Author Topic: Reuters: Musk shakes up SpaceX in race to make satellite launch window: sources  (Read 9881 times)

Offline Asteroza

  • Full Member
  • ****
  • Posts: 762
  • Liked: 121
  • Likes Given: 2
Quote
When contacted by Ars, SpaceX provided this statement:
The SpaceX Redmond office is an essential part of the company's efforts to build a next-generation satellite network that can link the world with reliable and affordable broadband service, reaching those who have never been connected before. Given the success of our recent Starlink demonstration satellites, we have incorporated lessons learned and re-organized to allow for the next design iteration to be flown in short order. This is a very similar approach of rapid iteration in design and testing which led to the success of Falcon 1, Falcon 9, Falcon Heavy and Dragon.
https://arstechnica.com/information-technology/2018/10/unhappy-elon-musk-went-on-firing-spree-over-slow-satellite-broadband-progress/

This would imply at least (and maybe no more than) one more Tintin  iteration flying before the main sat push in 2019. So 1 versus the previously mentioned desired 3 iterations. This, plus the persistent rumors of the current Tintins not doing well, pushes a lot eggs into 1 iteration basket.

Online docmordrid

  • Senior Member
  • *****
  • Posts: 5451
  • Michigan
  • Liked: 2839
  • Likes Given: 1
>
This would imply at least (and maybe no more than) one more Tintin  iteration flying before the main sat push in 2019. So 1 versus the previously mentioned desired 3 iterations. This, plus the persistent rumors of the current Tintins not doing well, pushes a lot eggs into 1 iteration basket.

Those rumors were denied in this report, and since.

Space Intel Report...
« Last Edit: 11/01/2018 03:19 am by docmordrid »
DM

Offline oldAtlas_Eguy

  • Senior Member
  • *****
  • Posts: 3607
  • Florida
  • Liked: 2060
  • Likes Given: 246
By EOY 2020 production rates should be about 500 sats per year.  2019 is the production startup. The R&D phase is ending. Production hundreds of units is very different than producing a few units. It requires a different managerial goals and experience.

Offline Asteroza

  • Full Member
  • ****
  • Posts: 762
  • Liked: 121
  • Likes Given: 2
>
This would imply at least (and maybe no more than) one more Tintin  iteration flying before the main sat push in 2019. So 1 versus the previously mentioned desired 3 iterations. This, plus the persistent rumors of the current Tintins not doing well, pushes a lot eggs into 1 iteration basket.

Those rumors were denied in this report, and since.

Space Intel Report...

I wouldn't necessarily call that a complete denial. Though to be fair, with the rumors still persisting after that, either there is a problem of some kind (perhaps partial failure, such as one antenna going bad out of several, or god forbid a bad firmware update hosing a workable piece of equipment), or there is an active FUD campaign by a competitor stakeholder (effectively overblowing problems one would normally encounter in a test sat)...

Offline Robotbeat

  • Senior Member
  • *****
  • Posts: 28850
  • Minnesota
  • Liked: 9025
  • Likes Given: 5793
There's an active FUD campaign by a single Twitter guy who is a satellite industry "analyst." He hasn't stopped, and he probably will never stop.
Chris  Whoever loves correction loves knowledge, but he who hates reproof is stupid.

To the maximum extent practicable, the Federal Government shall plan missions to accommodate the space transportation services capabilities of United States commercial providers. US law http://goo.gl/YZYNt0

Online ncb1397

  • Full Member
  • ****
  • Posts: 1484
  • Liked: 651
  • Likes Given: 7
>
This would imply at least (and maybe no more than) one more Tintin  iteration flying before the main sat push in 2019. So 1 versus the previously mentioned desired 3 iterations. This, plus the persistent rumors of the current Tintins not doing well, pushes a lot eggs into 1 iteration basket.

Those rumors were denied in this report, and since.

Space Intel Report...

