Follow

Probably time to revisit the air-dropped WiFi mesh node idea

The recipe so far:

- Cut-away drogue
- Flip-up 5V solar panels
- 18650 battery pack
- Hacked travel router
- Spray painted PVC tube
- 2 End caps
- Spring-loaded, sheet metal legs

+ Maybe an ESP32 to handle maintenance like flipping the solar panels down and up again via solenoid to get rid of snow/debris accumulation

I’m envisioning the sheet metal legs to pop out similar to the air breaks used in high-speed low altitude air-dropped munition used by the U.S. Air Force

They’re colloquially called “Snakeyes” and were a common during the Vietnam war. I figured the principle can at least be used for something more constructive than war

Show thread

some thoughts on the idea implementation-wise 

@cypnk snow is heavy, a solenoid wouldn’t be the right tool for the job. Motorised flip-out mechanism also adds many more points of failure. Could it put its panels on an angle to slough off snow naturally? Is there a reason it couldn’t be occasionally cleaned if necessary?

some thoughts on the idea implementation-wise 

@s0 @cypnk enclose in a tetrahedron case with every face a solar panel so nothing's flat and orientation doesn't matter much?

some thoughts on the idea implementation-wise 

@mike @s0 The fewer and simpler the components, the less there is to break/interfere

The only flat pieces are the solar panels, which are held in place with a spring lock. To flip down, it just has to unhook and a solenoid brings it back up to hook again

2 Solenoids can be actuated by simple relays, which are already available for Arduino/ESP32 projects

some thoughts on the idea implementation-wise 

@cypnk @mike relays and solenoids are fine for DIY projects, but for a real high reliability project intended to be unattended for significant time, neither would be appropriate to use. solid state switching of a different electrical actuator with more efficient energy transfer (probably a motor attached to a rack & pinion or hinge & cam) would be much more reliable and crucially, more energy efficient.
But I'm not quite understanding when these would be useful -- wifi is short-range. where is there an area where wifi coverage is needed but access for maintenance can't be arranged?

some thoughts on the idea implementation-wise 

@s0 I’m counting on the snow being heavy, actually. Once unhooked, the panels will violently collapse and knock against the tube body

The solenoid is only to lift the panels back up. No motors at all

some thoughts on the idea implementation-wise 

@s0 Since this is for an “air dropped” node, the implication is that the area is inconvenient or impossible to reach for human intervention

I.E. Fire/flood/war region

some thoughts on the idea implementation-wise 

@cypnk In the case of these first two, your suggested device wouldn't work, as I understand it. It's neither fireproof nor floodproof (even if waterproof, floods just smash everything & cover it with mud). In the case of a war zone, how would it be better than manually deployed mesh nodes by people on the ground? It's not of much use somewhere where nobody can get to, as there's nobody to communicate with? Is the suggestion to be as a 'bridge' between two populated areas separated by a dangerous zone? I would guess that long-range (lower freq and/or directional) mesh networks would be more serviceable. ( @diodelass would be the expert on this)

some thoughts on the idea implementation-wise 

@cypnk I've realised I may be coming across very critical. Sorry if I took your idea of a fun project to make yourself (it absolutely would be) as more specific suggestion to solve a hypothetical war-zone-communications problem

some thoughts on the idea implementation-wise 

@s0 @cypnk good ol' 2.4 GHz can have surprisingly good line-of-sight range (kilometers) if you have a directional antenna and decent transmit power, but this sounds like a higher-caliber application than the usual for that.

you're describing something mass-produced, long-term reliable, and airdroppable, which is a very tall order already.

some thoughts on the idea implementation-wise 

@s0 @cypnk
And fireproof and waterproof and ... stealthy? oof.

so, moving parts are almost entirely off the table if you want this to be at all reasonable to design or fabricate (reliable robotics are TRICKY and your power constraints are very tight)

but anyway we were summoned to talk radio so we guess we are digressing

you probably want something in the VHF range for this.

some thoughts on the idea implementation-wise 

@s0 @cypnk we guess the scope is unclear for us here - upthread is talking about something that might be thrown together in a basement shop, but the later musing about disaster-relief type stuff sounds more on the level of military hardware.

