Urbana, Illinois


#22

We’re pretty much at ground zero. We have a total of 7 devices, 3 have been fashioned into stationary nodes, leaving a pair of my wife and I, plus a cuple more for other users. I’ve been using one of them for testing along with mine (which almost met its doom this afternoon, but which made a remarkable recovery after some poking at it. )

I have filled out the survey form. Urbana is both big enough to make a fairly extensive and diverse use case and small enough to clearly see the effects of the results of a concerted effort to engage people with goTenna Mesh…


#23

Updating a few things…
We raised our home node, UMESH1, up to ~28’ . This helped its reach some, and put it where it’s easier and safer to work on.

I attended an event with some old political activist friends who are helping a local county board campaign. This gave me an opportunity to introduce them to the goTenna Mesh. They were intrigued by its features and the fact that we’re building a local network to support its use.

I tend to set up to automatically set pins at one minute intervals for 15 minutes when leaving home on an errand or returning after being away. This allows me to check performance and signal at various locations. I leave the GTM on the dash , so it’s not an ideal location for reception, but about the best for mobile monitoring short of duct-taping it to the roof. In reviewing, last nght’s “take” I found that I was getting good reception at UMESH1 of the pin from the location of the town-edge venue I just described. This wasn’t a confirmed message, but was a new record for reception at 2.38 miles. This pointed out two issues I’m still working around.

Raising the home node made its reception and transmission noticeably better than for return paths from mobile GTMs. Eventually, fill-in stationary nodes off-axis from our intial 3-node string will help get mobile GTMs through, but the current lack of more than 2 hops and linear nature of the UMESH network means that mobile reception and transmission tends to lag.

Second, pings and pins seem to be more reliable than messaging in getting through, but don’t tell you much directly about stationary node performance since there’s no direct indication of any intermediate hops when these occur, unlike the “hopping” indication when a message arrives via a stationary node.

Whether it was the raising of the home node or something else, I’ve been getting noticeably fewer test messages that indicate these nodes are at work. In fact, I was at my current work site at the south end of the UMESH when I received an off to work message from my dear partner as she left the house to walk to work. This was apparently a direct jump without stationary node involvement per what was documented on her GTM. Impressive, but not reliable unless you can find and stay in that “sweet spot.” This and observations of the inaction of LEDs in its clear box on the home node when initially transmitting a message makes me wonder if the mesh protocol tries to send direct first, then only goes to a stationary relay when that initial direct comm fails?

Relative lack of sun (better than the week before, though), the relative weakness of signals from mobile GTMs versus those mounted well above ground level, combined with the current general opaqueness of stationary node operation (visible when it happens only to the sender, IIRC) makes you wonder if the stationary nodes are working at all at times. Messaging usually gets through where its expected to relatively easily, although some locations that would seem to have obvious stationary node coverage to assist do not awwm to assist at the same time they lack direct transmission LOS. Again, adding off axis nodes will help with this by providing alternative “looks” for a signal to hop to make it through, but in practical terms this remains dependent on expanding the total number of hops available for a message to move through on its way to its destination.

The final issue in testing is the small number of users. I “simulate” another user by leaving a GTM and phone at one location, but this is only convenient at the two end points. Without the capability to jump from one end to the other of the three-node network due to the limited number of hops currently available, coverage is spottier than it will be with that and a denser mesh network. So I’m working on expanding the user base, which will get easier as time goes on and capabilities come online for the goTenna Mesh.


#24

Try hanging your GoTenna from your rear view mirror instead of the dash. I found that it can really help improve the signal performance.


#25

My custom goTenna case.


#26

I suspect this depends on the vehicle. It will certainly help, and having vertical polarity is also a good thing. I’ve started using this as one of the places I test from and had mixed results so far in our Land Cruiser.

I actually liked using the dash as a resting spot because it’s a somewhat compromised spot. Most users aren’t radio-centric like many of us, so testing like this represents a use case that may be more in line with average reception. If you pick up a signal this way, you also know it’s better than the weakest possible, which also gives a more real-world test.

dennilee’s “case” is spiffy (does that thing even have seatbelts?:thinking:) but like fuzzy dice, hanging stuff from the rearview mirror can at times lead to unwanted attention :policeman:

The GTM is small and non-descript enough it probably won’t be an issue, but something to be aware of if you decide to hang it there full-time.


#27

Here’s an explanation of how the Mesh protocol works from goTenna:

"In the case of goTenna Mesh, the process starts with initial discovery. A user sends a message out in all directions with the intention of finding the intended recipient (or IR). If the IR is not in range, a burst broadcast goes to all other devices in range, essentially asking if anyone local knows the whereabouts of the IR.

Once a relaying goTenna Mesh confirms that it can reach the IR, the path to the IR will be defined and remembered, so messages between the original sender and the IR will follow the same path (again, the optimal path with the fewest number of hops). The networking protocols built into goTenna Mesh — which we’ve named Aspen Grove — will also actively avoid interference and obstructions by finding an alternate path if the established route is compromised."

So, yes, they will try direct comms first, and then try to relay if the recipient is out of range.