But the rumors are corroborated by this news. If there was no problems with the current sats, the project management wouldn't have been pushing for 3 generations of test satellites to incrementally fix the issues. If they were performing beautifully, the management team probably wouldn't get sacked. The claim (without access to the whole article) essentially boils down to that they are in the right orbit and that they are sending signals. Breaking that down into a.)being in the right orbit and b.)sending signals is a pretty low bar to clear. Those things are necessary but not sufficient for a global cross linked constellation of thousands of satellites with massive combined bandwidth and low latency. For one thing, the orbit is more or less pre-determined after initial deployment no matter how the satellites are performing barring a rare malfunction with uncommanded thruster firing.

Offline Coastal Ron

  • Senior Member
  • *****
  • Posts: 4645
  • I live... along the coast
  • Liked: 3330
  • Likes Given: 4550
By EOY 2020 production rates should be about 500 sats per year.  2019 is the production startup. The R&D phase is ending.

I'm not sure we know that, do we?

If Musk was firing staff in summer, then the fruits of that firing needs to be fielded before they can feel confident they have identified their final design. Unless the first two that were launched turned out to have met their total needs for the constellation.

Unless I've missed something...  ::)

Quote
Production hundreds of units is very different than producing a few units. It requires a different managerial goals and experience.

Right, and not just in manufacturing, but all of the other disciplines have to staff up. For instance, they need scheduling and production control (my area of expertise), contracts and procurement, QA and QC, inventory control, manufacturing engineering, facilities, and a production engineering staff. Just to name some of the big ones.

I haven't been watching their careers page, so we see them adding such people?
If we don't continuously lower the cost to access space, how are we ever going to afford to expand humanity out into space?

Offline EspenU

  • Newbie Spacegeek
  • Member
  • Posts: 79
  • Norway
  • Liked: 78
  • Likes Given: 13
But the rumors are corroborated by this news. If there was no problems with the current sats, the project management wouldn't have been pushing for 3 generations of test satellites to incrementally fix the issues. If they were performing beautifully, the management team probably wouldn't get sacked. The claim (without access to the whole article) essentially boils down to that they are in the right orbit and that they are sending signals. Breaking that down into a.)being in the right orbit and b.)sending signals is a pretty low bar to clear. Those things are necessary but not sufficient for a global cross linked constellation of thousands of satellites with massive combined bandwidth and low latency. For one thing, the orbit is more or less pre-determined after initial deployment no matter how the satellites are performing barring a rare malfunction with uncommanded thruster firing.
That would be true if these were fully operational production sats, but they are not. They are test units, which means that they may function as planned, but that they want to make improvements and launch another test set. I would be very surprised if there was no need to improve the first test sats.
Any disagreement on how many test runs are required can be entirely caused by difference in development philosophy.

Offline NWade

  • Member
  • Posts: 56
  • United States
  • Liked: 69
  • Likes Given: 550
But the rumors are corroborated by this news. If there was no problems with the current sats, the project management wouldn't have been pushing for 3 generations of test satellites to incrementally fix the issues. If they were performing beautifully, the management team probably wouldn't get sacked. The claim (without access to the whole article) essentially boils down to that they are in the right orbit and that they are sending signals. Breaking that down into a.)being in the right orbit and b.)sending signals is a pretty low bar to clear. Those things are necessary but not sufficient for a global cross linked constellation of thousands of satellites with massive combined bandwidth and low latency. For one thing, the orbit is more or less pre-determined after initial deployment no matter how the satellites are performing barring a rare malfunction with uncommanded thruster firing.
That would be true if these were fully operational production sats, but they are not. They are test units, which means that they may function as planned, but that they want to make improvements and launch another test set. I would be very surprised if there was no need to improve the first test sats.
Any disagreement on how many test runs are required can be entirely caused by difference in development philosophy.

ncb1397 - We don't know the "3 iterations" were specifically to fix issues.  It could simply be that there were competing ideas about how to package some components on the satellite. Or deal with thermal loads. Or check a novel array/antenna deployment mechanism. Or any number of things - developing physical products is harder and more-complicated than most people think (especially when you cannot inspect the items if/when they go wrong)!
 
