Just a few random comments on this thread from someone who is not a hardware builder:
I like the general idea of using a small, low cost, low power board for this purpose but I suspect that the Raspberry Pi 2 may not be the best platform. Updatelee has pointed out the problems in posts #31 and #32 in this thread. People should realize that there are MANY small computers out there. A Reddit user recently posted a Google Docs spreadsheet of
Multicore ARM dev boards, $100 and below and it is amazing just how many there are. Several have arguably better specs than the Pi 2. Another device that has piqued my interest a bit is the soon to be available Fitlet (
PC World review), though I'm still not clear what the final price will be on it.
If the device is to function primarily as a receiver/STB I would definitely want the ability to play 4:2:2. On the other hand, if it is to function as a PVR, then my wish would be for a well supported repository that allows doing seamless upgrades, without having to go through any extra steps that Linux guys don't blink an eye at but that frustrate other users, like having to go through a bunch of steps to install tuner drivers.
Regarding USB vs. PCIe for tuners - even if you don't mind having extra boxes for tuners and having to decide how you are going to plug in those cube power supplies, having external tuners is just plain messy if you have more than one. And the problem is that most USB devices only have a single tuner. If you look you might find a dual but that's as high as it goes. Whereas, you can easily buy a quad tuner PCIe card, and it's contained within the case so no messy external stuff to deal with. Many of us are running dual tuner LNB's and multiple dishes and we need tuners to connect all those LNB inputs to.
Now just as a digression, the one thing that could change everything would be if anyone ever develops an LNB that basically functions like a HDHomeRun device, except the streams it sends are from satellite, not OTA. The idea would be you'd run a single Cat 5e or Cat 6 underground cable to each LNB, and the LNB itself would be able to tune both polarities and on request send one or multiple streams down the network cable, and then a backend program or media player anywhere on the network could select a desired stream (channel), similar to a HDHomeRun dual paired with MythTV or TVHeadEnd. That would eliminate the need for separate tuners altogether and have the advantage that there would be no signal loss before the signal is converted to a digital stream. And you could put a ridiculous number of "tuners" in each LNB, so that you could stream a number of channels simultaneously (from either polarity and from multiple transponders) down the network cable - you would not have the expense of running a separate RG-6 cable for each tuner, and you would not have signal losses on longer runs. In fact you could put a dish hundreds of feet away from the receivers if you used an optical (fiber) converter at each end and ran fiber optic cable; you'd just need a power source near the dish to run the fiber converter at that end, and in any case you'd need a way to power the LNB's. And just as a backend system can utilize multiple HDHomeRun devices, it should be able to deal with multiple networked LNBs.
Even though no such system exists today (as far as I'm aware), I wonder if it could be approximated by putting something like a Raspberry Pi 2 (or other single board computer) and dual USB tuner at the dish itself in a weathertight enclosure. Once again you'd only need to run cat 5e underground cable, and a pair of wires to provide DC power to the Raspberry PI 2 or whatever you use. The small computer would power the USB tuner which would in turn power the dual output LNB. Then you could run software on the Pi 2 or whatever small computer you use that would do all the tuning and convert the output to a stream usable by any PVR program, such as MythTV or TVHeadEnd, or a media player such as Kodi or VLC. Though I personally wouldn't have a clue how to put something like that together (mainly because I would not have any idea what software would need to be used), the idea of having a digital stream come right from the dish itself would be quite appealing to me. I don't know if anyone else feels the same way, though.
And finally, I don't think many of you guys would care much for OpenElec, because despite the name, it's really not very open. They really don't want you running anything other than OpenElec and its included software so they don't give you basic tools found on most other Linux-based systems. IMHO they should have called it ClosedElec because from a user's point of view, it's a pretty closed system. I realize the "Open" part might refer to it being open source (not sure about that either) but it will definitely make your life more difficult if you want to do any tweaking. I've opted to use
RaspBMC (now to be called
OSMC) for that very reason. I know OpenElec is more popular but I didn't like it and I'm not anywhere near into Linux the way some of you guys appear to be. But then again, it's possible there's a secret "developer mode" or something that I didn't find; I gave up on it after about an hour.
P.S. It occurred to me after I posted this that another advantage to having a computer and tuner at the dish would be that you'd only need to open the weatherproof enclosure to gain access to the HDMI and audio outputs, and the USB ports. So you could have a "peaking mode" where if you connect a small speaker or headphones, it would play an audible tone that rises of falls depending on signal strength, whereas if you connected a small video monitor (and maybe a keyboard or mouse to enable making selections), it would show more detail about the strength and quality of the signal being received, right at the dish. So you would not need an external signal strength meter or anything of that nature, you would just need the software that supports the "peaking mode".
