Shared publicly  - 
Since getting back from New Zealand, the Project Loon team has been busy conducting a series of research flights in California’s Central Valley. The purpose of these flights is to allow us to research various approaches for improving the technology, like the power systems (solar panel orientation and batteries), envelope design, and radio configuration. We’ll be doing a series of posts on what we learn from these flights, and how the technology develops as a result.

On our most recent research flight we overflew Fresno, a nearby city, to get statistics on how the presence of lots of other radio signals (signal-noise) in cities affects our ability to transmit Internet. It turns out that providing Internet access to a busy city is hard because there are already many other radio signals around, and the balloons’ antennas pick up a lot of that extra noise. This increases the error-rate in decoding the Loon signal, so the signal has to be transmitted multiple times, decreasing the effective bandwidth.

This is like trying to talk to a friend at a loud concert. The the music interferes with your voice, so your friend might have to ask you to repeat what you said a few times in order to make sure she heard it correctly. This will result in a more basic conversation; instead of speaking about complicated topics in depth, you’ll have to spend a lot of your time repeating yourself just to make sure your friend can understand you!

One way to deal with this is to speak louder, or in the case of a Loon balloon, to increase the signal strength. Our tests over Fresno will help us understand how signal-noise interferes with our signals, so we can determine how strong we’ll need to make our signal in order to transmit it effectively. 
sabrina alam's profile photoSuzan GENCER's profile photoRubén RC Soto's profile photosyed umer ali's profile photo
We've seen the same thing on our balloons. As you get higher the horizon "moves farther" away and you see more signal sources. That increases the noise floor.
T Wong
How do mobile phone system deal with similar problems? The radio spectrum is equally noisy and the power output from the cellphone are very tiny, yet the cell tower can distinguish individual signals from many hundred mobile phone connections simultaneously. 
I need to. read it again. In the morning to fully understand the implications, forgive me it's 03:40 hrs and I'm not alert enough to give ig the attention it deserves . Saying it is amazing isn't enough at all!
Loon in California wouldn't be crazy. I live 45 minutes from Google HQ. People in my neighborhood can't get broadband. 
If you increase the signal strength, don't you reduce the up time of each balloon? 
+T Wong Actually, the cellular spectrum is very quiet compared to the unlicensed ISM bands.

At any given location, for any given frequency, one (and only one) carrier owns the airwaves on the licensed bands that cell service runs on. As such, they can coordinate to eliminate self-interference, decreasing the noise floor and thus increasing throughput.

By contrast, 2.4GHz and even 5.8 GHz are a noisy mess, because there's no telling who, when and where from transmissions are happening. So throughput decreases significantly.
+Project Loon I feel like I missed something here. I was under the impression that circumstances in which you had an elevated noise floor, i.e. adequate population (and device) density, there are more cost effective means of providing connections. Is this just an interesting aside, or are there actually circumstances where you'd consider providing service via Loon to a high population density area?
Where graft in a city made WiFi by telco too expensive for poor people to use.
So I'm reading endgadget this evening and I see googles project loon article and I always love reading about what googles doing.. And hey whaddaya know.. They're in my home town launching balloons. Awesome.. Need any help from a Big nerd fan?
Keep plugging away guys. This is cool.
+Narada Bradman
You're right that we're developing Project Loon primarily to provide access to rural and remote areas that are not currently connected. That said, we think there are interesting potential use-cases when it comes to disaster recovery, which could involve more urban settings. Ultimately, we'd like to build a service that works well in a wide variety of circumstances, and solving the signal-noise issues is a necessary part of that.
Google putting Fresno on the map! Do more there please. Great city but enjoying none of the prosperity of other parts of California.
Don't do too much.. I don't want to bring the cost of living up too
Seriously, will no one from +Project Loon justify this extravagant waste of helium, a limited resource that we are expected to run out of in the next few decades?
Run out of helium?  BS on that, that's like saying oxygen is a limited supply.  Drill for natural gas and you get helium as well.  Now if we let the Environmental Nazis continue to limit anyone from drilling for oil and natural gas you may have a point.  +Project Loon must continue so that everyone can have internet at reasonable prices everywhere including here in the USA.
+Barry Davis except oxygen is a heavy gas and doesn't leak off the planet. Helium is too light and every bit we leak into the atmosphere eventually disperses off the top of the atmosphere into space. And we can't manufacture He. We can only mine it, and when we run out of natural gas we run out of He. You think running out of fuel will suck? We can use other things as energy sources. We can't use anything else as an alternative for He. Filling up balloons so they float is a stupid waste of He, when there are plenty of other ways of making things float and fly.
+Jeremy Taylor you sound like you're loads of fun to be around. All gloom and doom. I've been hearing this "We are running out of energy " BS since the 1970s. The truth is we have enough energy to last at current consumption levels to last us longer than we will be on the planet. Whatever they use after that is their problem. Quit throwing cold water on +Project Loon , nobody likes a party pooper. 
+Barry Davis I didn't say we were running out of energy, in fact I said we had alternatives in that respect. We don't have alternatives to He as nothing else has it's unique properties. +Project Loon has noble ambitions to make wireless internet ubiquitous, but surely there are alternative ways of achieving that which don't involve consumption and dispersal of He? Think about this: He is currently stockpiled and is being sold very cheaply, below it's real value. That is what makes +Project Loon viable. As supplies dwindle, the economics will force the price of He up. Then what is going to happen to your free wireless internet?

We already have a low-cost wireless internet thanks to cellphone providers anyway. If someone wants wireless internet where there isn't any currently, the cost of installing a cell site can be as little as $10,000. Imagine how many cell sites in remote areas could be installed for the development cost of Project Loon?

You call me a party pooper. Hey, I like cool stuff as much as the next guy, and I can see why people think this is cool - but occasionally it is worth looking at the proposal from alternative perspectives, to make sure we have things in perspective.

I'll tell you what. You keep cheering for the king in his new lovely finery, and I'll keep on suggesting that maybe he hasn't got any clothes on.
I hardly call cellular internet broadband or anything near cheap. It must be nice living where you get cheap or free broadband, but the rest of us out here in flyover country can't wardrive off our neighbors open internet connection because he has the same problem we do.
Let's just build a death Star and be done with it.. That should be googles next focus!
+Barry Davis it is sufficiently cheap. Here in NZ we are known for having "expensive" and not particularly fast Internet, but at US$80 for 12GB of 3G data, or 1GB when roamed onto another network, it is sufficiently cheap to be productive on the move.

I wonder what free wireless Internet would be like. Tens of thousands of people simultaneously hitting the same node as it flys over... I have a flashback to the bad old days of all-you-can-eat dialup data plans.

Yeah, sure, I'd forgo any chance of an MRI in my old age to be able to get that experience back.
Wonderful work.  Please check out Sierra County, California and specifically, the Yuba River Canyon area with local communities of Downieville, Sierra City, Allegeny, and more.  Their extreme remoteness, tough internet access, and VERY spotty cell reception has contributed to schools closing, families moving away, and a lack of growth.
Hi Michael.  Perhaps a trip up our way to take a look at the situation first-hand is the first step.  Sierra County is a gorgeous area with high buttes and mountains, alpine lakes, and lush valleys.  It's a great place to take a little weekend vacation and you can get a sense of the opportunities and challenges we face with accessibility.  
Did you try using directional antennas? Or does the mobility and requirements to have both balloon-to-balloon, and balloon-to-ground links render that infeasible?

Are your other interference sources in the ISM band wide-band/narrow-band? Perhaps having additional (tuneable) analog filters on the WiFi front end, might help.
+Ian Littman , You can always use other spectrum available like UNII in the 5.2 and 5.4 Ghz bands, its still unlicensed but should be less noisy.  Another alternative would be the semi-licensed 3.65 ghz band.. 
The implementation of this project is a bold affront to conservation of natural resources (Helium)  when a cheaper and far more plentiful alternative is readily available, i.e. Hydrogen. No mention of Helium recovery is made because this project has no means to do it - it's completely diffused and irrecoverable.

Hydrogen is superior for this application - for example, weather balloons often use hydrogen due to cost, with side benefits being lower density and lower diffusion rates across the membrane; flammability is not a major concern.

I'm guessing the only things preventing its usage here are the bad optics associated with the Hindenburg, which forever stigmatized hydrogen when there's any need to get the public on board with something, even when it's used in an innocuous component of the project (balloons to hold TX/RX nodes), i.e. these are like weather balloons, which again use hydrogen and pose no public threat. 

I'm sure a legal team advised that it's worth an added expense to go with helium and avoid some bad optics, because that's all that's in play - not public safety, just Google's image.

Somewhat ironically, you can use Google to discover why there is a shortage of helium, and the many irreplaceable usages of it, including MRIs, superconducting magnets, industrial and medical CO2 lasers, and... oh, a little something called manufacturing of semiconductors -
something I'd imagine a Silicon Valley tech giant might be interested in preserving if nothing else.

Learn how the BLM manages reserve stock that is dangerously depleted, and how it's the main source of all commercial acquisitions of the element since WWII. 

Note that you can't recover helium from the atmosphere once it's diffused out of a balloon, the atmosphere simply doesn't retain it and it disperses into space.  The only known earthly means of producing helium is via incredibly slow nuclear decay that occurs in some natural gas wells over time frames that make millennia seem like minutes.

Seriously, this a disastrously wasteful project considering how many tons of helium Google is willing blow off the face of the earth on a continual basis to look safe and friendly in the eyes of the public and to really "hard sell" this idea to them, i.e. you and me.
Add a comment...