The items they wanted to iterate could have nothing to do with the satellite's main function. They could be related to making systems easier-to-manufacture or cheaper to assemble (and you want to test those things because they might carry some additional risk-of-failure, or never-used-in-space-before component). Heck, they may never have intended to launch those iterations - just use them as a means to vet the manufacturing process and/or check the consistency of their components/systems/standards/practices/logistics-pipeline/suppliers/etc.  [Did I mention that making physical products is complicated?]
 
One final thing to point out: As a thought-experiment, consider that perhaps the former management _actually_ thought they could technically do it with fewer iterations. Maybe they felt the need to add a safety margin to their approach - either to prove themselves to senior leadership, or to ensure that their solution wouldn't fail & embarrass senior leadership. As middle/upper managers they only have so much authority, and there are always issues of trust & office politics to consider both up and down the managerial chain (even the best of companies have some of this; its human nature). For those reasons, they might've felt it was necessary to pursue this multiple-iteration direction (and indeed it might have truly been necessary, at the time they started down that road). However, Elon has the ultimate authority and so he can come in and declare a specific approach where others simply would not have the authority or the political capital to do so. And by declaring this change himself, he absorbs some of the responsibility for whether or not that decision pans out. Staff running down the chain don't have to worry (or argue) about whether its right or best, or whether the people at the top are OK with it. The course has been charted and now everyone either rows in that direction or they get off the boat.

In an ideal scenario the company culture is great and the head of the organization can make those mandates known, watch them propagate outward & downward, and not have to fire people. But the real world is rarely that clean/clear-cut and it is doubtful that we'll ever know all of the various factors that went into the decision. The important thing to remember is that we don't know all of the details, and thus we should avoid interpreting too much detail from these tea-leaves.
 
--Noel

Offline wes_wilson

  • Armchair Rocketeer
  • Full Member
  • *
  • Posts: 136
  • Florida
  • Liked: 65
  • Likes Given: 106
But the rumors are corroborated by this news. If there was no problems with the current sats, the project management wouldn't have been pushing for 3 generations of test satellites to incrementally fix the issues. If they were performing beautifully, the management team probably wouldn't get sacked. The claim (without access to the whole article) essentially boils down to that they are in the right orbit and that they are sending signals. Breaking that down into a.)being in the right orbit and b.)sending signals is a pretty low bar to clear. Those things are necessary but not sufficient for a global cross linked constellation of thousands of satellites with massive combined bandwidth and low latency. For one thing, the orbit is more or less pre-determined after initial deployment no matter how the satellites are performing barring a rare malfunction with uncommanded thruster firing.
That would be true if these were fully operational production sats, but they are not. They are test units, which means that they may function as planned, but that they want to make improvements and launch another test set. I would be very surprised if there was no need to improve the first test sats.
Any disagreement on how many test runs are required can be entirely caused by difference in development philosophy.

Just curious, the FCC requires a number of satellites by a date for the launch spectrum license?  Do those satellites have to be 100% functional, simply present in orbit, or is there an acceptance criteria somewhere between in orbit and 100% functional to count?  Just curious if there could be a cost/benefit trade to getting less capable or even only semi-functional devices up just to preserve the FCC license?  Even if they have to be replaced soon afterwards.

edit: launches <> spectrum...
« Last Edit: 11/02/2018 09:49 pm by wes_wilson »
@SpaceX "When can I buy my ticket to Mars?"

Online ChrisWilson68

  • Senior Member
  • *****
  • Posts: 3875
  • Sunnyvale, CA
  • Liked: 2595
  • Likes Given: 3346
But the rumors are corroborated by this news. If there was no problems with the current sats, the project management wouldn't have been pushing for 3 generations of test satellites to incrementally fix the issues. If they were performing beautifully, the management team probably wouldn't get sacked.

No.  People getting fired is not evidence that the test satellites are not performing well.

The test satellites were never intended to have all the features the eventual production satellites will have.  So having them performing flawlessly doesn't mean there isn't a lot more work to do.  It's perfectly plausible -- likely, even -- that the firings, if they in fact happened, were about the future direction of the program, not about the test satellites already in orbit.

