Manufacturing constraints for Kuiper launches

[LouScheffer]

Well, I understand that Kuiper has to launch 50% of the fleet (1618) satellites by July 30th, 2026. If we assume that 9 Atlas V can launch 50 satellites, that's 450 birds. A mere 27.8% of their needs. They will need about 30 [...] EELV equivalent launches on top of that in the next five years.

So assuming that Kuiper can't get some sort of waiver, who can do the remaining launches?  New Glenn most likely won't be in volume operation in time.  Ariane is currently launching about 6 times per year.  Even if the could double their capacity, that will take time.  So they can't do it all.  Soyuz could have done it in its heyday, with perhaps one launch per week (even if can only lift half as much) but it would need to be launched from South America, I think, because of ITAR.  That's only one pad, and a long supply chain.  Atlas would be happy to get 40 launches in 5 years, but could they get enough RL-10s and RD-180s?  These appear to be quite long lead-time items, so ULA would need a confirmed order soon to order the needed engines in time.  China is out due to ITAR, and Japan and India don't seem to have the launch cadence needed.  SpaceX does, but is politically dis-favored.  So who will launch the remainder of the Kuiper constellation?

My guess is the Amazon will spend about 3 years hoping New Glenn hits its stride quickly or they can get a waiver.  If neither happens, they will spread the order over several vendors (whoever can offer quick turnaround), then swallow their pride and give the remainder to SpaceX.

[GWH]

Vulcan should be ramping up over the next few years, so more flights could be ordered off ULA. I'd assume Atlas V's were ordered to fly in the next few years where as Vulcan could take over 2024 to 2026.

Assuming that Atlas V 551 is maxed out in mass and volume Vulcan will let them fly a fair more. 7200 ft³ on Atlas V, 11200 ft³ on Vulcan (200 m³ vs 320m³); a 55% increase. Then mass is 18,850 kg to LEO on Atlas V551 and 27,200 kg to LEO for Vulcan Centaur 6; a 44% increase.

Given that the worst case is Vulcan could fly 72 per mission based on mass.
Mass is arbitrarily estimated at 18,850 kg / 50 satellites = 377 kg each.
Volume arbitrarily estimated at 200 m³ / 50 satellites = 4 m³ each.

New Glenn should be flying in this same time frame of 2024 to 2026. I estimated a long time ago that New Glenn has 580m³ of volume, which would make mass the primary constraint there at 45,000kg / 377 kg = 120 Sats.

In theory ULA could offer the 7 meter fairing 3 liquid core rocket they have teased. Taking at SWAG at the comparative payload increase between Delta IV 5,4 and Vulcan Centaur 4 that would be 48,000 kg to LEO. Roughly equal to New Glenn. A backup plan that I won't include.

2022-2024:
9 Atlas V 551's = 450 Sats

2024 to 2026:
5 Vulcan Centaur 6's = 360 Sats
7 New Glenn's = 840 Sats

Total: 1,650

For New Glenn it only would have to fly an average of 2.3 times a year. Once in 2024 then 3 times in 2025 and 3 times in the first half of 2026. That seems very achievable even for New Glenn's late arrival.

[Kiwi53]

Very interesting analysis, thanks, just a couple of questions:

Do you have any reference for the Kuiper satellites being in the 350-400kg class? I have looked around and found very little information.

Is there any evidence that Kuiper satellites actually exist yet? If they're going to start launching in numbers in 2022, I'd have expected there to be at least a pathfinder launch of 2 or 3 test satellites in Q3 2021 at the latest, but there's been no announced launch AFAIK

[GWH]

Very interesting analysis, thanks, just a couple of questions:

Do you have any reference for the Kuiper satellites being in the 350-400kg class? I have looked around and found very little information.

I have nothing. The original post throws out a wild number of 50, so I just divided Atlas V's lift capacity by that number. 

Of interest is this document: https://eurospace.org/wp-content/uploads/2020/11/information-note-amazon-kuiper_18112020.pdf

It's from November 2020 but states the first wave of satellites would be 578. That would be roughly 64 satellites per the 9 Atlas V's. 250-300 kg would be a better estimate in that case?

[butters]

The BE-4 production ramp is likely a significant constraint on Vulcan and New Glenn launch opportunities in the coming years, regardless of when their maiden launches occur. It's going to take Blue Origin some years to be able to crank out 16-24 BE-4s a year just to enable a "reasonable" launch cadence for Vulcan alone. They have no volume-production experience, and optimization for manufacturability was a luxury that had to take a back seat to delivering a functional engine. Neither of these launch system is likely to be firing on all cylinders from a cadence perspective until the late 2025 or 2026 timeframe.

I suspect that Kuiper will require at least three (probably more) different launch systems to deploy their initial constellation. There's no single system with the spare capacity to handle their demand. The Atlas V deal was a relative no-brainer because it's the most appropriate launch system that is available right now and currently operating nowhere near capacity. On other launch systems, they're going to have to take whatever they can get, like slipping in some Arianespace Soyuz launches during lulls in the OneWeb campaign (or outbidding them for future block-buys beyond the current deal), or dealing with the vagaries of non-BE4-based newcomers like the Mitsubishi H3 and Ariane 6.

