Starlink Teardown: DISHY DESTROYED!

In this video, I do a full, destructive teardown and light analysis of Dishy – the Starlink User Terminal (dish). This is the first full teardown, so I'm really pleased that I get to be the first to share this with the world. Huge kudos to the Starlink team – this is incredible work, and it must feel great to know that users are finally getting their hands on it.

This video was shot/edited quickly, so please forgive the quality. If you have any questions, leave them in the comments below, or hit me up on Twitter – I'm @kenkeiter.

STARLINK/SPACEX TEAM: If you have any concerns – including about the way I'm portraying your hardware, or incorrect details – please feel free to reach out to me directly! I have nothing but respect for the work you've done here. Leaps and bounds ahead of the competition. Also: I'm sorry for the massacre here – I didn't have the right tools to do this teardown as cleanly as I wanted to, and a piece of technology this nice deserves better treatment.

=====

📺 MEDIA COVERAGE

– https://www.businessinsider.com/spacex-starlink-satellite-dish-user-terminal-cost-stmelectronics-outsource-manufacturer-2020-11
– https://hackaday.com/2020/11/25/literally-tearing-apart-a-spacex-starlink-antenna/
– https://arstechnica.com/information-technology/2020/12/teardown-of-dishy-mcflatface-the-spacex-starlink-user-terminal/

=====

🔬 RESEARCH

– This paper (https://bit.ly/3fzcZ8g) describes the design and optimization of the circularly polarized slot patch antenna extremely similar to the ones in the video.

=====

🔧 CORRECTIONS

** Accuracy of the information presented here is important to me! ** I'll be maintaining the following list of corrections + additional info from viewers (prefixed by the name or username of the viewer):

– [@kenkeiter] Self-correction: I mentioned a high-power version of the PoE standard – this would not be IEEE 802.3at, but instead IEEE 802.3bt.
– [Doug Mohney] Starlink satellites are not technically Cubesats – which would imply that they're built to a specific standard. Optical crosslinks are not in v1.0 Starlink satellites.
– [@TMFAssociates] RF frequencies as stated are incorrect. They should be 10.7-12.7GHz Rx and 14.0-14.5GHz Tx.
– [Aaron Huslage] This is not FR4, but probably Rogers RO4000-series material or something with a low Dk.
– [nraynaud1] kindly reminded me that a differential gear configuration allows us to infer control of both azimuth AND elevation (🤦‍♂️), which I have verified with a bench power supply. Although Dishy currently only controls elevation, a future Dishy update should allow it to point any direction it pleases 📡
– [Ergzay] corrects the above correction (lol) and says that Dishy is _currently_ capable of changing its azimuth and elevation – a behavior I hadn't previously seen. The more you know 🌈

=====

💬 SUGGESTIONS

The accuracy of these suggestions can't be verified yet, but I'm sharing them here for everybody:

– [Marcel Dejean] LO_CLK would be local oscillator clock - the reference clock for the RF oscillator's phase locked loop.
– [Oleg Kutkov] Those copper elements above antenna patches look like a polarizer to me. Starlink uses circular polarization waves and these are probably the elements that "convert" circular wave to linear and vice versa.

=====

✏️ IC MARKINGS

** Chip markings that were not legible in the video ** (where '/' denotes a new line)

– the smaller ICs that I'm calling "phase shifters" or "switchers" are marked: "GLL / AA62 / 946P / AEBJ"
– the larger IC that I suspect is an integrated RF frontend: "VQ944833 / GLLBSUABBBA / HPAPUKOR GL"
– the IC that I refer to as a "clock distribution amplifier": "GLLBLU / HPAQR VQ / KOR 006 / ACBF (e2)"
– the "application processor" or "RFIC": "GLLCC0CA6BF"

Видео Starlink Teardown: DISHY DESTROYED! канала Ken Keiter