some thoughts on the idea implementation-wise 

@s0 @cypnk in either case, VHF and UHF gear is not so hard to come by! we're thinking about those cheap Baofeng handsets. the ham radio people have been demonstrating for decades the effectiveness of ad-hoc repeater networks for spanning long distances, and this'll be at a smaller scale than that. the expected range these days for VHF/UHF walkie-talkie style radios is basically the horizon; VHF is slightly better about obstacles

some thoughts on the idea implementation-wise 

@s0 @cypnk we're imagining a droppable payload that looks like a low tripod designed to keep a monopole antenna (a meter long or so) perpendicular to the ground. inside, you've got one of those radio handsets, your favorite computery devboard, and a sound card of some sort, maybe one of those little cheap USB ones, to let it talk on the radio.

some thoughts on the idea implementation-wise 

@s0 @cypnk as for talking on the radio... we're in the middle of writing some software to do exactly this! and definitely way over-engineering it. but there are a lot of ways to encode data over audio links; you're basically reinventing the landline modem. an actual modem may work decently well, even. audio frequency-shift keying is popular for this sort of thing. we're doing chirp spread spectrum, because we're silly and extra.

some thoughts on the idea implementation-wise 

@diodelass this project sounds super cool. If I didn't have 6000 other projects to work on I would be so getting into radio stuff now

some thoughts on the idea implementation-wise 

@s0 @cypnk then you'll have to implement your mesh on top of that, with all the usual nightmares of shared-medium multiple-access control and routing, but you signed up for that at the beginning.

it'll be slow, very slow, but that's a tradeoff you make for range and cost.

if you want to go up a notch (and double your BOM cost), throw a fancy SDR in there instead of the walkie-talkie and use whatever modulation scheme you please.

some thoughts on the idea implementation-wise 

@s0 @cypnk we recommend the walkie-talkie because it's cheap power with convenient electrical hookups. a typical Baofeng costs somewhere in the middle tens of USD and gives you around 5 watts of transmit, using a typical FM narrowband modulation scheme with a 12 or 20 kHz channel width. it's not designed for digital, and there are more expensive ones that are, but it's still a capable radio frontend.

some thoughts on the idea implementation-wise 

@s0 @cypnk as for specific frequency selections, well, it's sort of tricky actually

the thing is that the spectrum's pretty crowded these days and there's hardly any space that won't have licensed users getting mad at you for getting in the way.

it's also a lot harder if you want to stay legal under whatever your jurisdiction's rules are.

some thoughts on the idea implementation-wise 

@s0 @cypnk if you're in the urban USA, we have been supposing lately that the VHF television allocation (175 to 200 MHz or so?) might be a fairly safe playground for narrowband stuff, since:
a) actual airwave television seems pretty uncommonly used in the highly-developed USA in 2020
b) narrowband transmissions are less likely to interfere with wideband digital television than they are with other narrowband users

some thoughts on the idea implementation-wise 

@s0 @cypnk
we think the worst place for pirate radio is anywhere that licensed users are talking on the same mode as you. if you're narrowband FM, avoid anyone using narrowband FM, like cops, emergency responders, the FRS/GMRS walkie-talkie bands, and so forth. If you're ultra-wideband, go somewhere besides the 1-2 GHz range where military telemetry happens.

some thoughts on the idea implementation-wise 

@s0 @cypnk if you can follow the rules on the ISM allocations, that's a great idea (900 MHz is a cool one for this sort of thing). if you can't, though (e.g. you need more power than the one watt they give you), it's probably best to avoid those, because they're full of sensitive receivers and low-power transmitters that might hear you even if you can't hear them.

some thoughts on the idea implementation-wise 

@s0 @cypnk oh and
do not under any circumstances get in the way of cell phone signals, because that'll get you in big trouble.
same for GPS and aviation, and probably be careful with maritime too.

bad idea re: some thoughts on the idea implementation-wise 

@diodelass @s0 @cypnk go down to HF and put it on 14.313 MHz upper sideband

nobody will notice with all the jamming that happens on that frequency anyway

(this is a bad idea both because it’s illegal unless you can operate under amateur radio rules, but also, using a frequency that’s commonly jammed means you need more power to get through)

some thoughts on the idea implementation-wise 

@s0 @cypnk we think the criteria we've come up with are like
find a frequency that:
1. avoids the really important stuff
2. is not allocated for use by the same mode as you
3. has mostly high-power transmitters
4. has mostly low-sensitivity receivers
5. doesn't have very many receivers per square kilometer

some thoughts on the idea implementation-wise 

@s0 @cypnk commercial bands are actually pretty good for this, because airwave stuff is declining in popularity there, and it's also just very easy to dodge the local channels and find a quiet gap. may be less so in a big city, but elsewhere there is usually plenty of space. and since you're talking at a few watts instead of kilowatts, anywhere that's quiet to you probably has no transmitters with coverage zones intersecting yours.

