First of all I’d like to thank Outernet project creators for getting people interested in radio, satellites and out of the box thinking. It is a very cool project to DIY (or should I call it “assemble it yourself”?) and it’s a perfect “gateway drug”.
Short version of the story: fully autonomous receiver is a very expensive project, it’s going to cost you hundreds dollars if not more, so for now I decided to focus on better reception giving up the autonomy idea. Below I list the challenges I’ve met, perhaps somebody could share their experience solving them.
Now the long version.
I got interested in radio in general and SDR in particular recently and thanks to RTL-SDR blog learned about Outernet project. I decided to make fully autonomous receiver because it sounded like a very cool idea.
I’ve ordered a DIY kit without a power bank (because I’ve already had two), a waterproof case and a 20W waterproof solar panel.
My findings so far:
The receiver works 5-6 hours on a 10 000 mAh power bank. Therefore if you want your receiver to be autonomous and you’re not living too close to the poles of the Earth, you’ll need at least twice as big power bank or multiple power banks. I think that the biggest power bank RAVpower sells (it’s 22 000 mAh) is still slightly below adequate capacity. To be on the safe side you need two such power banks, so it won’t be very small or light, but the price is reasonable (below $100 for two).
I’ve read that in cloudy weather solar panels work at 10%-25% of their maximum power rating. In reality 20W solar panel can’t charge any power bank or phone I’ve tried in cloudy winter weather, even if it looks like it’s bright enough outside. It shows the charging sign but the current is very low so it doesn’t charge at all (or charging so slowly that it doesn’t make any difference). Perhaps it can charge a power bank specifically designed to work with low power solar panel but I’m not sure about that. I’ve bought a USB dongle to measure current, I’ll post my findings as soon as it arrives. Unless I have a faulty solar panel, at the moment I think that you need at least 200W solar panel to fully charge a small (20000 mAh) power bank in a cloudy day. Such panel is going to be pretty heavy, big and expensive (few hundred dollars). It’s probably better to give up on solar-charging solution and better think about big enough power banks that you can swap and charge in some other ways.
Cheap ($15) waterproof case (made of polycarbonate) works excellent. Rain or snow, no problem at all. Can be a problem in sunny summer weather though - the RTL-SDR is getting really hot even in sunny winter weather. Perhaps a white colored case would be ideal, but the clear one is OK for occasional use. A black one is probably going to get really hot. With $2 stand for a tablet from dollar store it’s really easy to point it where you want.
Reception outdoors (far away from buildings) is not a problem at all, pointing is easy, the reception is good enough in any weather. I’ve tried it when it was raining with real heavy clouds on the sky and I was getting 5-7 dB SNR (signal to noise ratio). In my area reception indoors or close to a building is pretty much impossible for me. I’ve made a small research and learned that there is so much RF noise indoors in all frequencies and particularly 1540 MHz range that it makes reception impossible. I thought that the unshielded LNA is the problem but my tests proved that all the noise is coming through the antenna, so there is really no way to filter it out. What is worse, that my RTL-SDR dongle (a shielded V.3 version) still “receives” something without antenna attached. Surprisingly, the shielding completely removes broadcast FM but doesn’t remove strong signals in higher frequencies (like 900 MHz range). Some of that could be internal noise, but some is clearly GSM and pager signals. Not sure how this problem can be solved. I still have to investigate how much noise is coming through the USB port, but anyway it’s silly to expect $10000 performance from $30 dongle. I’m going to try a FM band-stop filter to see if it affects anything at all. It would be nice to have a band-filter just for L-band. Just out of curiosity I’ve compared RTL-SDR V.3 with no antenna attached and SDRplay RSP with no antenna attached and they both shown similar noise, SDRplay having some FM broadcast visible. Wrapping SDRplay tightly in aluminum foil didn’t affect anything at all.
CHIP is a great tiny computer, everything works surprisingly well and it boots super quickly. I wonder how much power it looses through the bright white led, it looks almost like a small flashlight. it would be nice to have it blinking once every 5 seconds instead of being always on. The only problem is that the W-Fi antenna is tiny and weak so it can be a bit tricky (but not too hard for sure) to place all your components in a small case without something interfering with the Wi-Fi connection. Just make sure you don’t place anything over the Wi-Fi antenna (it’s on the PCB side opposite to USB ports, on the “far end” of the CHIP board).
Overall I am very happy with the project, I’ve learned many exciting things while working on it. Taking into account how much Inmarsat equipment and subscriptions cost, the Outernet is a great way to get familiar with satellite communications without investing much.
Just as a side note, I have photos and screenshots of all the stuff described above, so if there is any interest I could add them here. For now here is a photo of my receiver:
This is how RTL-SDR behaves without LNA and antenna attached. That spike around 1539 MHz is always there, I believe it can interfere with the Outernet signal. It can grow bigger sometimes but never goes smaller than on the picture.
And this is with LNA and antenna attached. So much noise! And as you can see on the waterfall graph, it changes all the time, getting worse from time to time. I’m not sure what causes it, but there are many potential causes as I live in a big plywood building that has many apartments, and every apartment uses fluorescent lights and obviously lots of electronics.
Those spikes on the graph are not satellite signals as you might think, those are just spikes, no matter how close you zoom. From time to time the whole range “jumps” and that results in the horizontal line on the waterfall graph. It doesn’t happen when the antenna is not attached.
Update on accessories for the DIY kit.
Zebra case for CHIP is amazing, looks great and feels sturdy and well made. Now I am not afraid of damaging CHIP with static electricity. It’s made and sold by C4 Labs. I wish there was a case for Outernet LNA!
I’m planning to connect the components inside the case with flexible short cables so I can place components inside the enclosure more efficiently. I’m also planning to test the kit with L-band pass filter, it’s still on it’s way from Europe.
And here is the solar panel I’ve tested. It’s a great solar panel (not cheap too), charges power bank faster than wall charger when the weather is sunny, but doesn’t work very well when it’s cloudy. It has no fabric anywhere outside, all covered with polymer.
Yesterday I’ve finally measured how much power DIY Outernet receiver consumes and it somewhat upset me.
CHIP + RTL-SDR + amplifier = 3.5W (0.7A at 5V). That’s not too bad, but more than 2W claimed. Finally it’s clear why it drained my power banks so quickly. Not a deal breaker I suppose, but should be taken into consideration.
Because of the current draw of the RTL-SDR and CHIP, and associated heat dissipation, there really is no practical combination of battery and solar panel that will support that hardware operating for any significant time interval in a portable enclosed yet weatherproof case.
Even a larger 22w solar panel will not provide enough to operate the receiver and charge the battery for overnight reception.
Physics is our enemy with this hardware in the context of portable usability.
The question is why is the SDR using so much power? Is it the TCXO? My 1st gen DVB-T dongles never ran that hot. They were not E4000 all 2832 based. I ought to dig one out and see how it works in relation to the new V3 dongle.
It is odd. SDRPlay draws only 130mA while a FlightAware Pro pulls 310mA. Seems all these inexpensive sticks are power hogs, while nowhere near the performance of the SDRPlay.
Now, truth is, the extra 150 or so mA isn’t killing the system, but it all adds up.
Have not looked at what the GPS receivers have in the way of detector and demodulator… Wonder if it could be used instead of RTL.
edit- scratch that… GPS receivers are 1 bit.
However, this brings even more fun to the RTL-SDR mix… It can be used as a GPS receiver. Not sure why one would go to that trouble, but it can.
k5ted, it’s not about price, it’s about what chip it’s built on. AirSpy is even more power hungry than RTL-SDR as far as I know. SDRPlay probably uses the least power of all SDRs, and as far as I know it’s the only SDR using Mirics chip.
However it’s not just SDR that is a power hungry. My measurements show that CHIP itself uses around 2W even without any SDR. Which is not bad at all for a computer running Linux, it’s just a bit too much for autonomous device.
I think the ultimate solution is to configure the receiver to work only when there is enough energy coming from solar panel, then it wouldn’t drain energy during the night and cloudy conditions and this would make the receiver truly autonomous.
My system, with CHIP, RTL-SDR, and LNA draws between 800mA and 1.1A after startup. Most of that is the CHIP. Some is the USB hub I use to give some more ports for flash memory, or someday, a better WiFi radio.
This centralized Outernet receiver for distributing content around a village could be done at present with a 45W solar array and large capacity battery.
The system could be configured so that the receiver turns on per a schedule. CHIP only during the day, serving up stored content, and RTL+CHIP at night running off the stored energy. Upside of this is the CHIP + RTL only run at night when it’s coolest outside, ensuring best receiver performance.
A 45W solar panel gives about 3A at 12VDC. The CHIP by itself only draws about 600mA. That’s 2.4A, plenty left over to charge a large battery. Better yet, craft the panel to, instead of 12VDC, be an array of separate 6VDC panels and use multiple 6V storage batteries to eliminate the waste of the voltage converter. CHIP and RTL will probably run fine at 5.9V.
This is obviously going to cost more than $100
RTL - $20
LNA - $20
CHIP - $10
PATCH - $25 (in many areas could substitute a $8 Active GPS antenna)
6W 6V solar panel - $60
2 x battery 5AH - $40
Enclosure - $25
hardware - $10
$210
Could probably be done for much less. If size is not a big issue, 45W 12V solar charging systems are usually about $129 at Harbor Freight. That’s a lot of power, and the panel frames could be used to “house” the other components.
)