- What the to flat metal plates in the upper corners of the image are
It's possible they are phased array gateway antennas. The per-sat throughput of an all mobile service sat is going to be WAY lower than fixed broadband, so the gateway link capacity doesn't need to be as high; its possible the complexity/cost/packaging of [necessarily lower capacity] flat/fixed phased panels beats the the steerable parabolic gateway antennas that we see on the fixed service starlinks.
Its also possible they're some thermal radiator, though typically one wouldn't aim a radiator right at earth...
It's possible they are basically fixed, non-phased gateway antennas (so, radiating a really wide cone), with some thermal feature on them (the plates that we're looking at).
It's also possible they are small fixed service arrays...if so, likely used for some degree of testing out the D2D functionality. (There's not a lot of value to starlink services have a significantly reduced capacity sat providing production level service.)
- The cylindrical section opposite the hall effect on the other side of the array, assume another thruster...
The thrusters have the cylindrical red cover on it in the lower right. The ISLs have flat red covers, center right and far lower right. The earth sensors have a flat red cover on it, lower left--they're the ones at a ~45deg angle to the orthogonal. The black cylindrical section running from the thruster all the way to the other side of the sat is the prop tank/structural backbone.
- The object oriented at an angle in the upper left with a round attachment. Wonder if this angle related to an device for orientation/positioning?
The smaller black pancake-ey cylinder in the upper left (just above the white panel) is likely a motor housing for the ISLs on that side of the sat.
- The Reddit thread mentions these may be arrays that can unfold.
I don't believe the user array unfolds. As noted upthread I don't think there's a proper fixed internet payload on these sats [which would be covered up by an unfolding D2D array]. Among other things, the thickness of the sat just isn't there for multiple folds of the D2D array. We're looking at probably 25-30mm for a user array panel, when stacking up 1) the structural plate (the black plate under the silver user array PCBs with what appears to be orthogrid ribbing, 2) the board itself (the silver covered panels under the round element disks), and 3) the raised round elements that are precisely air-gapped from the boards.
I also don't see any hinge mechanisms. I'd speculate that the two silver/grey brackets on the lower right edge are either lifting features or clamping features for the
stack restraint bars.
Also--noting that its fair to figure SX might not adhere to this logic for a test sat, which these presumably are--phased arrays that serve a hemispherical geometry generally want to be as two-axis symmetric as possible. The geometry of any beam more of less takes on the shape of the array, and you generally want a symmetrical beam pattern on the ground (that of course goes oblong depending on elevation angle). So...if this were to unfold, it would really want to unfold to 4x panels. Only 2x panels would create a very high aspect ration beam cone that would really stretch out at low elevation angles...that increases the difficulty/complexity of traffic scheduling/routing.
I didn't notice the American flag reflection on the array initially.... cool "easter egg" shot.
Classic space industry there. I'm pretty sure its a contractural requirement for every American built satellite to have a picture taken with a reflection of the highbay flag...
I'm curious to know what the tabs on the left and right edges of the antenna platform are for. Are they just a production artifact?
I'd guess handing features used during the manufacturing of those structural plates through final assembly of the array onto the sat. It appears as though those black ortho plates are long and thin; it appears as though there are 5 of them that makeup the array. Its reasonable to speculate that the PCA is also that wide ("8 elements wide"), and there are probably 2-4 array PCBs in each of those 5 sections (so the user array is actually ~10-20 individual PCAs)