I thought I'd share some pictures of my receiver setup.
I have two dishes, one to pickup the blockstream single on G18, and another to pickup the signal on Eu113-- most of the US are covered by both of these signals.
Once some upgrades are made to the sending side use of both dishes will allow half the delay when sending blocks (and, obviously, more reliability against obstructions or equipment failures)-- pretty handy in places that happen to be covered by multiple signals. At the moment I'm just using one at a time, though both work.
The dishes I'm using are 76cm Winegard DS2076. I paid $45 each for them on Ebay. If I hadn't found such a good deal on this I probably would have used 90cm geosat pros (which are about $100). For my location the 76cm is adequate, though I've had some outages during bad weather-- heavy rain attenuates the 12GHz signal a LOT.
I understand blockstream is going to be making some signal changes that should improve reliablity, and also some modem changes that will make it easier to get pointing really peaked out. The current tools give a really noisy SNR measurement which swings over a few dB in the space of seconds even when you aren't changing anything, this makes it really hard to dial in the pointing and get a really perfect alignment. E.g. my polarization could be off by as much as 30deg and I wouldn't have any idea, because the changes just weren't visible on the background noise.
For feed horns and downconversion I'm using MK1 PLL LNBs on the dishes which were an astonishingly low price of $8 on amazon. These are the appropriate devices for the americas signals, and they seem surprisingly good. Europe and Asia need different LNBs.
The dishes are connected back to my equipment room with a ~250 foot coax run using some fairly low loss cable that I reclaimed from a CATV temporary lateral that was abandoned on the under the grass after the put a permanent install up on the polls. The reported SNR looks the same both with and without the cable, so I guess it's not too long. YMMV esp with less heavy duty coax or with LNBs that have less gain or lower voltage power inserters: I kinda expected these to be too long.
Then I'm using Direct TV "swim" power inserters. These cost $7. They are a little bulky but the only real complaint I have is that they're 120VAC only-- all my computer gear is 240v for efficiency reasons, so these being an odd ball out is a bit of a pain. But I am probably the only US user who is weird enough for this to be a problem.
Finally, I'm using the recommended $24 nesdr USB RTL dongles as the SDRs. Not much to say about these things. They're inexpensive and they work. I contemplated using a nicer SDR (I have a couple to choose from...) but I figure my bug reports are more useful if I use common hardware.
I use USB extension cables to hook the SDRs up to the host (otherwise the SDR is a fragile wart on the back of the computer).
These feed an older 3.4GHz quad core E3-1230 1u box that runs the fibre-enabled Bitcoind.
I've encountered a couple bugs which blockstream has been fixing as I've found 'em. In particular
the pull-req to store out of order blocks is essential. I've had a couple internet outages where the sat signals have successfully kept my Bitcoinds receiving blocks. Success!
The biggest issue that I had with the install is that multiple times I used a laptop for initial pointing that was too slow. And the blockstream modem software really doesn't give you a usable warning if the computer is too slow. When the computer is too slow. It _looks_ like its working, but that there is no signal. My small laptop that was easy to haul up onto a ladder was just too slow, and even when I switched to a faster one it was too slow while not plugged into AC power (so it was fine on the ground but when unplugged to drag it up onto the ladder it started returning junk). I wasted _hours_ due to this one problem. Since pointing is a little tricky, esp if you haven't done it a bunch of times it isn't surprising if it takes you a bit, which just makes it take longer to realize that the lack of success is due to a slow system.
On with the pictures. First, a wider show to show how I have the dishes placed:
I decided to wall mount the dishes: Compared to ground mounting them above where anyone would walk in front of them and block the signal or knock them out of alignment. Compared to roof mounting they're somewhat protected from the wind, I didn't have to worry about causing any leaks, and they're less conspicuous.
Here is a close-up of the dishes:
Like most small dishes these are offset fed, so they point much higher than they look like (about 24 degrees, in the case of these dishes).
You can see the dishes are aimed 'cross eyed', there is a power poll that gets a little close to the line of sight of the signal, so I wanted the dish that was pointed more towards the poll to be the dish that was further away from it. I could have located the dishes elsewhere but this location has the advantage of being invisible inside the buildings unless you push your face right up to the glass.
And my equipment room with the power inserters and SDRs... and a bunch of unrelated stuff. (I have the wall opened up at the moment due to unrelated work) To the right is the top of the rack that has the computer in it that handles the signal.
In the pictures I also have a pair of 1268 MHz-center 35MHz wide bandpass filters between the power inserters and the SDR. They're not required, though I find they do improve SNR by about a half dB or so (the inexpensive SDRs don't have very selective front ends). Mostly I have them just to avoid any issues with high power transmitters that I have getting picked up by the long coax run.
Beyond my issue with the slow computers making me falsely believe my aiming was wrong, the setup was really easy. (Though, I do have a non-trivial amount of experience with radio, SDRs, and Bitcoin (obviously)).
Poll
