search

p-v-o-s / pioneer-valley-open-science.github.com

9 stars 2 forks source link

LEAFFEST (Low Elevation Aerial Flights For Earth Sensing Technology) Meetup -- September 22nd, 2012 #11

Open dwblair opened 12 years ago

dwblair commented 12 years ago

(Click here to navigate back to the PVOS wiki home page)

This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License

Image

Background

For the official LEAFFEST announcement and to get on the participant list, see http://publiclaboratory.org/notes/cfastie/9-5-2012/l-e-f-f-e-s-t

First draft of LEAFFEST general schedule. Then at the bottom, click "Google Calendar" to open it in for real. On "Four Day" view, navigate to September 21.

Topics for LEAFFEST

Aerial Photography

Fastie commented 12 years ago

For the official announcement and to get on the participant list, see http://publiclaboratory.org/notes/cfastie/9-5-2012/l-e-f-f-e-s-t

dwblair commented 12 years ago

Dear PVOS-ers:

What would we like to prepare for the Sept 22nd meeting?

Based on recent email threads, I'm going to put this out there:

  1. Hydrogen balloon inflation in the field using e.g. garbage bags, water, sodium hydroxide, and aluminum foil. Goal: enough lift to carry a typical aerial photography load. (Let's ask Chris F. what that weight would be).
  2. Xbee communication over a 1000+ foot line-of sight.
  3. A way of monitoring battery level or camera operation (as per Matt's idea).
  4. Dream: Combining 1,2, and 3: a camera based on a Hydrogen balloon, with battery level monitored on the ground.
  5. Ideas and background for H2S sensing

Good stuff? Others?

I sense that Matt is pretty close to being able to do 2 & 3. I feel I'm on my way to doing 1. Shall we keep track of our progress in this thread? And then post to PLOTS once we have something somewhat coherent?

dwblair commented 12 years ago

Oh, also forgot: an Old FART demo of an aerial camera operation eavesdropping device, using the walkie talkie listen-in-on-camera-operation method.

I have some photos from my first (failed) attempt to make a garbage bag Hydrodgen balloon that I'll be posting on Flickr soon. Will link here, and then put on PLOTS too.

mbockmann commented 12 years ago

I think we'll get 100 meters line of sight communication with my 2 mW Xbees - so I guess about 300 ft? That's shy of our 1000 feet goal. I know you can get more powerful radios, does anyone know whats good for 1000+ feet? Maybe there could be a good discussion on what antenna type would be best too.

dwblair commented 12 years ago

Matt -- is the difference between 300 ft and 1000 ft simply a matter of purchasing a more-powerful flavor of the XBee radio / antenna, but otherwise the design and protocol for the setup and usage are the same? If that's the case, then I think you don't even need to worry too much about the issue for the sake of LEAFFEST; being able to demo this solution over 300 feet, and then noting that one simply needs to, e.g., "swap out antenna model XJ3001 with antenna model XJ3002", is already really great, I think!

dwblair commented 12 years ago

Okay, the new list of PVOS demos (so far as I understand):

-- Jeenode compost pile temp monitoring -- Don -- Jeenode pressure + temperature KAP monitoring -- wireless + SD card -- Don -- Xbee transmission of data, 300 ft -- Matt -- Xbee notification of battery status (possibly) -- Matt -- (Matt -- maybe we can use your battery-low device with the Jeenode, too?) -- Raspberry PI + webcam + WiFi video transmission demo (possibly) -- Craig -- Hydrogen generation via a) sodium hydroxide + Al and b) electrolysis (pending safety concerns) -- Don, Craig -- Hydrogen balloon using garbage bags (pending safety concerns) -- Don, Graig -- Some nefarious combination of all of the above as a warm-up to controlling Earth's electrical grid.

Additions / subtractions / modifications?

dwblair commented 12 years ago

Also: in order to write values to an SD card for the airborne Jeenode demo, I'm going to buy at least one OpenLog from Sparkfun. Does anyone else want to go in on an order from SparkFun for LEAFFEST (via expedited shipping)?

