I was on the “Cape” way back when. Our building was about as close as you could get for a Shuttle launch and still be allowed to be there.
Anyway…we are all on the roof for the launch…somebody has a radio going…3…2…1…main engine ignition…holey shit this is the stuff baby!..white clouds billowing outa the bowels the beast…but wait!..main engine shut down…blah blah blah…giant assed groaning sounds…I’m looking for a place to jump off the roof when parts start flying…good times…
I imagine they have the option of giving up on densified LOX and still getting the satellite into orbit, but giving up on 1st stage recovery? I wonder how many tries they get before they resort to that. (Though it’s not clear if the trouble today is related to super-cooling of the LOX)
For the SES-9 launch I don’t think SpaceX has an option of skipping propellant densification, at least for the second stage. The reserve impulse to land the first stage on the autonomous “drone ship” is actually quite small (as little as 2% to 3% of the liftoff propellant load, although that translates into more delta-V at the vehicle level because it is the last amount of propellant used) and has no impact on second stage other than the small amount of extra mass from densification. SES-9 is a high geosynch orbit that is just at the ragged edge of the capability of the Falcon 9v1.1 vehicle. Given the pressure for the SES-9 launch–the flight has been delayed five months from the original manifest launch date and is crucial to SES meeting their 2016 revenue forecast–I think SpaceX would forego densification for recovery to make up schedule at this point if it was feasible.
Stranger
Densification has some particular advantages for SpaceX, though. If you’re trying to eke out a bit more performance, one possibility is to stretch the stage. The F9 FT did this, but there are limits–it already has quite a high aspect ratio and they’re about as high as they’d like to be in that regard. Making the stage wider is not an option, not only because of tooling costs but because they’ve made it a priority for the F9 to be road-transportable, and it’s already as wide as it can be there. If you can’t go longer or wider, you have to go dense.
Also, subcooling has fewer downsides if you’re already focused on pad efficiency. They have other reasons for not hanging out on the pad for hours, and if you can manage that, you don’t have to worry so much about the propellant warming up.
Another thing is that densification gives them improved thrust. The F9 was already on the low side as far as acceleration at launch goes (IIRC, around 1.2 gees). Increasing propellant mass without increasing thrust could actually lower overall payload. Some rockets get away with adding high-thrust strap-on boosters, but that’s not SpaceX’s design philosophy. Subcooled propellant is still at least the same across all stages.
Still, we did see one downside today–some idiot in a boat can always cause a range violation, which almost certainly means a 30+ minute delay, which is apparently long enough for the propellant to warm and trigger an abort. I wonder if they could have two rockets worth of chilled propellant on hand, so that they can drain and refill if necessary.
Densification is not a new thing that SpaceX invented from whole cloth; several other contractors have attempted to develop systems around dense propellants (especially SLH[SUB]2[/SUB]/LO[SUB]X[/SUB] propulsion systems) and have found the very modest performance improvement to not be worth the cost and complexity of developing and using such a system. Cryogenic densification has numerous functional issues beyond the limited duration of post-loading launch availability, including thermal stratification which can lead to CTE overstress, water hammer induced shock, increased susceptibility to pogo oscillation, propellant utilization measurement errors, the need to carry additional inert mass in the form of insulation, et cetera. It is the only way to get an increase in total impulse out of a set propellant volume without changing engine performance but it comes with a very significant impact on availability, reliability, and processing effort.
The desire to minimize time on pad is understandable for numerous reasons, but the reality is that there are many uncontrollable environmental, security, and safety considerations that may prevent an expedient launch effort, including sudden increase in ground wind speed or wind shear aloft, lightning conditions, anomalies in the countdown sequence, communication issues, and of course, unauthorized ships or aircraft in the flight hazard zone. “Some idiot in a boat” (or a light aircraft, or a hot air balloon…yes, that has happened) is always going to be a potential issue, and frankly will be an even worse problem in the Brownsville site where SpaceX will have no legal authority to evacuate the downrange hazard zone even in US territorial waters, much less Mexican territorial waters. Launch recycles are also very costly; for Atlas V, a one-day recycle costs around US$500k in direct labor alone not withstanding downrange monitoring costs, refueling costs, et cetera, which probably sum up to well over US$1M per day spent on pad. I don’t know how much it costs SpaceX to recycle a vehicle, but even assuming a much lower labor wrap rate it is still likely in the mid-six figures for the all-up cost of a recycle, and also with impacts on their anticipated launch tempo.
I strongly suspect that SpaceX will end up limiting cryogenic densification to only the missions requiring highest performance (and with a price premium as a special service) if not abandoning it entirely.