#28

Makes sense. I need to do some more testing to see what the current system status is. Here at home UMESH1 is working but it’s on house power not solar. It’s also easy to check, unlike the other 2 that require some effort. So UMESH 2 and UMESH 3 are on uncertain status. They’re hard to test due to the limited number of users at the midpoint node and the inability to test end to end until the new firmware makes the mesh more “hoppy.”

I’ve been too busy to check things lately and the trend seemed to be against the nodes staying up due to the lack of sun and relatively limited recharge capacity of my beginner’s luck solar systems. Had a tooth pulled this morning so may try some testing later this afternoon if I get to feeling better. Right now it’s a bummer :hear_no_evil:

I have been testing bigger battery packs (I have several that are rated at 8,000 mAh and one that is around 25,000!), and they’re ready to go in service next time the nodes get back to ground. And I think bigger solarpanels are in the future for more reliability. I’m thinking about pairing up the panels I already have in inventory once I research how best to do that (they’re Voltaics), so could end up with 2 W +2 W = 4 W and 3.5 W + 3.5 W = 7 W. I’m actually hoping that I can make two sets of 3.5 W + 2 W = 5.5 as that seems optimal and I could do by buying one more 2 W panel and get UMESH 2 and 3 running full-time. Then I just buy larger ones.


#29

It looks like Voltaic makes a combiner box to pair up panels for the V15:

https://www.voltaicsystems.com/2-panel-6v-12v-circuit-box


#30

Excellent! I was looking at that thinking it might be what’s needed. That should set me up for the network rebuild. I suspect 5.5 watts into a 8,000 mAh battery box will give a much larger margin of error. Wth a little luck, it might be adequate to last through January’s around here (and November, December, and February, too! :crazy_face:


#31

Some updates for Urbana’s UMESH…

UMESH 3 is back on the air and better than it ever was. It’s the first node to be rebuilt to the upgraded standards that seem necessary here. It has a Voltaics 6 watt solar panel and a generic 8,000 mAh battery pack that is always on to deal with the long gray winters here.

I salvaged the parts from the original node. Its 3.5 W solar panel is now paired with 2 W panel and a bigger battery pack. Packaged in the medium sized HF “Apache” case to accommodate the larger panel aize, it’s mostly foam inside so plenty of room for other potential e-cohabiitants. Cooling may become an issue (150 F high operating limit IIRC for GTM units) but we’ll figure that out then I suppose.

I did run some heywhatisthat profiles for various sites. It looks like a theoretical possibility that the folks further south in Urbana with a stationary relay node might be able to pick up UMESH 3 now that it’s back on. It’s almost a mile from that approximate location to UMESH 3.

I used only approx. locations here in all instances and will not be posting the data other than to say it’s to the best of my ability using self-published info, I want to respect everyone’s privacy based on the potential to create over overly precise interpretations of what is often fuzzzily degraded data by users interested in not being too precise about their location for privacy reasons.

That said, the profiles show interesting places where service can be available at longer distances than expected, typically 1/2 mile in our non-dense but still urban environment. One is that another house we have access to between Lincoln Square and the high school would fill a hole in coverage for west of Lincoln Square on Green Street. There is good coverage from UMESH 1 east of Lincoln Square on Green St. to Cottage Grove. And we still have plans for a downtown Urbana node once we get this other stuff squared away.

Testing UMESH 3 today showed it solid at Anderson and Colorado. Coverage should be up to about Washington but didn’t test it fully.

I will get some pics soon. If anyone else uses it, let us know if it works. Since there are no central servers, etc, only initial user feedback tells you much is my very limited expectation, but suggestions and comments are welcome. I just want something neat and useful, so beyond that it’s all bonus utility that serves the community in many ways we’re still not sure the extent of yet.


#32

Some pics to catch up. The new Mark 1 Standard Stationary Node (SSN) is built in a Harbor Freight medium “Apache” (#63926) case. I paint it white to help hide it and limit heat buildup. I drill one hole in it to pass through the line from the solar panel, plus a few more to attach the panel mounting brackets. The window weights hanging from the case handle are slung over the roof ridge and counterbalance the weight of the case.

This view shows the Voltaic 6 watt panel. It’s mounted to a wood platform that is supported by brackets. I don’t pay a lot of attention to getting exact solar angles, just point it towards the sun and go.

The case opened shows a curious cat (Wally!) and just the top of the goTenna. The case is tall enough so the GTM can stand up for good vertical polarity on a flat roof. For pitched roofs,a wider opening is made and scraps of foam are used to angle the GTM inside the case in order to offset the roof angle to keep the fractal antenna vertical.

This pic show the internal arrangement of the 8000 mAh always on-capable battery pack connected with Voltaic cabling between the solar panel and the GTM. This pack can provide about a week’s worth of power when fully charged on its own. Seems like overkill, but decided to try it to avoid downtime during winter gloominess. There’s still plenty of room for other equipment.


#33

I’m thrilled to report that UMESH achieved end-to-end connectivity today around noon.
:joy:

It can be a little spotty to go end to end with it, because there are 3 stationary nodes, but you can max out the hop count with just two plus the end user devices. There is coverage overlap of the nodes that makes this possible now that everything is working. Once the max number of hops is increased this will go away as a significant user issue unless you’re right at the network fringe.

