That link is correct; Clyde Space won an award from the UK Space Agency to construct satellites for Outernet, but we’re unable to launch them.
Clyde Space was not involved with anything we were doing on Ku.
So right now outernet get 15Mb a day in data to the file system, off of a BPSK 2.4K/sec encoded stream. If they wanted to double that then they would need to switch up to QPSK for a 4.8K/sec stream.
This means that the bandwidth of their bent pipe on each satellite would have to double (pesky shannon’s law) and the power distribution would be spread out over that larger bandwidth so our RX antennas would have to have somewhere arround 3-4 db more gain. Forget the 170mm flat panel antenna, we would all be winding and pointing helix antennas.
That is just a small taste of going tor 30Mb/day,
-Cecil
curious as to which Inmarsat service Outernet is using.
We don’t use any Inmarsat service; we operate our own on an independent channel.
1 Like
What is the max bandwidth of the channel you have allocated to you? Just curious if a speed upgrade could be done someday.
-C
I am curious as well. Could Outernet ask Inmarsat to burst to twice the normal data rate say for 1 hour and would the current configuration of Outernet receiver automatically cope.
What is the cost for the channel and can we contribute money to help support it?
I personally don’t see a reason to increase the bandwidth at this time. I think once the Weather package is removed from the carousel there will be plenty of room for the other stuff. 20MB a day is after compression which should be plenty for all the news, weather etc daily updates plus some requests and community content.
[Skylark][root@skylark:/mnt/downloads]$ du -h .
0 ./preprocess/sop
0 ./preprocess
3.0M ./News
108.0K ./Requests/Inmarsat - Wikipedia_files
252.0K ./Requests/L band - Wikipedia_files
708.0K ./Requests
5.6M ./Wikipedia
8.0K ./Games/Connect Four
8.0K ./Games
256.0K ./Amateur Radio/APRS/APRSAT
256.0K ./Amateur Radio/APRS
16.0K ./Amateur Radio/AMSAT/News
16.0K ./Amateur Radio/AMSAT/ARISS
32.0K ./Amateur Radio/AMSAT
288.0K ./Amateur Radio
8.0K ./Community Content/Bolivia
260.0K ./Community Content
0 ./grib2
2.1M ./Weather/data/weather/2017/01/05
2.1M ./Weather/data/weather/2017/01/07
2.0M ./Weather/data/weather/2017/01/08
2.1M ./Weather/data/weather/2017/01/09
8.1M ./Weather/data/weather/2017/01
8.1M ./Weather/data/weather/2017
2.1M ./Weather/data/weather/current
10.2M ./Weather/data/weather
5.1M ./Weather/data/oscar
15.9M ./Weather/data
340.0K ./Weather/grib2/gfs.2017010500
664.0K ./Weather/grib2/gfs.2017010600
1004.0K ./Weather/grib2
17.5M ./Weather
156.0K ./opaks
424.0K ./.cache
27.9M .
[Skylark][root@skylark:/mnt/downloads]$ uptime
07:29:00 up 2 days, 7:21, load average: 1.07, 0.86, 0.83The outernet SDR is set up to decode 2 phase signals (bpsk). The receiver in software would have to be able to shift from a two point decoding constellation to a quadrature format (QPSK) on the fly to do this. As it is they have to decode 4 streams to find the correct phase. Processing overhead would jump to 16 decoder streams to do Quad decode, unless there are a few other SDR tricks for quad decode. I am still on the low rungs of using GNU Radio so I may be missing some details.
The 15M/Day is because of packet overhead, that is why I said to the disk.
-C
@CRCasey We are on a 5kHz channel, which is the maximum we can afford now. A speed upgrade is possible, it’s just a matter of paying more money–lots more money. The lots more money is why it’s really not possible to bootstrap this type of business; there’s no way around needing investment capital to start off.
