I believe the plan is for phased arrays, at both ends, at substantially lower prices than are currently commercially available.

Since SpaceX has already built the "first batch" of satellites (delivered to launch site as of a few days ago), one has to imagine they've largely solved the manufacturing problem and reduced the cost from 250k per antenna.

IIRC, OneWeb's initial plan was to use phased arrays as well, but it proved too expensive and they switched to mechanically steered antennas (at least for first gen).

Phased array is definitely the way to go if you can get cost down. No moving parts means fewer hardware/mechanical failures and less noise

What is the difference between phased array and MIMO? Isn't MIMO a solved problem and shouldn't the technology be transferable to this use case?

A difference is that phased arrays need sub-wavelength tolerances in phase between emitters to work

I can easily imagine economies of scale in MMICs (Monolithic Microwave Integrated Circuits) bringing the price down enough to make phased arrays affordable. I sincerely hope it happens.

Why are electronic phased arrays so expensive? I am no radio expert, but it seems that the pricey bit might be the engineering required rather than cost of production. From what I remember of my polar calculus, it sounds like a lot of painful calculus written once and fed henceforth into a computer.

Presumably you need a high-end FPGA or a custom VLSI chip to process the signals.

Both of these are very expensive in low-volume.

You do not need a high-end FPGA; you need a bunch of small antennas (potentially extremely cheap, as these kinds of antennas are often microstrip antennae etched into PCBs), and a bunch of microwave-rated components such as mixers, phase-shifters and capacitors. The phase shifters allow you to add up the signals in such a way as to form a coherent beam in a certain direction, this is known as beamforming [0], and with some admittedly complex precalculation, an electronically-steerable antenna array needs nothing more than a simple microcontroller to set the phase shifters for the desired array orientation. Of course you would want to adaptively steer it based on received signal, but this can be done without the need for fancy FPGAs; your receiver signal chain will include all the mixers, demodulators and signal processing elements needed already baked in silicon; similar to how a WiFi chipset does channel adaptation by analyzing the received signal without resorting to high-end hardware, the kind of system that SpaceX is building does not require particularly exotic hardware (although it would be custom, and expensive to develop).

Speaking of which, there's a neat project that sneaks into a particular Intel chipset and extracts the CSI (Channel State Information) from the wifi hardware so you can actually do things like estimate which frequencies are being attenuated [1]. It's out of date, but still shows the kind of live, complex, signal analysis that is possible with commodity-priced hardware.

In short, I think it is certain that a base station can be built for <$1000, and likely that it can be built for <$150, although without knowing the precise specifications, it's hard to say. We did a case study on this in my antennas course in grad school 3-4 years ago, and our ending numbers on our prototype designs (not fully built and tested, of course, but simulated and spec'ed out with real-world pricing) were around $150. I am certain that our design as it was would not work, but something similar to it likely would. ;)

[0] https://en.wikipedia.org/wiki/Beamforming [1] https://dhalperi.github.io/linux-80211n-csitool/

$1000 won't recoup the cost of their constellation, so they better hope you're way off.

That's the build cost, which needs to be as low as possible. The cost charged to customers is a different issue.

It also needs to be rad-hard for space. Terrestrial FPGAs are subject to occasional "SEU" (single event upsets)[1] which can arbitrarily affect behavior because gates can flip. Robust terrestrial FPGA systems have to handle SEUs; you have to imagine that these events are more frequent in space and need to at least be handled.

(This may something you already had in mind, summarized as "high-end FPGAs;" I just wanted to add some interesting color some folks might not be aware of.)

[1]: https://en.wikipedia.org/wiki/Single_event_upset

In LEO it doesn't need to be rad hardended, especially for something that isn't life critical (compared to say, the control system of a space shuttle). At this altitude, the atmosphere and magnetic field provides a lot of protection still.

At the very least SpaceX would test the devices to see what effects the higher radiation would have. However it's not unreasonable to just have the software detect and reset itself, if a high energy event is detected. There are going to be a lot of these satellites and a lot of redundancy in coverage, so if one is down for a few minutes it's not that big of a deal.

Yes, as long as radiation only causes upsets and not permanent failures — satellites are somewhat expensive — and as long as software can detect faults and reset.

4000 satellites is enough to hit economies of scale. Per piece they are probably among the cheapest satellites in orbit.

Sure, but that isn't saying an awful lot; most satellites are very expensive. SpaceX would want to weigh the cost of hardening the system against radiation versus the cost of another launch to replace failure(s). I'd guess radiation-hardening of some kind (at least sufficient to prevent most permanent damage) would win out in that calculation, but I'll happily admit I don't have any particular factual basis for that guess.

Ground station cost is actually more of a concern than satellite cost, from market analysis I've heard. Going to need more of them.

Good to know, thanks. I suppose at the scale of SpaceX, ASICs may be an option.

I'm not sure 4000 is enough to break even on ASICs, but I'll admit I don't know much about the costs of either high end FPGAs nor ASIC production at quantity 4000.

From what I am reading here, a large part of the cost is up-front with chip design and setup. I suppose it depends on pricing. They also may wish to sell more after those four thousand units.

http://blog.zorinaq.com/asic-development-costs-are-lower-tha...

The microwave devices are the expensive part. The signal processing and software are straightforward with modern components like DSPs and FPGAs.

Are hand-offs really all that complicated? Cell networks have to handle it.

I don't think you need to move the antenna, but you need a wide footprint and probably overlapping coverage between satellites.

How does Iridium work? I'd think this has already been done(quite a while back even).

LEO satellite handoff is tough because you effectively need two antennas: One points at the "old" satellite while one points at the "new" one. You can achieve this with a single phased array antenna divided in half and controlled by software, but each half must still be large enough to give the needed S/N ratio.

Iridium uses an omnidirectional antenna which is fine if all you want is 1200bps. But if you want 5Mbps, you have to point the antenna. So Iridium is not applicable here.

Cell hand-offs don't happen nearly as frequently as they would for an LEO constellation.

The orbital period for a satellite at 550 km is ~01h 36m. That means any given satellite is going to be visible for a very short period of time, which means hand-offs will have to be extremely frequent.

Approximately every 2.5 minutes in fact.

That's per satellite; each satellite is going to have multiple beams, so you'll have to do a beam handover like every 15 seconds and a satellite handover every couple of minutes

> Are hand-offs really all that complicated?

The fundamental problem is this:

   https://en.wikipedia.org/wiki/Two_Generals%27_Problem

Don't know how cellular handles this.