Fastie commented 12 years ago

-- Jeenode compost pile temp monitoring -- Don

Will you have components that can be dedicated to this for the weekend? If so, we should stream it live to Cosm.com as a demo. Get here early on Friday and we can turn the pile and emplace the thermistor and then the world can watch the result of millions of bacterial cell divisions.

-- Jeenode pressure + temperature KAP monitoring -- wireless + SD card -- Don

Keeping track: this is one instrument package that could fly. Which sensors do you have? Should I buy a humidity sensor on your Sparkfun order?

-- Xbee transmission of data, 300 ft -- Matt

This might be a ground based demo? Or is there some aerial data to transmit?

-- Xbee notification of battery status (possibly) -- Matt

Matt, did you get the battery detector working? This might be a ground based demo?

-- (Matt -- maybe we can use your battery-low device with the Jeenode, too?)

sounds like an experiment to try.

-- Raspberry PI + webcam + WiFi video transmission demo (possibly) -- Craig

Wow, this will be a live video downlink? Can't wait. This would be a good one to fly. We could mount this on a standard rig with a Powershot, or just send it up by itself for testing.

-- Hydrogen generation via a) sodium hydroxide + Al and b) electrolysis (pending safety concerns) -- Don, Craig

There is a time slot open for this on the evening of the bonfire.

-- Hydrogen balloon using garbage bags (pending safety concerns) -- Don, Graig

I have a bunch of Mylar sleeping bag covers (huge space blanket condoms) that might be bigger and hold the H2 longer than garbage bags.

-- Some nefarious combination of all of the above as a warm-up to controlling Earth's electrical grid.

Forget the warm-up, let's do it.

So maybe just two or three payloads that should be flight tested? It would be great to post a research note or two at PLOTS this week of prototype/planning documents like the remote temperature compost note but for other payloads. All you need is one illustrative image for the "main image" and some notes about the plan. Don, did you forget Old-FART again? Or have you moved on to the next thing? :)

mbockmann commented 12 years ago

Wow lots of exciting things happening!

Re [Chris] ":-- Xbee notification of battery status (possibly) -- Matt Matt, did you get the battery detector working? This might be a ground based demo?" [Matt reply] The battery detector works around 3V (it's a little finicky), but not much higher than that. I need to get a better understanding of the IC, tweak it around some more, and consider other ideas like a very tiny battery and a comparator. I'll have a ground based demo for Leaffest.

Re [Don]: "-- Xbee notification of battery status (possibly) -- Matt -- (Matt -- maybe we can use your battery-low device with the Jeenode, too?)" [Matt reply] Yes I think we should consider this too. If you have a microcontroller on the kite or some means to sense and send data without an additional microcontroller I think we can do it.

dwblair commented 12 years ago

Going to reply in depth to above, but quickly:

-- Matt -- do you happen to know what electronics I should buy (from e.g. Digikey) to allow for multiplexing several thermistor measurements? Just so I can put in an order tonight / tomo ...

-- Chris -- re: Sparkfun order -- humidity sensor already gotten (I have 10 that were intended for another project).

mbockmann commented 12 years ago

Hi Don, I think I led you astray with what I said in my truck. I went back and looked at the drawing I worked on for multiple analog inputs with the multimeter, and it called for 1 BJT transistor and 1 reed relay per input. I've had good luck with this SPST reed relay # 306-1062-ND on digikey, however what I had breadboarded was more focused on output, and I did not spec the correct Demux chip. Basically the idea is to select which relay you want with the demux, and energize the correct relay with a BJT. I think I have the basic idea down, but debugging would still be necessary. Sorry!

dwblair commented 12 years ago

No Matt, your idea is fantastic -- no worries if there will be debugging involved! We might not get multiplexing to work before LEAFFEST (because my background in this stuff is so poor), but we can at least demo 4 sensors on a JeeNode this weekend, then design an expansion to 12 on our own afterwards that we can e.g. publish to PLOTS.

However: does it make sense to order the stuff you mentioned:

before LEAFFEST, so we can prototype when we're up there all together? Or are there more details to figure out before we order?

