SpaceX Dragon XL

[tbellman]

Yes, in the commercial marketplace you can sell anything for any price. No one is disputing pricing in the commercial marketplace.

So what happens when the US government tries to buy something, and the lowest bidder is suspected of inflating their profit margin?  In a hypothetical situation where the USG wants to buy a bespoke frobnicator, the lowest bidder (company A) asks $2B, the second lowest bid (from company B) is $3B, and the USG personnel believe that company A could actually sell the frobnicator for just $1B.  The USG can try to haggle, but if company A is stubborn and don't want to lower their bid, they can't be forced to sell at a lower price, can they?

I suppose the USG can withdraw and redo the tender, in the hope that they will get a better bid next time, but they then risk that company A will not bid at all, or raising their bid, leaving the USG with bids higher than they originally got.

You could threaten the bidder that they will never ever get another contract unless they lower their price, but as public procurements are pretty strictly regulated in law (at least they are here in Sweden and the EU, and I assume they are in the USA as well), and can be challenged in court if a seller thinks they are treated unfairly, that's a rather empty threat.

[Karloss12]

Yes, in the commercial marketplace you can sell anything for any price. No one is disputing pricing in the commercial marketplace.

So what happens when the US government tries to buy something, and the lowest bidder is suspected of inflating their profit margin?  In a hypothetical situation where the USG wants to buy a bespoke frobnicator, the lowest bidder (company A) asks $2B, the second lowest bid (from company B) is $3B, and the USG personnel believe that company A could actually sell the frobnicator for just $1B.  The USG can try to haggle, but if company A is stubborn and don't want to lower their bid, they can't be forced to sell at a lower price, can they?

I suppose the USG can withdraw and redo the tender, in the hope that they will get a better bid next time, but they then risk that company A will not bid at all, or raising their bid, leaving the USG with bids higher than they originally got.

You could threaten the bidder that they will never ever get another contract unless they lower their price, but as public procurements are pretty strictly regulated in law (at least they are here in Sweden and the EU, and I assume they are in the USA as well), and can be challenged in court if a seller thinks they are treated unfairly, that's a rather empty threat.

If there are late penalties for this job then competitors to SpaceX will have done a commercial risk analysis and increased their price because they don't even have a launch vehicle at the time of bidding and can't be sure one would be available in time for launch.  Considering that SpaceX was the only awarded contract, the competition possibly didn't bid at all because they know they can't deliver on time.

The result is that SpaceX is in the advantageous position which tbellman describes above.

Government cost reviews can protect against a contract being 10 times overpriced.  But there are always easy explanations to get a project 3 times overpriced passed these cost reviews.  Everyone knows the bid is overpriced and there is nothing you can do about it because the project must go ahead. :-)

[edkyle99]

All evidence points to 8 tonnes being for flyback boosters and downrange core recovery.  The only other mode offered on the SpaceX web site and on the NASA NLS 2 mode description is fully expendable.  Just look at the first Block 5 Falcon Heavy result.  6,465 kg Arabsat 6A to a 327 x 89,815 x 22.96 degree supersynchronous orbit.  It all lines up, and has for awhile now.

 - Ed Kyle

Ah, I see you are making the classic Ed Kyle mistake again. That is: not believing the max. performance figures of a launch vehicle until that launch vehicle has - in fact - launched that max. mass to orbit.

We've had this discussion before, and it was cut short by the mods for all the right reasons: your line of reasoning is flawed.

No.  I believe the number, the 8,000 kg number given by SpaceX for reuse mode.  I also believe the much larger expendable mode numbers, though I wouldn't be surprise if some more vehicle improvement will be needed to get there.  I'm just pointing out that the actual flight result is entirely consistent with the number.  They put a smaller mass into orbit than 8,000 kg on that flight, but boosted it to a higher than standard GTO.  Falcon 9 results aligned with company stated performance this way too.

 - Ed Kyle

[Coastal Ron]

Yes, in the commercial marketplace you can sell anything for any price. No one is disputing pricing in the commercial marketplace.

So what happens when the US government tries to buy something, and the lowest bidder is suspected of inflating their profit margin?