Offline woods170

  • IRAS fan
  • Senior Member
  • *****
  • Posts: 9033
  • IRAS fan
  • The Netherlands
  • Liked: 6361
  • Likes Given: 2192
Don't know why are people getting so rattled about managers getting fired.
It is normal on fast paced companies. A manager can be very competent, but unfit for a specific strategy.


Yes. I've seen that happen multiple times in the business I operate in (IT industry).

Offline IRobot

  • Full Member
  • ****
  • Posts: 1296
  • Portugal & Germany
  • Liked: 297
  • Likes Given: 262
But the rumors are corroborated by this news. If there was no problems with the current sats, the project management wouldn't have been pushing for 3 generations of test satellites to incrementally fix the issues. If they were performing beautifully, the management team probably wouldn't get sacked.

No.  People getting fired is not evidence that the test satellites are not performing well.

The test satellites were never intended to have all the features the eventual production satellites will have.  So having them performing flawlessly doesn't mean there isn't a lot more work to do.  It's perfectly plausible -- likely, even -- that the firings, if they in fact happened, were about the future direction of the program, not about the test satellites already in orbit.
Musk takes a lot of inspiration from the Apollo program and I am sure this was an Apollo 4 moment, when an "all-up" test was decided instead of an incremental approach, to meet deadlines.

Managers pushed it for incremental approach, they got fired. That's how I see it.

Personally as a manager in that situation I would do the following:

a) if it possible to reduce the number of generations to keep the deadlines, albeit with more risk, I have to work torward a plan that makes it happen and attempts to reduce the risk.

b) If I know for sure this won't work, then I have to properly communicate to top management and wait to see if they are reasonable persons.

So the question here is who was stubburn: Musk, because he refused to hear his manager's concerns, or the managers, because they refused to draw up a plan with more risk but that could fulfill the deadlines?

Offline JamesH65

  • Full Member
  • ****
  • Posts: 1009
  • Liked: 719
  • Likes Given: 8
But the rumors are corroborated by this news. If there was no problems with the current sats, the project management wouldn't have been pushing for 3 generations of test satellites to incrementally fix the issues. If they were performing beautifully, the management team probably wouldn't get sacked.

No.  People getting fired is not evidence that the test satellites are not performing well.

The test satellites were never intended to have all the features the eventual production satellites will have.  So having them performing flawlessly doesn't mean there isn't a lot more work to do.  It's perfectly plausible -- likely, even -- that the firings, if they in fact happened, were about the future direction of the program, not about the test satellites already in orbit.
Musk takes a lot of inspiration from the Apollo program and I am sure this was an Apollo 4 moment, when an "all-up" test was decided instead of an incremental approach, to meet deadlines.

Managers pushed it for incremental approach, they got fired. That's how I see it.

Personally as a manager in that situation I would do the following:

a) if it possible to reduce the number of generations to keep the deadlines, albeit with more risk, I have to work torward a plan that makes it happen and attempts to reduce the risk.

b) If I know for sure this won't work, then I have to properly communicate to top management and wait to see if they are reasonable persons.

So the question here is who was stubburn: Musk, because he refused to hear his manager's concerns, or the managers, because they refused to draw up a plan with more risk but that could fulfill the deadlines?

Think its more likely to be Musk realising that although it could be done in one generation, the management were incapable of implementing it due to their previous experience being in more lengthy timescales. ie entrenched attitudes.

Offline speedevil

  • Senior Member
  • *****
  • Posts: 3306
  • Fife
  • Liked: 1767
  • Likes Given: 2143
Think its more likely to be Musk realising that although it could be done in one generation, the management were incapable of implementing it due to their previous experience being in more lengthy timescales. ie entrenched attitudes.
Spending a hundred million on a 50% gamble to accelerate the project by a year may be quite justified at this point.
This is not the kind of decision most managers would take.