On Sun, Sep 16, 2012 at 8:27 PM, Matthew notifications@github.com wrote:

Hi Don, I think I led you astray with what I said in my truck. I went back and looked at the drawing I worked on for multiple analog inputs with the multimeter, and it called for 1 BJT transistor and 1 reed relay per input. I've had good luck with this SPST reed relay # 306-1062-NDhttp://www.digikey.com/product-search/en?x=11&y=21&lang=en&site=us&KeyWords=306-1062-NDon digikey, however what I had breadboarded was more focused on output, and I did not spec the correct Demux chip. Basically the idea is to select which relay you want with the demux, and energize the correct relay with a BJT. I think I have the basic idea down, but debugging would still be necessary. Sorry!

— Reply to this email directly or view it on GitHubhttps://github.com/Pioneer-Valley-Open-Science/pioneer-valley-open-science.github.com/issues/11#issuecomment-8602064.

voice / SMS: +1-651-252-4765 skype: dwingateb

dwblair commented 12 years ago

Yo! Sorry for the delay in answering some of the prior points ...

Will you have components that can be dedicated to this for the weekend? If so, we should stream it live to Cosm.com as a demo. Get here early on Friday and we can turn the pile and emplace the thermistor and then the world can watch the result of millions of bacterial cell divisions.

I believe I'll have the proper components for this by Friday, yep! For four sensors, at least.

Keeping track: this is one instrument package that could fly. Which sensors do you have? Should I buy a humidity sensor on your Sparkfun order?

Humidity sensors, I'll bring -- meanwhile, I'm about to put in my order for the SD card logger. If there's something you'd like to buy from Sparkfun (e.g., your first Arduino?), lemme know (I'm thinking of putting the order in tonight).

I have a bunch of Mylar sleeping bag covers (huge space blanket condoms) that might be bigger and hold the H2 longer than garbage bags.

Perfect! Craig was looking into finding some mylar balloon material ... is there a place to buy these cheaply online? In any case, if you have several, that's great. Part of the hydrogen balloon idea is the fantasy of creating a kit that would e.g. be useful in Alaska -- enormous space blanket condom, aluminum foil, sodium hyrdoxide, water, and you've got a balloon for carrying a payload.

So maybe just two or three payloads that should be flight tested? It would be great to post a research note or two at PLOTS this week of prototype/planning documents like the remote temperature compost note but for other payloads. All you need is one illustrative image for the "main image" and some notes about the plan. Don, did you forget Old-FART again? Or have you moved on to the next thing? :)

Re: the PLOTS note, I'll try to put something together. Haven't forgotten about Old-FART (though forgot to mention it -- sadly, seems like the Old FARTs are often eventually overlooked / fade away ... do not go silent (but deadly) into that good night! and etc ... )

Fastie commented 12 years ago

I have a few of the four pack Mylar bags 84"x36" so we have plenty for LEAFFEST. You might have to fill three or four of these to lift more than a little camera, but I have never done it so I'm not sure.

Math update: A five foot diameter balloon holds about 65 cu ft and barely lifted a minimal two-camera rig at Lee, NH. One Mylar bag holds about 14 cu ft. so four of them might lift this rig. That's a lot of gas.

The only helium supplier I could find in Middlebury is fresh out and won't have any for a few weeks.

The current forecast is for very light winds this weekend, but the rain is supposed to hold off until Sunday.

dwblair commented 12 years ago

The only helium supplier I could find in Middlebury is fresh out and won't have any for a few weeks. The current forecast is for very light winds this weekend

HYDROGEN FTW!

cversek commented 12 years ago

Just heard a segment about the helium shortage affecting party stores on NPR's Marketplace. These trying times call for drastic measures.... hydrogen party balloons.... well they said it was going to be blast!

mbockmann commented 12 years ago

Wow those Mylar bags will look awesome when they float.

Re Don on parts for multiplexing: I have plenty of BJTs, and I should be able to dig up a couple of SPST relays. I don't have a good demux, but I may have a shift register on hand. Either should work now that I think about it. I'm not sure which would need less pins.

dwblair commented 12 years ago

The logistics for all of this are getting fun. I'm going to need a new pen.

On Mon, Sep 17, 2012 at 9:23 PM, Matthew notifications@github.com wrote:

Wow those Mylar bags will look awesome when they float.

Re Don on parts for multiplexing: I have plenty of BJTs, and I should be able to dig up a couple of SPST relays. I don't have a good demux, but I may have a shift register on hand. Either should work now that I think about it. I'm not sure which would need less pins.