For commodity items that already exist, and there is competition, the government doesn't care what the profit margin is of the supplier. They are looking for the best value to the American taxpayer - which is their job.

When the U.S. Government is asking contractors to build something unique, like a cargo transport vehicle for the Lunar Gateway, the U.S. Government has an interest in making sure they are not over paying.

I suppose the USG can withdraw and redo the tender, in the hope that they will get a better bid next time, but they then risk that company A will not bid at all, or raising their bid, leaving the USG with bids higher than they originally got.

We don't have to get into hypotheticals. In the case of the Gateway Logistics Services contract there would have been multiple bids, so the government would have first determined which were qualified to bid, and then they would have reviewed the bids submitted from a number of different aspects, including technical, cost, schedule, etc. Cost is not the only consideration, but it is an important one.

You could threaten the bidder that they will never ever get another contract unless they lower their price...

You are going WAY overboard here. As long as there is competition then the bidding process is usually pretty straightforward. And since everyone bidding on the Gateway Logistics Services contract likely has YEARS of experience bidding on government contracts, there is usually no drama. Everyone knows the rules.

...but as public procurements are pretty strictly regulated in law (at least they are here in Sweden and the EU, and I assume they are in the USA as well), and can be challenged in court if a seller thinks they are treated unfairly, that's a rather empty threat.

Losing bidders can sue, for various reasons. And sometimes they win. And it is safe to say that pretty much every aerospace company has sued the U.S. Government at one time or another over lost bids. Luckily those are the exceptions, and so far we haven't seen anyone protesting this contract award to SpaceX.

[Robotbeat]

Different in amount, not in kind. LEO is same as deep space, just different dosing rates. Same order of magnitude rates.

It's different in kind, too.  You don't get much galactic cosmic radiation inside the Van Allen Belts.  GCRs are highly energetic heavy ions.  That's certainly a different kind of hardening for electronics--and people.

I didn’t misspeak. You get plenty of GCR in LEO, too. Same order of magnitude. And it’s not the Van Allen belts, it’s the magnetic field that does the shielding. And that shielding is ineffective for the highest energy particles anyway.

[mikelepage]

Cross posting from the CRS2 thread.
Since NASA & SpaceX announced Dragon XL for gateway could they use this for ISS first?

Makes you now wonder about the private Dragon flight if it could dock to Dragon XL it could have an awesome week in space.

Miniature space station, like Tiangong.

I do wonder how many other customers there would be for this kind of thing.  Surely the creation of a class of small, temporary and private space station(s) in LEO would help SpaceX fund SS/SH development?

1) Anyone care to speculate/confirm whether a Dragon 2 can dock to a Dragon XL?
2) Could it be done autonomously? (e.g. so crew of two could just be payload specialists, or spaceflight participants).
3) Just because Dragon XL is not designed to re-enter the atmosphere need not make it "expendable" if it could be reused by multiple visiting vehicles.
4) If Dragon XL is being designed to host a Canadarm, could it retrieve payloads from the trunk of a visiting Dragon 2?
5) Would it be possible to use the visiting Dragon 2 as an airlock (i.e. shut the hatch between D2 and Dragon XL and open and close the main side hatch)?

There's probably more, but my point is, is SpaceX so confident that SS/SH is going to be ready for crewed flights in near future that they would overlook this potential goldmine using what will be heritage hardware?

EDIT: Most of the responses to this are in this new thread:
https://forum.nasaspaceflight.com/index.php?topic=50541.0

[woods170]

If DragonXL does a slow transfer to NRHO (months), how did SpaceX comply with the unique capability of a fast transfer to Gateway (<30 days)? If NASA request this unique capability.

The standard DragonXL launch and travel is not a fully expendable Falcon Heavy? And the fast transit is a fully expendable?

I' m a little bit lost.

Can you point to the specific requirement for a <30 days transfer to Gateway? I can't seem to find it.

File: Attachment_01_GLS_SOW.pdf
5.0 MISSION UNIQUE CAPABILITIES (SUBCLIN 103) [pages 29-30]

Fast Transit to Gateway: Provide end-to-end cargo delivery services withcargo transit time of 30 days or less. NASA shall have the ability to order this capability within the time defined in the GLS Space System Architecture contract attachment.

Perfect! Thank you. It is sometimes challenging to find specifics when an RFP has a kazillion attachments

It is a mission unique capability. Here is the first paragraph, which basically explains it all:

5.0 MISSION UNIQUE CAPABILITIES (SUBCLIN 103)

The contractor shall provide Mission Unique Capabilities for requirements over and above
logistic service requirements obtained under subCLINs 101 and 102.

These capabilities may be added throughout the life of the contract or they may be incorporated in
response to specific task order requirements.

So, these are non-standard capabilities which MAY be - someday - required by NASA. As such, they are not part of the baseline solution offered by SpaceX.
What I think (but I'll have to check with my sources) is that SpaceX is perfectly capable of sending Dragon XL on an expedited trajectory to Gateway, but at the cost of reduced cargo up-mass.

Another solution would probably be to to fly FH in fully expendable mode. Which will result in the launch being more expensive to NASA and thus the total number of launches under the fixed price cost cap decreasing.

[smoliarm]

#1
Different in amount, not in kind. LEO is same as deep space, just different dosing rates.
#2
Same order of magnitude rates.

It's different in kind, too.
#3
You don't get much galactic cosmic radiation inside the Van Allen Belts.  GCRs are highly energetic heavy ions.  That's certainly a different kind of hardening for electronics--and people.

I didn’t misspeak. You get plenty of GCR in LEO, too. Same order of magnitude. And it’s not the Van Allen belts, it’s the magnetic field that does the shielding. And that shielding is ineffective for the highest energy particles anyway.

The numbered statements (in bold) are inaccurate / incorrect (#2 and 3) or misleading (#1)

#1
Well, the border between *difference in amount* and *different kind* is pretty fuzzy, and it's a philosophical issue.
However, I can tell you for sure that statement #1 will make unhappy ALL editors/reviewers of "Meteoritics and Planetary Science" or "Geochimica Cosmochimica Acta". (In second case - may be somewhat angry

#3
Van Allen Belts have little effect on GCR, and almost all the shielding of GCR is done by... PLANET. So, in Low Earth Orbit we get one half from GCR compared to deep space, because GCR comes from EVERY direction, but in LEO half of the sphere blocked by Earth.
The same thing will be in Low Mars or Low Lunar Orbits.

#2 and #3 -- Van Allen belts
- they do interfere with Solar Wind - they are, in fact, composed of *solar wind matter* (solar protons and electrons).
 Van Allen belts work as filters for Solar electrons, and for some energy bands they are very effective filters.

The bottom line - for all types of radiation - LEO compared to deep space:
1. LEO has GCR reduced by ~ 50% (Earth's shielding);
2. LEO has Solar Wind reduced by 100% when in shadow (Earth's shielding);
3. LEO has Solar Wind selectively filtered when out of shadow (Van Allen belts)

[pochimax]

If NASA get over their aversion to Earth orbit rendezvous. SpaceX could just assembled and pre-positioned a large vehicle stack in LEO with enough hypergolic propellants for a fast transfer from LEO to the Gateway. The Dragon XL flies up on a Falcon 9 and docked with vehicle stack then head for the Moon.

Maybe the vehicle stack could be a modified Dragon XL variant only carrying hypergolic propellants with a vacuum optimized de-rated SuperDraco in place of the docking port on the bottom and a docking collar on the top.

I think it is too much complicated.

Easier if you can have a slow transfer with FH reusable (partially) and fast transfer (<30 days) with an expendable FH

(If it is the right and true FH lauch configuration of both options)

[TGebs15]

I am one of those technical and commercial experts who works with multi-million pound projects in various industries including Aerospace.  So this is a case of me (an expert) disagreeing with other so called experts.
I generate an accurate estimate for the design, fabrication and management of equipment at a customers site and a standard profit margin is applied to my estimate to get a standard bid price.  If we know that my company is the only bidder then the project manager will take my standard quote price and multiply it by 2-3.  There are other multipliers for whether you want the job.

Yes, in the commercial marketplace you can sell anything for any price. No one is disputing pricing in the commercial marketplace.

When you throw politicians into the mix like with Aerospace, the price just gets even worse.

Actually in the vast amount of government contracting politicians are not involved, and politicians can only really be involved through legislation, like with the SLS and Orion programs.

I think people on this forum over exaggerate the governments ability to intergate a suppliers price.

Many of us have first hand experience with government contracting, and what can and can't be done. Every American should feel good about the fact that government procurement professionals really are good at what they do, and it takes incompetence in agency leadership, or specific legislation by politicians (i.e. pork) to allow gross profiteering.

The most recent article says that Dragon XL is a Variant of Dragon 2.
Going from the Dragon 1 variant to the Dragon 2 Variant cost over $2bill.

SpaceX received a $2.6B contract for Commercial Crew, and as of October of 2019 Elon Musk felt they would not exceed that contract value.

I think with a lack of competition and a tight schedule squeeze, a similar price could be justified for the Dragon XL varient.

Lack of competition? SpaceX responded to a public RFP that NASA issued last year for a 15-year, $7B commercial supply services IDIQ contract. They had competition. And all the finalists would have had a cost review so that NASA could ensure that the eventual winner wasn't low-balling the bid - which does happen, and which is why the government does contract audits.

No doubt SpaceX had an advantage due to their extensive spacecraft experience, and due to the fact that they can launch their own vehicles - that cuts out one layer of extra profit.

The marketplace drives competitive pricing, not so for limited competition government contracts. The thing about government contractring is that it is controlled by contracting officers, not engineers, and lots of the time is specialized enough that three bids is considered big competition.

When RFP's are put out for systems that are needed and you've got one, two, or three options you are pretty much stuck with those options. A review of the options will be done. For smaller contracts that often is just a contracting officer looking at the bids and selecting the lowest of the group, if they have questions then they go to the engineers. For big contracts engineers are utilized and a better sense of value can be determined. However, when there is limited competion you have to pick from the options available. Even if an option is "overpriced" for what is being done, if they can do what they say they can do and that price is far under the competition the government has a duty to go with it since it is still best value. Government can't dictate price, they can only walk away and that is of rare occurance.

Also just as a side note, regardless of the capability of the individuals that work for the government they are hindered by the system that moves slow and inefficiently.

[dror]

All evidence points to 8 tonnes being for flyback boosters and downrange core recovery.  The only other mode offered on the SpaceX web site and on the NASA NLS 2 mode description is fully expendable.  Just look at the first Block 5 Falcon Heavy result.  6,465 kg Arabsat 6A to a 327 x 89,815 x 22.96 degree supersynchronous orbit.  It all lines up, and has for awhile now.

 - Ed Kyle

Ah, I see you are making the classic Ed Kyle mistake again. That is: not believing the max. performance figures of a launch vehicle until that launch vehicle has - in fact - launched that max. mass to orbit.

We've had this discussion before, and it was cut short by the mods for all the right reasons: your line of reasoning is flawed.

No.  I believe the number, the 8,000 kg number given by SpaceX for reuse mode.  I also believe the much larger expendable mode numbers, though I wouldn't be surprise if some more vehicle improvement will be needed to get there.  I'm just pointing out that the actual flight result is entirely consistent with the number.  They put a smaller mass into orbit than 8,000 kg on that flight, but boosted it to a higher than standard GTO.  Falcon 9 results aligned with company stated performance this way too.

 - Ed Kyle

Bold mine
Please stop saying that unless you can show for it. 
As much as we all saw, Spacex gave this number for a specific flight price. They never mentioned the flight profile or manifesting for that spesific price.

[edkyle99]

No.  I believe the number, the 8,000 kg number given by SpaceX for reuse mode.  I also believe the much larger expendable mode numbers, though I wouldn't be surprise if some more vehicle improvement will be needed to get there.  I'm just pointing out that the actual flight result is entirely consistent with the number.  They put a smaller mass into orbit than 8,000 kg on that flight, but boosted it to a higher than standard GTO.  Falcon 9 results aligned with company stated performance this way too.

 - Ed Kyle

Bold mine
Please stop saying that unless you can show for it. 
As much as we all saw, Spacex gave this number for a specific flight price. They never mentioned the flight profile or manifesting for that spesific price.

No one wants to believe, and yet ....
https://forum.nasaspaceflight.com/index.php?topic=50467.msg2063374#msg2063374
etc.
https://space.skyrocket.de/doc_lau_det/falcon-9-heavy_b5.htm

 - Ed Kyle

[Robotbeat]

#1
Different in amount, not in kind. LEO is same as deep space, just different dosing rates.
#2
Same order of magnitude rates.

It's different in kind, too.
#3
You don't get much galactic cosmic radiation inside the Van Allen Belts.  GCRs are highly energetic heavy ions.  That's certainly a different kind of hardening for electronics--and people.

I didn’t misspeak. You get plenty of GCR in LEO, too. Same order of magnitude. And it’s not the Van Allen belts, it’s the magnetic field that does the shielding. And that shielding is ineffective for the highest energy particles anyway.

The numbered statements (in bold) are inaccurate / incorrect (#2 and 3) or misleading (#1)

#1
Well, the border between *difference in amount* and *different kind* is pretty fuzzy, and it's a philosophical issue.
However, I can tell you for sure that statement #1 will make unhappy ALL editors/reviewers of "Meteoritics and Planetary Science" or "Geochimica Cosmochimica Acta". (In second case - may be somewhat angry

#3
Van Allen Belts have little effect on GCR, and almost all the shielding of GCR is done by... PLANET. So, in Low Earth Orbit we get one half from GCR compared to deep space, because GCR comes from EVERY direction, but in LEO half of the sphere blocked by Earth.
The same thing will be in Low Mars or Low Lunar Orbits.

#2 and #3 -- Van Allen belts
- they do interfere with Solar Wind - they are, in fact, composed of *solar wind matter* (solar protons and electrons).
 Van Allen belts work as filters for Solar electrons, and for some energy bands they are very effective filters.

The bottom line - for all types of radiation - LEO compared to deep space:
1. LEO has GCR reduced by ~ 50% (Earth's shielding);
2. LEO has Solar Wind reduced by 100% when in shadow (Earth's shielding);
3. LEO has Solar Wind selectively filtered when out of shadow (Van Allen belts)

I don’t think you realize this, but... You confirmed my points.

You confirmed GCR is still present, as I said. Just reduced in dose rate. 50% reduction is still same order of magnitude.

Didn’t deny Earth shadowing. I was responding to the claim the “Van Allen Belts” shield (they don’t).

The main point is this: people falsely assume that LEO experience is useless in deep space. False, this is one of the same falsehoods that Moon hoaxers hold. A difference in dose rates by less than an order of magnitude means you STILL have to understand and address the issue and gain experience with the impact of that radiation type.

Thank you for confirming my point.

[dror]

No.  I believe the number, the 8,000 kg number given by SpaceX for reuse mode.  I also believe the much larger expendable mode numbers, though I wouldn't be surprise if some more vehicle improvement will be needed to get there.  I'm just pointing out that the actual flight result is entirely consistent with the number.  They put a smaller mass into orbit than 8,000 kg on that flight, but boosted it to a higher than standard GTO.  Falcon 9 results aligned with company stated performance this way too.

 - Ed Kyle

Bold mine
Please stop saying that unless you can show for it. 
As much as we all saw, Spacex gave this number for a specific flight price. They never mentioned the flight profile or manifesting for that spesific price.

No one wants to believe, and yet ....
https://forum.nasaspaceflight.com/index.php?topic=50467.msg2063374#msg2063374
etc.
https://space.skyrocket.de/doc_lau_det/falcon-9-heavy_b5.htm

 - Ed Kyle

So why didn't you just say so ?
 
Thanks,
and sorry

[TheRadicalModerate]

#1
Different in amount, not in kind. LEO is same as deep space, just different dosing rates.
#2
Same order of magnitude rates.

It's different in kind, too.
#3
You don't get much galactic cosmic radiation inside the Van Allen Belts.  GCRs are highly energetic heavy ions.  That's certainly a different kind of hardening for electronics--and people.

I didn’t misspeak. You get plenty of GCR in LEO, too. Same order of magnitude. And it’s not the Van Allen belts, it’s the magnetic field that does the shielding. And that shielding is ineffective for the highest energy particles anyway.

The numbered statements (in bold) are inaccurate / incorrect (#2 and 3) or misleading (#1)

#1
Well, the border between *difference in amount* and *different kind* is pretty fuzzy, and it's a philosophical issue.
However, I can tell you for sure that statement #1 will make unhappy ALL editors/reviewers of "Meteoritics and Planetary Science" or "Geochimica Cosmochimica Acta". (In second case - may be somewhat angry

#3
Van Allen Belts have little effect on GCR, and almost all the shielding of GCR is done by... PLANET. So, in Low Earth Orbit we get one half from GCR compared to deep space, because GCR comes from EVERY direction, but in LEO half of the sphere blocked by Earth.
The same thing will be in Low Mars or Low Lunar Orbits.

#2 and #3 -- Van Allen belts
- they do interfere with Solar Wind - they are, in fact, composed of *solar wind matter* (solar protons and electrons).
 Van Allen belts work as filters for Solar electrons, and for some energy bands they are very effective filters.

The bottom line - for all types of radiation - LEO compared to deep space:
1. LEO has GCR reduced by ~ 50% (Earth's shielding);
2. LEO has Solar Wind reduced by 100% when in shadow (Earth's shielding);
3. LEO has Solar Wind selectively filtered when out of shadow (Van Allen belts)

I accept your correction on my "VAB" nomenclature.  I was thinking "magnetic field" but "VAB" came out of the typing fingers.

The real question here, per the original issue raised, is whether SpaceX needs to harden their electronics for Starship above and beyond what they do for F9 and FH.  I'm guessing that a GCR hit on an integrated circuit is pretty catastrophic, so a doubling of GCR flux comes pretty close to a difference in kind.  ("Quantity has a quality all of its own.")  But I agree that the dividing line between "kind" and "degree" here is pretty fuzzy.

[Robotbeat]

Separate topic: Dragon XL (without the pressure vessel, with some body-mounted panels and stretched tanks) would make a good recoverable third stage for Starship.

Would allow Starship to do single-launch fully recoverable launches to direct GSO.

The Dragon XL with the payload on top would stage at GTO out of Starship, do the burn into GSO, separate payload, then go back to GTO (phasing may take time?), and dock with Starship/Chomper and come back to Earth. Full reuse, single-launch direct to GSO. Could probably do escape missions this way, too.

Likewise, this is how Dragon XL could be launched and recovered from Starship for Gateway missions. Those folding solar panels (necessary for independent power on Gateway—body panels aren’t enough) will need to be either foldable or expended.

[TheRadicalModerate]

Separate topic: Dragon XL (without the pressure vessel, with some body-mounted panels and stretched tanks) would make a good recoverable third stage for Starship.

Would allow Starship to do single-launch fully recoverable launches to direct GSO.

The Dragon XL with the payload on top would stage at GTO out of Starship, do the burn into GSO, separate payload, then go back to GTO (phasing may take time?), and dock with Starship/Chomper and come back to Earth. Full reuse, single-launch direct to GSO. Could probably do escape missions this way, too.

Likewise, this is how Dragon XL could be launched and recovered from Starship for Gateway missions. Those folding solar panels (necessary for independent power on Gateway—body panels aren’t enough) will need to be either foldable or expended.

I think you're massively overestimating the delta-v capabilities of the DXL.  Just to do the GTO-to-GEO portion of the trip, you'd need 1800 m/s of delta-v, and then another 1800 m/s to return to GTO empty.  In contrast, the NASA requirements don't require any more than about 700 m/s, and that's pretty generous.  SpaceX isn't going to size the tankage on the DXL for 3600 m/s mission.

BTW, I did a quick back-of-napkin on what you could do with an Isp=300 system if Starship can take 21 t to GTO, assuming a 3 t dry mass for DXL.  To do GTO-GEO-GTO, your max payload is 5.8 t, and you need 9.6 t of prop. 

In contrast, the same 5.8 t payload with the same 3 t dry mass can go loaded to NRHO (430 m/s) and dispose of itself (270 m/s), for 1.4 t of prop.

[Robotbeat]

Separate topic: Dragon XL (without the pressure vessel, with some body-mounted panels and stretched tanks) would make a good recoverable third stage for Starship.

Would allow Starship to do single-launch fully recoverable launches to direct GSO.

The Dragon XL with the payload on top would stage at GTO out of Starship, do the burn into GSO, separate payload, then go back to GTO (phasing may take time?), and dock with Starship/Chomper and come back to Earth. Full reuse, single-launch direct to GSO. Could probably do escape missions this way, too.

Likewise, this is how Dragon XL could be launched and recovered from Starship for Gateway missions. Those folding solar panels (necessary for independent power on Gateway—body panels aren’t enough) will need to be either foldable or expended.

I think you're massively overestimating the delta-v capabilities of the DXL.  Just to do the GTO-to-GEO portion of the trip, you'd need 1800 m/s of delta-v, and then another 1800 m/s to return to GTO empty.  In contrast, the NASA requirements don't require any more than about 700 m/s, and that's pretty generous.  SpaceX isn't going to size the tankage on the DXL for 3600 m/s mission.

BTW, I did a quick back-of-napkin on what you could do with an Isp=300 system if Starship can take 21 t to GTO, assuming a 3 t dry mass for DXL.  To do GTO-GEO-GTO, your max payload is 5.8 t, and you need 9.6 t of prop. 

In contrast, the same 5.8 t payload with the same 3 t dry mass can go loaded to NRHO (430 m/s) and dispose of itself (270 m/s), for 1.4 t of prop.

But propellant is cheap and few payloads would need more than 5.8 tons direct to GSO. Delta IV Heavy can do 6.75to direct GSO. Also, GTO is really high for Starship as Starship has super high dry mass, so staging earlier might work.

As well as an extended bell for Dracos to improve Isp.

[envy887]

[...]
And on the way back in "garbage truck" mode... take a month or two to reenter in a controlled manner.
[...]

Why would it need to reenter for disposal? BEO or Moon-crashing would most probably be cheaper, delta-v wise.

It's always bothered me when "we" leave our "stuff" in places they should not be left... JMHO...
I smile every time a spacecraft saves enough fuel, or puts itself in a <90day to re-entry orbit...to clean up after itself... 
Delta-v be damned... figure out how to make it work before you launch it... or just don't...
Sacrifice payload or as a last resort, expend the rocket if you have to...
Because someday, someone will have to clean up the mess we have made over the last 60+years in space...
Again... JMHO... 

Space is big, I mean really big.

The same kind of reasoning was used when some countries started dumping nuclear waste in the oceans in the 1960s.

What a civilized bunch of people is supposed to do is properly clean up the mess they made. Just randomly dumping Gateway waste on the lunar surface is NOT the way to go. Neither is sending it into lunar orbit or solar orbit.

Solar orbits are fine for disposal. If we have to start recollecting them for some reason, we will also have to collect trillions of asteroids.

[Nathan2go]

Separate topic: Dragon XL (without the pressure vessel, with some body-mounted panels and stretched tanks) would make a good recoverable third stage for Starship.

Would allow Starship to do single-launch fully recoverable launches to direct GSO.
...

A third stage would increase the single-launch versatility, but I think direct GSO is a bad application.  It turns out that every single geosynchronous satellite needs station-keeping engines in order to be geostationary;  all they need is a little more fuel, and they can do their own apogee burn for little extra cost.  Solar-electric propulsion makes this even more attractive.  Commercial companies all understand this.  The fact that the Air Force is still asking for GSO missions is a sign of bureaucratic in inefficiency.

The use of super heavy (exploration class) vehicles for launching light cargo means more ride-sharing.  For payloads needing escape trajectories, that implies they should be designed with a kick motor, so they can be dropped off in GTO with everybody else (e.g. the Nusantara/Beresheet mission).  In fact, we may see payloads being designed to be dropped off mid-way between LEO and GTO, as this boosts the payload capacity to escape orbits substantially.