A little of this and a little of that. Maybe it'll add up to a suitable launch rate for deploying their constellation. Maybe.

[Kiwi53]

Very interesting analysis, thanks, just a couple of questions:

Do you have any reference for the Kuiper satellites being in the 350-400kg class? I have looked around and found very little information.

I have nothing. The original post throws out a wild number of 50, so I just divided Atlas V's lift capacity by that number. 

Of interest is this document ... It's from November 2020 but states the first wave of satellites would be 578. That would be roughly 64 satellites per the 9 Atlas V's.

250-300 kg would be a better estimate in that case?

250-300kg would be much better than 350-400kg from Kuiper's point of view.
Launch cost is more or less proportional to mass, so the Kuiper satellite being 40% heavier than Starlink would put them at an immediate costs disadvantage.

[GWH]

The BE-4 production ramp is likely a significant constraint on Vulcan and New Glenn launch opportunities in the coming years, regardless of when their maiden launches occur.
.... snip...
Neither of these launch system is likely to be firing on all cylinders from a cadence perspective until the late 2025 or 2026 timeframe.

It's 2021. The big shiny engine factory they broke ground in in 2019 is built and full of tooling.  Another 4 to 5 years to hit cadence?  Even for Blue that seems pretty unreasonably pessimistic.

Too bad ULA had been dragging their butts on engine recovery and reuse.

[TrevorMonty]

The BE-4 production ramp is likely a significant constraint on Vulcan and New Glenn launch opportunities in the coming years, regardless of when their maiden launches occur.
.... snip...
Neither of these launch system is likely to be firing on all cylinders from a cadence perspective until the late 2025 or 2026 timeframe.

Too bad ULA had been dragging their butts on engine recovery and reuse.

NOAA and NASA have finalized an agreement to fly a reentry technology demonstrator as the sole secondary payload on a weather satellite launch in 2022.

This demostrator that ULA needs as part of SMART development program.



Sent from my SM-G570Y using Tapatalk

[GWH]

Too bad ULA had been dragging their butts on engine recovery and reuse.

NOAA and NASA have finalized an agreement to fly a reentry technology demonstrator as the sole secondary payload on a weather satellite launch in 2022.

This demostrator that ULA needs as part of SMART development program.

I know, I looked that up before posting 

A 2022 reentry demo means an actual engine reuse likely won't be until 2025 or 2026. Vulcan doesn't have the separation hardware built into it from the 1st flight, they have a long ways to go until then sadly.  These potential Kuiper flights would be an ideal opportunity to try it and reduce production bottlenecks, however none of this would be in time.

[TrevorMonty]

Too bad ULA had been dragging their butts on engine recovery and reuse.

NOAA and NASA have finalized an agreement to fly a reentry technology demonstrator as the sole secondary payload on a weather satellite launch in 2022.

This demostrator that ULA needs as part of SMART development program.

I know, I looked that up before posting 

A 2022 reentry demo means an actual engine reuse likely won't be until 2025 or 2026. Vulcan doesn't have the separation hardware built into it from the 1st flight, they have a long ways to go until then sadly.  These potential Kuiper flights would be an ideal opportunity to try it and reduce production bottlenecks, however none of this would be in time.

Be surprised if they didn't design engine pod to be easily detachable from day one. My guess for next step is separate engine pod during one of earlier missions. Want prove this works before adding expensive inflatable reentry shield.

They also need todo some MAR tests with for size engine pod mockup and heat shield. Similar to what RL have done.

Sent from my SM-G570Y using Tapatalk

[butters]

The BE-4 production ramp is likely a significant constraint on Vulcan and New Glenn launch opportunities in the coming years, regardless of when their maiden launches occur.
.... snip...
Neither of these launch system is likely to be firing on all cylinders from a cadence perspective until the late 2025 or 2026 timeframe.

It's 2021. The big shiny engine factory they broke ground in in 2019 is built and full of tooling.  Another 4 to 5 years to hit cadence?  Even for Blue that seems pretty unreasonably pessimistic.

Too bad ULA had been dragging their butts on engine recovery and reuse.

If it's any indication of BE-4 ramp-up expectations, NASA announced last week that Vulcan is being onboarded into LSP and that ULA will be eligible to bid on NASA launches beginning June 2025.

[baldusi]

The BE-4 production ramp is likely a significant constraint on Vulcan and New Glenn launch opportunities in the coming years, regardless of when their maiden launches occur.
.... snip...
Neither of these launch system is likely to be firing on all cylinders from a cadence perspective until the late 2025 or 2026 timeframe.

It's 2021. The big shiny engine factory they broke ground in in 2019 is built and full of tooling.  Another 4 to 5 years to hit cadence?  Even for Blue that seems pretty unreasonably pessimistic.