@Seasalt No, we can only lease capacity on a long-term basis. That is why we have our own channel. The alternative would be to use one of their existing services, which are not broadcast channels. The current receivers can cope up to about 20kbps.
@kf4hzu Inmarsat requires that we keep pricing under confidence. Yes, the community of users can absolutely help to support our exceptionally high bandwidth costs. I always thought that the best way to support the free service is by buying receivers or spreading the word, which includes helping us bug test the software, develop new applications, and build our reseller network.
Yes, increase bandwidth would require a commensurate increase in power, in order to maintain the same energy per bit (and antenna gain requirement). Some day, our energy per bit will be far higher than it is right now, so we can close the link to a 3dBi antenna. This would allow for reception everywhere without pointing. But for now, it is what it is.
We are definitely looking into optimizing the demodualator and decoder.
1 Like
Would be interesting to try a two element patch antenna for wide beamwidth, and stack another LNA on, piggyback power from the first one, just to see what it looks like.
How much more money? And for what extra amount of capacity?
I just wanted to reiterate, that I’m not talking Outernet down, just lamenting what has been lost, and I do understand the reasons.
Thank you Syed for the reply! I understand the confidentiality agreement with Inmarsat. I’ve been talking about and showing many people this service. I’ve also bought several components from your Amazon sales items so hopefully that helps a little. Maybe have a PayPal or Bitcoin donation link for people that want to donate directly? I think you run a for-profit company but you could always take the money and send starter kits to people that can’t afford them or Internet access.
We really want an easy way to enable donations, but what we don’t have is a partner who knows where the donated received should go.
You may consider that this actually works against your mission statement, “a library in every pocket”. Availability implies cheap, cheap means commodity hardware, commodity hardware means low profit and lots of competition. I think that the basic economics of the situation means you are either catering to those with expendable income that are interested in the service because it seems neat (i.e., ham radio enthusiasts) or actually providing it inexpensively, and not both.
I’m wondering what the cost reductions could be if the receiver were designed from scratch as an integrated single device incorporating the LNA, tuner, and file server?
One day, maybe broadcom might include something like MAX2121B functionality inside of an appropriate processor if it were popular enough.
Great idea.
What about making the RTL dongle and the LNA and putting them on a small board that plugs onto the CHIP connectors?
There is an Arduino SDR shield, but it’s HF and relatively expensive. It would certainly be possible to make a similar “shield” for a CHIP, complete with filtered LNA, but the price would likely exceed the cost of discrete SDR and LNA modules. The positive side of this is that we could do away with the USB interface and other extraneous components.
Significant cost reduction. We’ll have more on this topic in a few weeks.
1 Like
For DVB-S, there already exist LNB devices that output the received stream on an ethernet port.
They are used in modern digital TV systems where the same dish can receive up to 8 different TV
programs and distribute them to IPTV receivers on the local network. I presume a Raspberry Pi could
be connected to such an LNB and serve as a processing- and file server. In addition you would only
need a dish. Of course all of this is way more expensive than the L-band setup and requires a larger
antenna, but that same system can also be used to watch TV. So it could be already present in
some environments (although TV systems usually are a bit more primitive).
In the past, I have experimented with receiving data streams on my Dreambox DVB-S receiver.
One can connect to a network socket on this receiver and get the content of a selected stream.
10 years ago, internet providers sometimes sent unencrypted data this way and it could be captured
in a similar way the current Outernet receiver does (putting received file fragments in disk files).
I heard about Outernet only recently and I understand this is an already passed station, but I would
think it would be possible to get some space on a DVB-S transponder to send data, maybe it could
even use the left-over capacity because of the variable-rate encoding of the TV programmes, maybe
quite cheap when someone has the proper “connections”. Such a stream would be 1000 times
faster than the current L-band stream. Receivers for DVB-S are $50 today, of course it requires a
specific type to be able to receive the data via the network but many Linux-based cheap DVB-S
receivers should be able to do that.