Over on Twitter and his github.io page, Pieter Noordhuis (@pnoordhuis) has shared details about his low cost RTL-SDR based GOES satellite receiving setup. GOES 15/16/17 are geosynchronous weather satellites that beam back high resolution weather images and data. In particular they send beautiful high resolution 'full disk' images which show one side of the entire earth. As the satellites are in geosynchronous orbit, they are quite a bit further away from the earth. So compared to the more easily receivable low earth orbit satellites such as the NOAA APT and Meteor M2 LRPT satellites, a dish antenna, good LNA and possibly a filter is required to receive them. However fortunately, as they are in a geosynchronous orbit, the satellite is in the same position in the sky all the time, so no tracking hardware is required.
In the past we've seen people receive these images with higher end SDRs like the Airspy and SDRplay. However, Pieter has shown that it is possible to receive these images on a budget. He uses an RTL-SDR, a 1.9 GHz grid dish antenna from L-Com, a Raspberry Pi 2, the NooElec 'SAWBird' LNA, and an additional SPF5189Z based LNA. The SAWBird is a yet to be released product from NooElec. It is similar to their 1.5 GHz Inmarsat LNA, but with a different SAW filter designed for 1.7 GHz GOES satellites. The total cost of all required parts should be less than US $200 (excluding any shipping costs).
Pieter also notes that he uses the stock 1.9 GHz feed on the L-com antenna, and that it appears to work fine for the 1.7 GHz GOES satellite frequency. With this dish he is able to receive all three GOES satellites at his location with the lowest being at 25 degrees elevation. If the elevation is lower at your location he mentions that a larger dish may be required. It may be possible to extend the 1.9 GHz L-Band dish for better reception with panels from a second cheaper 2.4 GHz grid dish, and this is what @scott23192 did in his setup.
For software Pieter uses the open source goestools software that Pieter himself developed. The software is capable of running on the Raspberry Pi 2 and demodulating and decoding the signal, and then fully assembling the decoded signal into files and images.