Too bad ULA had been dragging their butts on engine recovery and reuse.

You are making 2 assumptions:
1) That Blue are experts at actual mass-scale (for space applications) production. They aren't, they haven't been producing anything at a high rate. Having a factory and actually hitting a stride, are two different things and the latter is very hard to do it. It might take a few years.
2) That BE-4 is ready for high production. We've all heard how BE-4 was the actual pacing item for Vulcan development and they still don't have "flight ready" set of engines. Even after that, it might be "kinda ready" but need a lot of work to actually be ready for mass production.
Just look when there was a problem discovered on engines like the Chinese YF-77 turbopump (the LM-5 hydrolox sustainer). It will require an all-hands effort to do it in a year, and that would be work detracted from the mass production effort.
And even after all this, you have lead time. Lead time and certification are the killer problems here.

[baldusi]

Vulcan should be ramping up over the next few years, so more flights could be ordered off ULA. I'd assume Atlas V's were ordered to fly in the next few years where as Vulcan could take over 2024 to 2026.

Assuming that Atlas V 551 is maxed out in mass and volume Vulcan will let them fly a fair more. 7200 ft³ on Atlas V, 11200 ft³ on Vulcan (200 m³ vs 320m³); a 55% increase. Then mass is 18,850 kg to LEO on Atlas V551 and 27,200 kg to LEO for Vulcan Centaur 6; a 44% increase.

Given that the worst case is Vulcan could fly 72 per mission based on mass.
Mass is arbitrarily estimated at 18,850 kg / 50 satellites = 377 kg each.
Volume arbitrarily estimated at 200 m³ / 50 satellites = 4 m³ each.

New Glenn should be flying in this same time frame of 2024 to 2026. I estimated a long time ago that New Glenn has 580m³ of volume, which would make mass the primary constraint there at 45,000kg / 377 kg = 120 Sats.

In theory ULA could offer the 7 meter fairing 3 liquid core rocket they have teased. Taking at SWAG at the comparative payload increase between Delta IV 5,4 and Vulcan Centaur 4 that would be 48,000 kg to LEO. Roughly equal to New Glenn. A backup plan that I won't include.

2022-2024:
9 Atlas V 551's = 450 Sats

2024 to 2026:
5 Vulcan Centaur 6's = 360 Sats
7 New Glenn's = 840 Sats

Total: 1,650

For New Glenn it only would have to fly an average of 2.3 times a year. Once in 2024 then 3 times in 2025 and 3 times in the first half of 2026. That seems very achievable even for New Glenn's late arrival.

You can't take the LEO launch mass. You have to consider the inclination and altitude, plus add any payload adapter. I was the one who stated 50 because some where assuming 40. But let's focus on Atlas V a bit.
After looking into an old (2001) Atlas Payload Guide, I found out that Atlas V 552 (that second RL10 is important for LEO performance, and Centaur III DEC has flown with Starliner), can do about 19tonnes to 350km x ~39deg. The big question is if they have managed to make a self stacking design like StarLink, because if not, they will need at least 10% of their payload in payload adapter and probably lose a lot of volume.
So, we can do a few assumptions, as needing 40 (450kg), 35, 30 and 25 launches. That would correspond to 440kg, 390kg, 334kg and 277kg per satellite.
So, they already have 9 Atlas V launches. If we assume that 2022 will have 2 launches for a new LV, but only 1 dedicate to Kuiper, I get the following dedicated launches list.

Year	|2022|	2023|	2024|	2025|	2026|	Total
Dedicated launches|	1	3	8	13	6	31
Dedicated launches|	1	3	6	9	7	26
Dedicated launches|	1	2	5	8	5	21
Dedicated launches|	1	2	3	5	5	16

If we assume they won't have dedicated launches in 2022 and 2023, it kinda compresses to require that many extra launches in 24/25/26.

[Nomadd]

Be surprised if they didn't design engine pod to be easily detachable from day one. My guess for next step is separate engine pod during one of earlier missions. Want prove this works before adding expensive inflatable reentry shield.
They also need todo some MAR tests with for size engine pod mockup and heat shield. Similar to what RL have done.

If they can keep buying engines for $7 million each, I doubt if reuse will ever be a serious factor.

[bstrong]

How many sats could SLS launch? 

[TrevorMonty]

$14 for pair of engines plus most of valueable components and manufacturing effort from booster will be in pod. If they can half cost of booster its worth recovering especially as payload hit it is small.




Sent from my SM-G570Y using Tapatalk

[Nomadd]

How many sats could SLS launch? 

Zero

[TrevorMonty]

How many sats could SLS launch? 

Zero

Could. Lot more than Zero
Would. Zero

Sent from my SM-G570Y using Tapatalk

[bstrong]

How many sats could SLS launch? 

Zero

Could. Lot more than Zero
Would. Zero

I definitely agree, but I think it's interesting/ironic that at the $500M price (IIRC) that was floated for commercial launches, it would probably be competitive with most of the other options in this thread