Stranger
Most of the things you’ve mentioned are design issues, though, which presumably SpaceX has already worked through–it’s not an additional cost.
The exposure to delays (winds aloft, range violations, etc.) is a biggie, though. They can improve it by streamlining their pad ops even further, but I don’t know how much headroom they have. Currently they start fueling at ~T-30:00 and finish by T-10:00. The remaining time is taken by system checks. Could they improve this to, say 15 minutes total? No idea.
I think you’re probably half-right. I doubt they’ll abandon densification completely, but will work out a system that allows them to “soft fail” in case they spend too much time on the pad. Lighter payloads will fall back to undensified propellant and SpaceX will just accept the additional probability of a failed stage recovery.
There is an interesting tradeoff here, which is that since most GEO birds these days have ion engines, there is some wiggle room in performance. The satellite operator can evaluate the cost of a few extra months of insertion time and some loss in total lifetime against that of having the bird in orbit now (plus any discounts that SpaceX may choose to apply).
That would depend on SpaceX having a good model for performance as it’s affected by pad conditions over time. It seems likely that they have such models but I wonder if they’re “in the loop” to the extent that they can influence a go/no-go call.
The next launch attempt is scheduled for today, March 4, at 6:35 EST, when this post is about 4 hours old. Link.
Live feed just kicked in. Not sure if the launch is a go yet or not.
13 minutes to planned launch time.
The mission integration engineer is hot.
Just sayin’
The launch and satellite deployment was successful!
No final word on the landing, though the camera on the barge caught a few frames lit by the descending first stage. As in previous launches, the stream cut out right before the landing, what with everything tending to nudge directional antennas out of alignment…
I like her, too! Lauren Lyons is her name. She’s so smiley and friendly!
I am guessing that that first stage did not land on the barge like they were hoping? Seems odd that they haven’t been able to confirm that, some thirty minutes later, though.
#Falcon9 booster did not survive landing, confirmed by #SpaceX. #SES9
No official confirmation, though. Supposedly they used three engines instead of one for their “hoverslam”. They were very fuel-constrained, and maneuvers like this (hoverslam, suicide burn, etc.) are more efficient the more thrust you use. However, even with one engine they had a tiny margin of error, and with three the problem is that much worse.
All this to get a silly GTO?
Damned kids today!
Back in my time, we had all the REAL GTOs we needed!
And we Liked it!, Gol-durn it!
Does anyone know how the first stage and the barge do their “dance”?
I have to believe there is some pitch on the deck ,and there has to be some kind of feedback between the two.
Now, if only the engines were instantaneous and had infinitely adjustable thrust.
And then all you have to do is get the two to agree which side is “top”, “left”, etc.
When they do get the silly things to mate, I will be impressed.
But, just sitting here and watching a satellite being pushed away from a second stage rocket 150+ Km up over Africa, viewed on a machine that is cheaper than the TV my parents had AND has a larger screen. Using a technology they never ever dreamed of?
Nah, not at all impressive.
I was going to make a comment about clocks in automobiles, but I suspect there are about 5 of us who would understand it.
“We can put a man on the MOON, but we can’t get a car clock to work!”.
This may be the last time that expression is written.
More official:
Rocket landed hard on the droneship. Didn’t expect this one to work (v hot reentry), but next flight has a good chance.
For better or worse, there’s no longer a barrier to entry for software. Which means that if you’re only willing to pay the absolute minimum for your software, you’ll get minimum quality as well.
I take it you are not one of the 5 
I’ll wait a bit to see if any of the 5 cares to explain the “car clock” problem.
Hint: no software involved.
Apparently not :). I assumed you were talking about the Feb 30 thing.
OK - All five of then had their chance:
The first (built-in at the factory) auto clocks were electric motors - running on 6vDC supplied by a generator.
DC motors vary their RPM with fluctuations of current.
To say that those electrical systems “fluctuated” is a huge understatement.
So: Yes, there was a time when people were on the moon but no car clock in the country could keep time.
Hence:
“We can put a man on the MOON, but we can’t get a car clock to work!”.
Not by the voltage regulator?
It would have to see voltage when the engine wasn’t running.
The problem was:
Battery - which can vary all by itself
Generator - which varies with speed
I suspect the introduction of 12v and, later, the alternator, got the current MUCH more constant.
But a clock’s motor really can’t vary in speed AND keep anything resembling accurate time.
The “… and we STILL can’t make a car clock” became a standard throw-away line.
7/21/69 saw the first man on the moon. And the 1970 model year cars STILL had crappy clocks.