Coverage is basically as described previously, from the east side of Lincoln Square to around Florida & Philo Road. There may be coverage further south of there due to another stationary node located south of Sunnycrest who I do not know other than from the pin on the imeshyou map.

We would be thrilled if there are any others who can report back on their experience with UMESH. Your comments and suggestions are welcome!


#34

I did quite a bit of testing today, all with the GTM hanging from the passenger side sun visor.

Keep in mind the system topology, a string of 3 stationary nodes, but only two of which can be used as intermediary hops. You can be right next to the end node, but what the device needs to do is hit the intermediary node, then the other end node, and finally to the device addressed. Fortunately, the end nodes are spaced just about the right distance that this does work if all three nodes are up.

Here’s how this works in practice. UMESH1 is atop the house, then UMESH 2 is half in between (about 1/2 mile), and finally UMESH 3 is a block or so away from Test Device Two, but another 1/2 mile beyond IMESH2.

This will be easy once I can use more than 4 hops. Right now, one end of the network of links between the transmitting and the receiving parts of the devices must be within range of the middle node. Think if it as an extra tall top step. Nonetheless, despite the seeming radio voodoo, it mostly works.

But on my end here at home, right now I have to hang the GTM from a certain spot behind the basement window looking towards the southeast, set it on top of a file cabinent in my office in the southeast corner of the house or set it in front of the south facing wndow behind my wife’s desk. The rest of the house is unable to get a good enough siting on the antenna at the middle node.

If there was another hop possible, the signal could instead go more easily up there to a stationary node there first with its much clearer view of the radio horizon, then the signal could be beamed down to our GTM units where it would strong enough to reach throughout the house. Just one more hop would result in a much better signal here, for instance, but having many more will be great! For now, the restricted number of hops is more nuisance than anything, but it does mean that users have access to only part of the system’s capabilities until it becomes fully operational with multiple hopping possibilities.

Part of my testing was verifying that meshing was going on at the south end of the network. The stationary node located there predates my bumbling around with UMESH. I don’t know them, but my hat’s off to these local pioneers of meshing. Sure enough, there’s good signals down around E. McHenry and S. Cottage Grove where it would otherwise be slims pickings at the tail end of UMESH. It’s all nicely seamless and just happens to fall into place.


#35

Great testing! Do you have any visuals of your nodes in the field? I’d love to see the set up. We’re hoping to get goTenna Mesh hops up by Late April / Early May! That should really help this relay!


#36

Thanks!
Don’t have many pics of the installed nodes, beyond the one about halfway down the page above. I’ll try to get some. It will be a good excuse to fly the drone :smirk:


#37

Didn’t need the drone for UMESH1, sitting on top of UMESH HQ lashed to one of several antenna mast here with giant tywraps.

This is the only non-solar-powered node in the UMESH constellation. A 10’ long micro USB cable reaches up to the node from the attic for charging purposes. If need be for a lengthy outage, the USB cord could be unplugged inside the attic from its wall wart power supply and a battery pack could be plugged in to power it until service is restored.

Plans are to rebuild this node to current solar-powered node status. It’s currently housed in a medium sized watertight box whose rim I notched carefully to pass the power cord into. It’s not clear plastic, but is transparent enough you can read the status LEDs on the GTM unit.


#38

They are great! If I wasn’t terrified of being on my roof I’d try spinning one up.


#39

Thanks!

Yes, be safe first. If you’re uncertain about being on the roof, then discretion is the better part of valor.

This happens to be safer than the pics indicate. The antenna is mounted using an old Radio Shack gable mount kit, attached to the end of the roof. Rather than being 20’ or so to the ground, it’s only about a 3’ drop onto the garage roof, which itself provides relatively safe access via a 10’ step-ladder. I can stand on the house’s roof ridge and work on the box, so it’s attached about 6’ above the ridge itself.

Still, you need to keep your balance, as a slip could be disastrous, with a long roll following a fall, then you’d be [temporarily] flying as you leave the eave for an even rougher landing.


#40

Last night I went to a meeting of CU2600 a group that is interested in all sorts of computer-driven and other technologies. They’re a new group (this was the 4th meeting since they started in December 2017) but seem to be finding ways to experiment with and educate about new technologies. About two dozen attended, a pretty good sized crowd and one where most were active participants in the discussion.

I made a brief announcement that UMESH achieved IOC (initial operational capability) this week and described the current service area for them using the imeshyou map. There was lots of interest, including from a person who said she believed her family’s home was located on the highest point in Urbana (even better, she’s a Ham and has been thinking about putting up a tower for a long time :wink: ) Another fellow said this seemed like an ideal solution for friends who have land and a small community in Hawaii that doesn’t have cell service, but would like some sort of reliable communications.

I plan on staying in touch with CU2600. Once the firmware update to increase the allowable hops comes through and is applied, I will do a formal presentation on UMESH and the goTenna Mesh that makes it work at a future meeting.


#41

This is amazing!! Thanks for the report — hope folks in your community join up here on mesh community