Show more

some thoughts on the idea implementation-wise 

@s0 @diodelass I'm sorry, I'm not understanding what you're expecting, but let me try to address some points

> fireproof nor floodproof
Numbers + Cost > Durability
500 - 1K per deployment
~$150 per unit

> long-range
Core is a router running OpenWRT ad-hoc mesh (802.11s). 2.4GHz @ 20dBm

2 Antennas, 0.3 km range omni (users) and 2km directional (node2node/uplink)

> bridge
Yes, temporary, for emergencies. 30 Days expected lifetime

some thoughts on the idea implementation-wise 

@s0 @diodelass I think maybe you're assigning this a far greater scope than it actually is. Maybe more can be clarified once the prototype is designed

@cypnk I really like the image of the air brakes doubling as solar panels. But I'm guessing that wouldn't really work.

@borup There are no breaks besides the drogue chute for slowing the device down. The solar panels don't do anything beside providing power and the legs don't do anything besides keeping it upright

There are no double-duties for any of the major components

@cypnk now you got me curious what you are talking about lol.

@jordan31 Many, many, many moons ago, I had a chat here with a bunch of other hackers about creating an auto-syncing, auto-connecting, self-healing WiFi mesh network

The air-dropped node idea came out (wasn't my idea originally) and it seemed the most feasible in a disaster area where human intervention wasn't possible

Basically, you can drop a few of these which will link up to each other and provide an ad-hoc hotspot for others. The node closest to the internet backbone will provide backhaul

@cypnk so reestablishing communication to the outside world during a disaster?

@cypnk well as you said it would also create a network inside ground zero for communication between those affected. Its an interesting idea.

@jordan31 Yes, and even if for some reason the solar panels don't deploy well, or if there's obstruction or some other issue, the design is such that it can stay powered for at least a few days

@cypnk make it hamster powered lol. Or use thermal energy. Though that would have some real pains to work through on using it safely, and would be expensive.

@jordan31 Yup, exactly

And the second reason is that a lot of places that can benefit from internet are also impoverished. I've seen copper theft happen in a lot of places where Internet is newly installed. A mesh node is cheap to replace and there's nothing to dig

@cypnk wouldn't sat still be best for this since they are outside of the area, and can handle communication within and to outside the area? Though they would be expensive to launch unless they used something cheap like cube SATs.

@jordan31 For long-term, absolutely. I'm really hoping SpaceX Starlink becomes a thing soon. I think the costs will come down after mass-production takes hold

For short-term connectivity where no receiver is available and the time needed to launch a sat isn't practical, these can be deployed within 24 hours at most

@cypnk @jordan31
Cool idea. I've wondered if mesh networking could provide a benefit in my area; DSL in many places, nothing in others, places with rough terrain and relatively few people.

Something like Scuttlebutt (scuttlebutt.nz/) appeals, where it doesn't need regular, real-time connectivity. Get a message out, store-and-forward style, occasionally by sneakernet.

Maybe you'd want some sort of gateway to the world, opportunistically, not aimed at browsing and Netflix.

@cypnk
Raspi or orange pi zeros are great candidates for that .. if airdrop is not a requirement you can add some 433Mhz yagiis pointing at each other for long communication support ;)

@gled I settled on a dedicated travel router: GL-MT300N (V.2). Air-drop is absolutely essential for the "Air-dropped Mesh Node" ;)

The biggest reason was power consumption. This has to run for days on batteries if no sun is available and the Pi draws way too much current even when idle

Also the MT300 supports OpenWRT natively, but the Orange Pi uses a much larger image that may cause issues in the longer run. I can also reconfigure its properties via over-the-air update

@cypnk
'Idle' is less than 500mA for those boards but depends on the need and where you are for sure. I have a bunch of both on carefully selected treetops running for quite a while now, hence the answer ;)

@cypnk you could do precision landing if you use an RC paraglider
That could be a great sewing project
:D

Sign in to participate in the conversation
Mastodon

Server run by the main developers of the project 🐘 It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!