— Reply to this email directly or view it on GitHubhttps://github.com/Pioneer-Valley-Open-Science/pioneer-valley-open-science.github.com/issues/11#issuecomment-8636719.

voice / SMS: +1-651-252-4765 skype: dwingateb

mbockmann commented 12 years ago

I think I'll have the resistors and BJTs needed, but only a couple SPST relays so only a couple of thermistors can be done with what I have in mind unless we get more relays. I need to draw it out, I hope someone can improve on the idea to use less relays. I need 1 SPST relay per input.

dwblair commented 12 years ago

Matt -- cool! -- can I quickly order a bunch of SPST relays from digikey, and get it before Thurs? Do you know what to look for on digikey?

On Mon, Sep 17, 2012 at 9:43 PM, Matthew notifications@github.com wrote:

I think I'll have the resistors and BJTs needed, but only a couple SPST relays so only a couple of thermistors can be done with what I have in mind unless we get more relays. I need to draw it out, I hope someone can improve on the idea to use less relays. I need 1 SPST relay per input.

— Reply to this email directly or view it on GitHubhttps://github.com/Pioneer-Valley-Open-Science/pioneer-valley-open-science.github.com/issues/11#issuecomment-8637038.

voice / SMS: +1-651-252-4765 skype: dwingateb

cversek commented 12 years ago

Don -- The electrolysis (water splitting) method is safer but trickier, than the aluminum hydrolysis reaction. I'm going to run some tests tomorrow...powered by my car's cigarette lighter (with engine idling, so I don't need to get a jump again, \wink). For a given set of electrodes plus electrolyte we need an idea of the typical running current. This current will be proportional to the amount of H2 gas produced per unit time, ala Faraday's law of electrolysis. I'm going to try the carbon arc electrodes with sodium or potassium hydroxide in the water for added conductivity (depends what the hardware store has).

Chris -- The aluminized mylar bags sound like they might work...but I would imagine they're a bit thicker than optimal from a weight perspective. Do you have an estimate of what weight your camera rig is? A very smart guy told me that hydrogen can lift ~15X it's own weight; so with the info from the above experiment...maybe we can estimate how long it will take fill. BTW I am working on a Raspberry Pi wireless webcam rig that hopefully will be less than 300g.

Fastie commented 12 years ago

Craig,

The dual camera NIR/VIS mapping rig weighs 415g with everything aboard. It’s shown here without electronics. We lifted this with a five foot helium balloon that could have been a bit more full. NIR/VIS KAP rig from top

My pan/tilt rig with S95 and all electronics as shown here weighs 590g. This is easy to lift when there is a 10 mph wind. Without the camera this weighs 400g. SunOrchRig-20120429-133-8

I have another minimal picavet rig which weighs 150g with no payload. Add your Pi payload and it will be flyable in an 8 to 10 mph wind or with 60 to 90cubic feet of H2 or He.

Each Mylar sleeping bag cover weighs 82g.

I updated the bad math in my comment above: A five foot diameter balloon holds about 65 cubic feet and barely lifted a minimal two-camera rig at Lee, NH (although the balloon could have held more He). One Mylar bag holds about 14 cubic feet, so four of them might lift a camera rig.

Fastie commented 12 years ago

Here are three Mylar sleeping bags lifting a small camera in a soda bottle rig: https://plus.google.com/photos/111415749895648787719/albums/5544785685010575201/5544785865953576146?banner=pwa

Other photos there are of Jeff Warren' workshop at the Gigapixel conference at CMU two years ago.

cversek commented 12 years ago

Here's the good news... we can produce small amounts of hydrogen gas safely using electrolysis (with sodium carbonate "washing soda" electrolyte) and an old laptop power supply drawing 5 amps max. And I have a fun demo planned for this.

Here's the bad news, a quick calculation using Faradays' law of electrolysis and the rule that H2 can lift 15X its mass for this power supply says we would need ~90 hours to make enough gas to lift a payload of 500g. Even if we boost the power using, say a desktop computer supply, it would still take unbearably long and the heat load could become dangerous.

If there is to be any lifting gas produced... it looks like we are back to Al + H2O in strong base. A reasonable reactor set up and safety precautions should make this safe enough for those who know what kind of dangers abound.

mbockmann commented 12 years ago

What rate of hydrogen generation did you calculate with electrolysis?

mbockmann commented 12 years ago

....5.6 grams an hour. I should think before I type :)

cversek commented 12 years ago

actually the H2 mass 1/15th of that....what you calculated is the maximum payload mass that we can lift given an hour of gas generation with that 5A current source

cversek commented 12 years ago

I just got a live webcam feed from my Raspberry Pi working. It is set up as a wireless access point using a cheap USB dongle. Then anyone with a standard wireless interface (and the secret password) can connect then view the feed in their browser! I have a battery pack that should last for 5 hours, hopefully we can get something into the air this weekend.

BTW, I followed the instructions on this guy's blog and combined it with the first step of this post. Note, compiling large source packages like ffmpeg on the RPi is not for the impatient - though easy to set up, it took over 6 hours to compile - but is well worth the investment.

Don - We are now so close to that LEADzeppelin prototype!

dwblair commented 12 years ago

Holy COW. I'm getting back online by tonight 630 - Craig can you gather photos / vids of all of this awesomeness (hydrogen, raspberry pi) for a brief pre-leaffest PLOTS research note? As Chris suggested, it'll be a nice progress report, as well as a preview / advertisement re: leaffest ... dude, you're on FIRE! On Sep 19, 2012 2:43 PM, "Craig Versek" notifications@github.com wrote:

I just got a live webcam feed from my Raspberry Pi working. It is set up as a wireless access point using a cheap USB dongle. Then anyone with a standard wireless interface (and the secret password) can connect then view the feed in their browser! I have a battery pack that should last for 5 hours, hopefully we can get something into the air this weekend.

Don - We are now so close to that LEADzeppelin prototype!

— Reply to this email directly or view it on GitHubhttps://github.com/Pioneer-Valley-Open-Science/pioneer-valley-open-science.github.com/issues/11#issuecomment-8701749.

Fastie commented 12 years ago

Craig, nicely done. That sounds like it will have all sorts of possible applications.

About how much does the Pi + WiFi dongle + battery weigh? What battery are you using? What range would you expect for transmission? Can I have one? Can we pick up the WiFi stream and make it available to anyone with the correct URL?

The forecast is changing by the hour, but the rain is now mostly during Saturday night, and more than 10 mph wind is predicted for part of both days.

dwblair commented 12 years ago

Craig is documenting the setup as we speak; but meanwhile, I think I can answer some of Chris's questions:

What battery are you using?

Craig is using the battery pack recommended on the Adafruit website. He has some opinions about it. Maybe there will be other suggestions for lightweight battery packs this weekend?

Can I have one?

Bien sur. Everything Craig is using is available on Adafruit.com, I believe. Should we put together a shopping list?

Can we pick up the WiFi stream and make it available to anyone with the correct URL?

Yes -- including web-enabled mobile devices. At this point the prototype is streaming rather choppy, live video (but is still a little buggy -- sometimes the web serer craps out -- that's something that can be fixed with enough tinkering); it might be even more stable to take still snapshots and have them available online.

Conceptually, what Craig has put together is the equivalent of having a small laptop in a lightweight package -- this goes beyond what an Arduino could do. Pretty much anything that a linux laptop can do (in particular, a Debian linux laptop), this little baby can do. Apparently C was able to cobble together some available packages quite readily into his current prototype. The wireless is being transmitted via a USB wireless dongle of the sort that you'd plug into a laptop in order to provide it with a wireless connection to local WiFi.

Craig: one could also use, for example, a 3G/4G wireless dongle of the sort sold by Sprint as a way of getting a laptop online "anywhere", using the cell network; wouldn't that make the platform able to provide wireless imagery from wherever there is a 3G/4G signal? BOOM.

cversek commented 12 years ago

Here's an image of the prototype (much to left be optimized): Photobucket

weight: I would estimate ~300 grams, I'll get a better figure when I'm back in Amherst

Don pretty much covered what I would say. Rather than muck around with low-level circuitry and limited hardware options, there are a ton of USB capable devices many of which are plug and play. The webcam and wireless dongle were two things that I just had lying around, not purchased specially for this project. We could even write complex control software using an agile language like python and hook into many other open source projects. With Linux in the sky ...the sky's the limit.

Yeah, I guess 3G/4G could be a premium option...never used one of these dongles directly, so I don't know what the linux support is like.

mbthorn commented 12 years ago

Axiom: I don't really know what I am talking about

I'm hoping this may help explain why the video is choppy. It may be because of the wi-fi transmission speed, but I doubt it. I think the more likely culprit is asking that poor little raspberry pi to decode/encode the video from the webcam for transmission. That is a fairly intensive task unless you have a chip specifically designed for the task. This may give you an option for improvement. If you can buy a chip that encodes/decodes video over USB you'd unload the PI a bit.

In this case you may actually be better off not using software for the task. I think it is why these newfangled smart phones can send a video to a 32" TV and my fancy laptop heats up like a hotplate.

On Sep 19, 2012, at 8:43 PM, Craig Versek notifications@github.com wrote:

Don pretty much covered what I would say. Rather than muck around with low-level circuitry and limited hardware options, there are a ton of USB capable devices many of which are plug and play. The webcam and wireless dongle were two things that I just had lying around, not purchased specially for this project. We could even write complex control software using an agile language like python and hook into many other open source projects. With Linux in the sky ...the sky's the limit.

cversek commented 12 years ago

Good point about the encoding, this is definitely going on. But I'm not sure how one would go about adding and encoder without respinning the board; there are general purpose IO pins, but I am not sure if they can achieve a high bus rate - maybe worth looking into though? Apparently some of the fancier webcams have encoders built in, hence you can directly stream in mjpeg format. At any rate, this is just a demo of the platform and we are probably not going to do any great science with it yet. Doing a highest resolution time lapse image that is cached to disk and hosted via the Apache server might be a better strategy in the long run. I just slapped this together because someone had relatively straight-forward instructions that I could follow. We can save the web-dev for a future hackathon. I more interested in seeing if we can loft this thing with hydrogen and get good connectivity while it's 1000ft up (my guess is that the weak wireless dongle will crap out much lower than that). BTW, a more accurate weight is 210g.

On Fri, Sep 21, 2012 at 8:55 AM, mbthorn notifications@github.com wrote:

Axiom: I don't really know what I am talking about

I'm hoping this may help explain why the video is choppy. It may be because of the wi-fi transmission speed, but I doubt it. I think the more likely culprit is asking that poor little raspberry pi to decode/encode the video from the webcam for transmission. That is a fairly intensive task unless you have a chip specifically designed for the task. This may give you an option for improvement. If you can buy a chip that encodes/decodes video over USB you'd unload the PI a bit.

In this case you may actually be better off not using software for the task. I think it is why these newfangled smart phones can send a video to a 32" TV and my fancy laptop heats up like a hotplate.

On Sep 19, 2012, at 8:43 PM, Craig Versek notifications@github.com wrote:

Don pretty much covered what I would say. Rather than muck around with low-level circuitry and limited hardware options, there are a ton of USB capable devices many of which are plug and play. The webcam and wireless dongle were two things that I just had lying around, not purchased specially for this project. We could even write complex control software using an agile language like python and hook into many other open source projects. With Linux in the sky ...the sky's the limit.

— Reply to this email directly or view it on GitHubhttps://github.com/Pioneer-Valley-Open-Science/pioneer-valley-open-science.github.com/issues/11#issuecomment-8763596.

mbthorn commented 12 years ago

Yeah, your probably right. I'm not sure what your specific USB wi-fi unit will do, but I know mine typically craps out at about 50-60' (through walls though). I bet outside will be hundreds of feet, especially if it is an N type. I'm pretty sure you can get boosters for wi-fi signals up to a couple miles… though the FCC might not like it. How about a modem over walki-talki Don's working on… could sound old school--hipster hacking 101? That last suggestion is, of course, a joke.

On Sep 21, 2012, at 10:06 AM, Craig Versek notifications@github.com wrote:

I more interested in seeing if we can loft this thing with hydrogen and get good connectivity while it's 1000ft up (my guess is that the weak wireless dongle will crap out much lower than that).