This adds some color perhaps to the argument that this is for underserviced regions -- they don't mean third-world or impoverished even though it sounds like that, at least when I heard people defending Starlink.
Three random examples:
https://www.appliancesdirect.co.uk/p/kge36awca/bosch-kge36aw...
https://www.appliancesdirect.co.uk/p/ffu3dx1/hotpoint-ffu3dx...
https://www.appliancesdirect.co.uk/p/htf-540dp7/haier-htf540...
Of course, once you're doing the 'USA-megahome-sized' thing, it's pointless to bother with energy efficient fridges. And many people don't!
I know we are all used to 24 hour internet, but if I were energy-conscious or off-grid, I might want to turn it on for 3-4 half-hour sessions during the day.
I know a few people who have Starlink on an outlet timer or WiFi switched outlet, and only have it run during the day. It seems to be okay with that.
In other words is it 100W all the time because of physics or is it just "sub optimal" hardware that in theory can be mitigated through smarter protocols and fancier hardware?
The "all the time"-part is of course something that can be worked on even without changing the hardware. A simple switch on the outlet to turn it off during the night or if everybody is at work/school and thus doesn't need it would reduce power consumption a lot already.
"Fancier hardware" could also mean a timer for the outlet switch or a "smart home"-solution. This could mitigate connection time by turning the router back on before you arrive at home/wake-up.
There's a lot of potential for not having to use 100W all the time without hardware changes.
Talking to a satellite in space through the atmosphere has a lot of attenuation. That's for the physics.
For comparison, wifi is 0.03W for 100 meters, mobile phones are 2W for 35km range. A satellite using 100W is not out of the ordinary.
There's a balance to decide on, power vs bandwidth, noting that doubling power has little effect on signal (log2). They could probably half the power for little difference in operation, but a notable difference on the bill.
It's exactly the sort of thing that should be adjustable, there could be a setting to adjust power/bandwidth/quality. Both for customers and for support.
edit: the video mentions up to 300 Mbps and up to 200W power consumption during snow (extra power to melt the snow). They have a lot of margin to reduce power while still being able to watch YouTube 4k.
[1] Table 2.2.1 https://www.gov.uk/government/statistical-data-sets/annual-d...
[2] Claiming typical down speeds of 50Mbps-100Mbps
And yes, that's easily over a hundred dollar a year.
Also, if you aren't off-grid, you're apparently not so far from civilisation that previous utility suppliers couldn't be bothered to provide service to you. Maybe if this generation's utility suppliers got their act together you wouldn't need Starlink anyway.
People who are off grid might grumble at having to provision more solar, battery and inverter capacity.
But for people who have other Internet options, $18/month might influence their decision.
1. https://www.electricchoice.com/electricity-prices-by-state/
But even with insane generation rates, most of the electric bill is the power company service fee, not the generation fee.
For example, out of a $200/mo bill, maybe $50 is electric generation, the rest is the service fee. PG&E is insanely badly run.
Is it, though? The price drops have been significant, yes, but the efficiency of COTS panels hasn't seen any dramatic improvements in the past 20 years. Silicon panels have a theoretical limit of 29% efficiency and current models are close to 20%. There's not much room for improvement left there.
A 400Wp panel won't be able to provide 400W most of the time to begin with. Unless you're in the middle of a desert near the equator and have a sun-tracking installation, weather and varying daylight hours are a thing.
Under realistic conditions, a single 400W panel gets you between half and two thirds of the required power (e.g. up to ~1.5kWh/day, depending on location).
Just a quick reminder: 400Wp means the panel produces 400W of electricity when brand new (it'll degrade over time), at 23°C ambient temperature and an incident angle of 0° (i.e. sun directly over it). None of that is generally the case 8h a day on average, which is why your estimation won't work.
Depending on where you live, optimal sun hours vary between 3.5h/day and 6.5/day (continental US) or 2.9/h/day to 5.9h/day (most of Europe) [0].
You'd also need to plan for big batteries, since seasonal differences may be huge depending on latitude.
I'm sure that for a few hours during the summer, the solar irradiance in London is probably 90% of what it is at the equator. That doesn't help most of the year.
With heat inverter pump tech. taking over fridges, water heaters, AC, etc. and making them so much more efficient, it can give a lot of headroom on energy requirements.
My office has 5kW of panel hung, 10kWh of battery, and my Starlink terminal is on the house system (grid tied) because an extra 2.4kWh/day isn't workable with my system in the winter, short of a LOT of generator time - I'm severely power limited during inversions, and the Starlink dish more than doubles my office's "idle power draw" (it's the property network hub, one of our internet connections, inverter idle, sleeping computers, etc).
Yesterday, the 15.9kWh system on the house produced 78kWh for a "sun factor" of about 4.9, so a 400W panel would produce about 2 kWh. Whoops. That's not 2.4.
In the dead of winter, the same system can produce 2.5 kWh - yes, 2.5 kWh on a 15.9kW nameplate system. A 400W panel won't even power on the charge controller in those conditions.
To reliably power a 100W load, 24/7, in most areas, requires probably 1500-2000W of panel and 20+kWh of battery - or someone willing to light a generator, which is the far cheaper option. But "a 400W panel and batteries" (implied as a trivial thing to set up) definitely won't. It won't even run it 24/7 in peak sun most of the year.
... and that's before it tries to melt the snow off. From the blog post:
> During the heavy snowfall, Dishy quickly spiked up to 125W, peaking at 175W towards the end of the snowstorm.
Solar panels don't produce much covered in snow, either.
I deal with the realities of off-grid power in my office on a daily basis, and a vast majority of what's written about solar and batteries by people who don't have experience with them is simply wrong. I try to correct it where I can.
With a 400W panel and a few kWh of batteries, you could reasonably accomplish 14-16 hours of access during peak sun in the summer, 8-12 hours in spring and fall, and 0-3 hours in winter, except for 5-6 hours on sunny winter days.
Although I've heard that the newer Dishys use somewhat less power (50-70W), which does improve things.
The problem is that after 48h, then... what do you have? You need to have either enough panel area to both run loads and charge the batteries, often in dim conditions, or you need a lot more battery.
Doing reliable off grid power takes a lot more than most people who haven't dealt with it assume, especially if you're dealing with serious winters and don't want to light a generator.
I'm basing my estimates off what it would require for me to run 2.4kWh/day, year round, without generator, on my office system - and it takes a LOT. I have 5kW of panel, 10kWh of battery, a typical winter day load of around 3kWh (base loads + minimal compute loads for a work day), and I still require a couple gallons a year of gas to run this reliably. Depending on the winter, I go through 3-5 gallons of propane for heat as well, though I've been trying to move to kerosene - more efficient, as I can run a lantern as a combined heat/light plant on my desk, and heat my hands, which are the only real constraint on cold temperature work in there.
I know those numbers sound somewhat nonsensical for a 100W load, but I literally make a living in an off-grid office...