Offline RoboGoofers

  • Full Member
  • ****
  • Posts: 468
  • NJ
  • Liked: 342
  • Likes Given: 304
But the rumors are corroborated by this news. If there was no problems with the current sats, the project management wouldn't have been pushing for 3 generations of test satellites to incrementally fix the issues. If they were performing beautifully, the management team probably wouldn't get sacked. The claim (without access to the whole article) essentially boils down to that they are in the right orbit and that they are sending signals. Breaking that down into a.)being in the right orbit and b.)sending signals is a pretty low bar to clear. Those things are necessary but not sufficient for a global cross linked constellation of thousands of satellites with massive combined bandwidth and low latency. For one thing, the orbit is more or less pre-determined after initial deployment no matter how the satellites are performing barring a rare malfunction with uncommanded thruster firing.
That would be true if these were fully operational production sats, but they are not. They are test units, which means that they may function as planned, but that they want to make improvements and launch another test set. I would be very surprised if there was no need to improve the first test sats.
Any disagreement on how many test runs are required can be entirely caused by difference in development philosophy.

Just curious, the FCC requires a number of satellites by a date for the launch spectrum license?  Do those satellites have to be 100% functional, simply present in orbit, or is there an acceptance criteria somewhere between in orbit and 100% functional to count?  Just curious if there could be a cost/benefit trade to getting less capable or even only semi-functional devices up just to preserve the FCC license?  Even if they have to be replaced soon afterwards.

edit: launches <> spectrum...

that's my question too. why waste two more launches on a handful of test sats when you can send up ~25 per launch. the first sats will only stay in orbit for a short time and there will be so many of them, if the first 50 perform poorly or die early, no biggie. That's especially true if they count toward the total sat count for the license.
« Last Edit: 11/06/2018 04:26 pm by RoboGoofers »

Offline IRobot

  • Full Member
  • ****
  • Posts: 1296
  • Portugal & Germany
  • Liked: 297
  • Likes Given: 262
Think its more likely to be Musk realising that although it could be done in one generation, the management were incapable of implementing it due to their previous experience being in more lengthy timescales. ie entrenched attitudes.
Spending a hundred million on a 50% gamble to accelerate the project by a year may be quite justified at this point.
This is not the kind of decision most managers would take.
Most managers think first about their job, then much later about the company.

Or as it was said in the 70's, "nobody ever got fired for buying IBM".
Unfortunately (for them) these guys work for Musk, so this saying doesn't apply.

Offline JH

  • Full Member
  • ***
  • Posts: 335
  • Liked: 206
  • Likes Given: 65
that's my question too. why waste two more launches on a handful of test sats when you can send up ~25 per launch. the first sats will only stay in orbit for a short time and there will be so many of them, if the first 50 perform poorly or die early, no biggie. That's especially true if they count toward the total sat count for the license.

Well, they probably would have placed them as secondary payloads on launches that other people paid for, like they did with Tintin A & B.
« Last Edit: 11/07/2018 01:37 pm by JH »

Offline Herb Schaltegger

that's my question too. why waste two more launches on a handful of test sats when you can send up ~25 per launch. the first sats will only stay in orbit for a short time and there will be so many of them, if the first 50 perform poorly or die early, no biggie. That's especially true if they count toward the total sat count for the license.

Because test sats are, by definition, "test" vehicles. The FCC oversees and regulates satellite telemetry/command and payload radio-frequency emitting operations for testing and R&D differently than they do for operational/indefinite use vehicles and transmitters, especially when there's even the slightest risk that the test radio frequency emissions might interfere with other (terrestrial or space) licensees and users of the same or neighboring frequencies.
Ad astra per aspirin ...

Offline Nomadd

  • Senior Member
  • *****
  • Posts: 4033
  • Boca Chica, Texas
  • Liked: 13351
  • Likes Given: 433
 I've had experience with S, Ku band and Iridium. All were complete child's play compared to the system SpaceX is planning. The antennas and laser links and such will be impressive feats, but I'd bet dollars to donuts that software will be the real issue.
 Point of ramble being, I doubt if Elon means to launch a fully functional system all at once, but hardware that can be software upgraded to meet most of the goals. Antenna hardware will surely evolve, but the really fancy spot beam stuff will be more important when the customer base starts building.
Those who danced were thought to be quite insane by those who couldn't hear the music.

Tags: