and practiced it the day before during the course scout we knew it would be a learning curve for all parties involved. And it was. We all struggled early on as multiple Jeeps began needing airtime to report problems, net control needed airtime to dispatch Jeeps to specific locations etc. However, as the morning pressed on, everybody began to settle into the task at hand and things smoothed out considerably. In retrospect, the main issue was patience. Net control expected Jeeps to answer immediately and the Jeeps expected an immediate answer from net control. The reality is occasionally the Jeep operators are busy with another task or out of their Jeep assisting riders. Similarly, net control was not always immediately available due to situational chaos. To the credit of all involved, on air efficiency went up consistently throughout the day. By the end of the event, some 21 hours later, everybody was handling the radio traffic and dispatch protocol quite well.

It is unlikely that anybody anticipated the sheer volume of radio traffic. Looking at repeater reports, our local machine was keyed up, passing traffic for over 14 hours of the 21 hour event. We were incredibly busy.

DK, KCJC and EARS had several post event meetings to discuss all facets of the event. We had frank discussions on what we did well and what we could do to improve. Each volunteer took action items out of those meetings to fix things for the following year. As DK 2017 approached, we again met and discussed improvements to our system and processes. We have used that process repeatedly year to year to learn and provide a better service for the safety of the riders.

Unbound Gravel Communications v.2022

Posted on by Jim Belford

Background

Unbound has become one of the largest gravel bicycling events in the world. Once riders toe the start line of this 200 mile course, they are on their own. Riders can only accept help from non-participants at two designated checkpoints along the course. If a rider calls for assistance at any time outside of one of these checkpoints, they are disqualified and not permitted to continue. Given the distances and terrain of the Flint Hills gravel roads, mechanical issues are part of the day for many riders. Heat, wind, rain and severe weather have all contributed to this event, sometimes all in the same year. Much of the appeal of this event is the challenge in working through personal and mechanical issues on the course.

Objective

Kansas City Jeep Club (KCJC) and Emporia Amateur Radio Society (EARS) work together during this event to satisfy two main objectives. We provide a dispatch center and communications infrastructure system where any participants can call (if they have cellular coverage) if they are in an emergency situation. As our services become required, we are able to roll Jeeps on course to provide assistance as needed.

In emergency situations, it is important for us to get a Jeep crew into an area immediately to manage a situation until EMT/First Responders help can arrive. We have been involved a number of Ambulance and LifeFlight rides over the years but thankfully v2022 did not have a 911 type emergency come through our dispatch center. This is unusual but welcomed.

We also worked with the non-profit group Camp Alexander. This group volunteered their time to provide non-emergency race day support to riders requesting their services. EARS/KCJC, provided transportation and extraction services for those riders who wished to leave the course. The reason for the extractions varied from dehydration and hunger to cramps or a simple lack of motivation. Riders understood that once they accepted our support, their day on the bike was over and they were disqualified.

With over 4,000 riders heading South out of Emporia, Ks into the Flint Hills prairie, the sheer volume of riders guarantees a busy day for race organizers and volunteers. Parts of this 200 mile course are not covered by consumer cellular network providers so telephone simply cannot be counted on. We learned years ago that if we wanted to insure a communication path throughout this course, we would have to build it ourselves.

Planning

In order to provide these services we must be able to communicate effectively to, and travel to nearly all points on the course and the access roads in the area. The process of creating a plan we could execute on race day began months prior to the event. This is the 6th year we have been involved in this event so we had the benefit of past experience working in our favor. We knew we were again going to put our club trailer downtown near the start/finish line. This puts net control/dispatch in constant contact with race organizers and creates an office workspace for volunteers. This trailer was built to be versatile and for this event we use all three of the operating positions. Position #1 is used for our net control radio operator. He has access to a dual band radio, a reasonable antenna and our local repeater. Position #2 is his logger and dispatcher. As requests come in, and are resolved, this position logs information into the computer database and keeps track of dispatch tickets. Position #3 was used for administrative purposes. Riders are given an emergency contact number during their orientation and that number rings into our trailer where calls are triaged and dispatched as necessary.

Communication is pointless if there is nobody to talk to. KCJC and EARS have a symbiotic relationship and we rely on each others strengths. EARS provides the communications logistics while KCJC provides a fleet of 20 amateur radio equipped Jeeps and radio operators to roam the course and provide the eyes, ears and muscle along the route. Each group relies heavily on the other and neither could meet our objectives without the other. We all work tirelessly throughout planning and race day and both groups thoroughly enjoy the relationship.

Early in the process, race organizers provided us with the course which we laid down in Google Earth. Given the nature of the task at hand, our goal was to provide better than S7 signal across the course. This year, the route went South, directly into the footprint of our local repeater. Using SPLAT! software we generated a computer generated theoretical > S7 coverage map of our 146.985 repeater over the top of our bicycle course. As the route continues South however, we expected repeater coverage to become problematic and the SPLAT! propagation maps validated our concerns. So, portable crossband ‘drop repeaters’ were planned at strategic locations along the course. Each of these drop repeaters would provide a 15(ish) mile radius UHF footprint and crossbanded VHF link back to the main repeater. By choosing our drop locations carefully, UHF coverage into previously weak spots would be solid and provide our Jeeps with nearly 100% coverage across the course. This requires our Jeep operators to monitor different frequencies as they move along the course but that is covered in our pre-race meeting. We then began looking for favorable locations to drop our portable crossband repeaters to supplement our coverage. When looking for these locations we took several considerations into account:

  • Height above average terrain (HAAT)
  • RF access back to our 146.985 repeater
  • Serviceability
  • Landowner Permission

We have been very fortunate in the willingness of landowner’s to grant permission for us to use their land. The relationship between the race and the landowners is solid and is a testament to the founders of this race and their ideals. As with so many things, the relationship made our job much easier to execute.

Whenever possible we like to put our equipment near to but not on the course. This gives us the ability to get in and service equipment without traveling on the course. This allows for quicker access times and reduces the amount of vehicle traffic on the course making for a safer environment for the participants. Thankfully to date, the only real service we have been involved in during the race has been the re-fueling of the generators.

Studying the coverage maps across the course, we determined the soft spots in our repeater coverage and began looking for satisfactory locations for our drop repeaters. Topo maps were utilized to determine high spots and then theoretical SPLAT! coverages were created. In the end we determined three additional drop locations would serve our needs well. They were located near MM 45, MM77 and MM 100. The respective coverages are shown in the links.

Implementation

Surprisingly, the hardware for our drop machines is quite simple. We generally use standard dual band Kenwood V71 radios in a cross-band configuration. Power comes from small Honda generators and Astron linear power supplies. The entire setup (sans generators) is bolted together along with fans blowing across the heatsinks and is housed in a Tupperware type container outfitted with input and output forced airflow for cooling. The mass of the power supplies provide ballast for the containers and keeps them from blowing over (or away) in the Kansas wind. The fans provide the airflow needed to keep the equipment temperature within acceptable limits. The antenna systems are Diamond X30 fed with LMR400 cable.

Due to the relative small footprint needed and our drop location HAAT, we do not need much of an antenna support structure. We use a 5 to 10 foot mast attached to a fence post as the mounting for our antennas. It always amazes us how far radio waves fly when launched off of high ground.

Observations/Gotchas

It must be noted that using a standard dual band rig in a high duty-cycle crossband application is risky and has severe limitations. We run these dual band radios in low power and we blow a significant amount of forced air across the rigs to keep them cool. We also use fans to pull filtered outside air in through the bottom of the container and a complimentary output fan at the top of the other side as an exhaust.

For this race we used and tested two smaller Lexien radios and a homebrew Arduino “controller” as the crossband system for one of our drops. While not quite as compact or easy to program as the V71, this setup offers several advantages. By using two discrete radios we cut the duty-cycle of each radio in half as well as doubling the heat sink capabilities making it possible for us to run higher power. The Lexien radios are significantly cheaper than the name brand counterparts and we can custom 3D print fan enclosures to provide substantial airflow around the heatsinks. We also are testing two QYT radios in crossband mode connected via their proprietary cable which eliminates the need for the additional Arduino controller. The results of these tests were encouraging.

We have learned that race promoters enjoy putting their riders in the most beautiful and difficult spots imaginable. We have also learned that these challenging spots are often in gullies, behind hills or along creeks. These landscapes rarely lend themselves to good RF propagation which increases the need for the drop repeater systems. Given the relative simplicity of our drop setups, things worked very well. As expected, there were locations on the course where coverage was not optimum but those areas were few and far between.

You may have noticed we discuss location in terms of the mile marker on the course. Garmin and other GPS companies make sophisticated bike computers that give all types of race day metrics back to the riders. In events like this, nearly every rider utilizes these devices. Distance ridden is common on these units and nearly every rider knows exactly how far they have ridden at any given time during a race. When calls come into our dispatch phone, the first question the rider is asked is how far into the race (in miles) they have ridden. That gives us the course mile marker and Jeeps are dispatched accordingly. Jeeps, dispatch and net control all use Avenza Maps with the course and mile markers layered on a real time map. Given the very long stretches of course without any cross-streets or houses, we have determined the mile marker method is superior to using addresses. In rare cases, riders are unaware of their location and we ask them to drop us a “Google Pin.”

Prior to the race, each rider is given a unique “bib number” plate. This plate is much like a small license plate and must be displayed on the riders bicycle. This bib number corresponds to the rider for the day and is used as the primary form of identification.

Problems

Interestingly, most problems we discover are not technical. Our biggest issue is finding riders that call for extraction and then are nowhere to be found when we arrive. Often times we learn the rider was able to fix their problem or gain the motivation needed to continue before we could arrive. This happens quite often and we are forced to continue looking for these riders until we can confirm their safety.

Long Day

This is a long race and that makes for a very long day. Our day begins at 5A with our support meeting. The official race ending time is 3A the following morning with riders occasionally coming in past the deadline. Some of our crew run in shifts but others remain throughout the 22 hour event. We were able to break down our drop repeater locations once the race moved past their locations. In one case that occurred way after dark but it did allow us to get home before sunrise.

Conclusion

For us, this is what ham radio is all about. It gives us the opportunity to put our skills and resources to use while helping others. This event is a huge deal for our small Kansas town and being involved in such a hands on way is a lot of fun. The evolving relationship between EARS and KCJC is the “secret sauce” for our combined success. Generally, the calls that come into our trailer lead to a Jeep roll, minor first aid and a simple trip back to town for a rider, which is fine with us. We much prefer those casual calls to the serious situations that come up occasionally. When those serious emergencies emerge (and they do), our team is there to provide any assistance needed for as long as is needed. The overall goal is to make the event safer for all participants and spectators.