Falcon Heavy may be able to lend support to Starlink mass-to-orbit needs.
Could be, yes. The main thought experiment is where they are currently with the V2M and V2M+D2D constellation deployment plan: They’re at a place where their launch rate more or less equals their decommission rate, and that of course results in the painting-the-golden-gate-bridge problem—at their ~current rate, they basically are decommissioning the first sat as soon as the last sat launches 5 years later. That’s obviously not a solid business plan so for sure they need to figure out a way to increase satellite launch rate; obviously they’re not going to stand-still relative to today’s launch rate and its very reasonable to assume F9 launch rate will continue to increase. Hard to say where the practical limit is for F9 launch rate.
Extending the thought experiment to rolling out a full V2 constellation [on Starship] returns a much less favorable outlook, however. We’re easily years away from the full 4k satellite V2 network (let alone with D2D payloads), and probably looking at close to 2030…
FWIW there’s a few FH things to work through on the concept of augmenting F9 starlink deployments:
- Its kind of hard to say what the practical rate of FH launches could be and its unclear how a FH launch impacts overlapping F9 mission timelines/rates. The solution would be some blended manifest (90/10 F9/FH or whatever) that probably ends up not adding too much benefit vs just figuring out how to launch moar F9s.
- Launch cost per sat will go up due to the more expensive FH, though that could end up in the noise for SX.
- The biggest roadblock is that the F9 Starlink stack is really pushing the limits on lateral mode*** and so the ability to add sats (and thus height to the stack) is at best, a very diminishing returns kind of situation.
***Bit sideways here, but SX driving first modes down on Starlink is truly is one of the most amazing things SX has done in the history of SX, and something not a lot of people talk about. Usually launcher PUGs spec lateral mode to at least 6Hz (F9 is currently 10 I think). In conversations with most of them, they're willing to entertain the notion of going down to ~4Hz with analysis. BUT…because SX owns both the starlink stack and the rocket underneath and [obviously] have both very well structurally characterized, they can hyper tune the starlink stack such that they fly WAY below 4Hz. They’re REALLY pushing load boundaries and flying a way floppy stack in order to maximize the number of sats on each launch.
That low(er) frequency manifests as a huge mass savings vs designing to a “normal” frequency; SX can strip a LOT of mass out of the starlink stack that would otherwise be needed to keep the stiffness up. Bit of a WAG here, if spaceX had to keep the lateral mode above 4Hz, they'd probably only get ~2/3 the number of sats on there.
Just think of the constellation’s evolution over time if they were launching at 2/3 the rate. Further, think of the massive launch cost and satellite rate advantage they have over anyone else—even customers that fly on F9 will never get that tightly coupled stack+rocket; even an otherwise exact starlink replica constellation but from a paying F9 customer would only get 2/3-3/4 the launch rate that SX has with Starlink because SX would (understandably) not want to toe the line like they do internally.
Anyway, circling way back, the
actual point here is that there's not a lot of opportunity to make the Starlink stack much (if any) taller with more sats in FH because they’re basically on the limit of the starlink stack being too floppy for F9. The additional mass from additional sats [to say nothing about the mass being way up top] would quickly push the stack beyond the floppiness limit, so more mass would have to be baked in throughout the stack to bring the stiffness back in line…but of course all that extra baked in mass unfavorably impacts the floppiness of the stack also, so even more mass would have to get baked in to control that mass…and so on…
If we hand-wave FH reusable as 50% more mass than F9, I’d guess at best they could only bolt 25% more sats on FH vs F9. Not worth it IMO—just launch 25% more F9s.
What is all that mass used for?
Figure about 1/2 of that mass for the actual payload, 1/2 for the actual structure, and some of both for thermal (which gets a bit muddied). Some of it actually doubles as structural mass—for instance, if you need some thermal mass there but also need a structural or shielding plate there anyway, you just make the plate a bit thicker than you would if it were solely structural. Some of it is baked more into the payload parts themselves—more/thicker copper in PCA layers, for instance.
For Starlink's Ku payload, a major enabler is that the relatively high frequencies mean everything ends up pretty small, so it’s really easy to package a phased array antenna on a typical PCA. The lower mobile frequencies mean everything is bigger—especially the antenna elements themselves, in all three dimensions. A single mobile band element is easily 10-20x (or more) the area of a Starlink Ku element, and thickness also comes into play. So…the array becomes multiple PCAs, and so then you need interconnects between them, phasing needs to be matched between them, etc. For the Starlink Ku elements you can build a reasonably efficient radiating element within a practical number of PCA layers and overall thickness. For a C band element you might need many many mm of thickness for it to really work the way you want it to…manufacturing issues aside, think about how heavy an 8-10mm thick PCA is going to be…. (There are other ways to get the dimension than a solid PCA, but there’s still more mass involved than a “normal” sized PCA)
Structurally, obviously you need to survive the launch environment and then have a release mechanism once you’re up there, but even on orbit you need a rigid enough structure to keep the many-m2 area array pretty flat. There’s going to be some pretty significant thermal gradients and distortions on the array that’s just kind of hanging out there on the side of the main starlink satellite, and its largely going to have to be thermally self sufficient. There’s also likely material geometric distortions between the main satellite body and the deployed D2D array that you either need to solve with a super stiff (= heavy) deployed structure, or additional sensors on the D2D array that feed back its overall shape/position to the main satellite so the main satellite knows where the D2D array is pointed.
The difficulty in an "unmodified" phone being able to reach a sat is what I had talked about earlier... if you have to hold your phone up at the sky ("Can you hear me now?" lol), they are at the limit of their ability. And I suspect if attenuation from your body and/or careful positioning of the antenna are necessary, there are going to be lots of places /situations where service just may not work.
Yes, and, that’s the reality of any D2D service, pretty much for everyone. On the forward link (from the sat) the link is FCC/ITU regulated to a PFD, so you
can’t just turn the sat up to 11 to add dB’s. On the reverse link (from the phone), there’s really only so much power a phone can put out physically, and output of course is also regulated by various health and safety organizations. The shitty part is that every non-sat phone has multi-directional antennas, so unless the services are selecting specific antennas within the phone (eg, only the ones on the top of the phone) a bunch of the power is totally wasted spraying signal the wrong way. NBD on terrestrial networks for that tower a few clicks away. Real big deal when the “tower in the sky” is 500-1000+km away.