Author Topic: Was N1's failure due to the number of engines? (seeking clarification)  (Read 11542 times)

Offline Pipcard

  • Full Member
  • ****
  • Posts: 433
  • Liked: 94
  • Likes Given: 92
With the recent reveal of SpaceX's plans for a 42-engine heavy-lift rocket, some people are worried that this would be N1 all over again. But what actually caused its failure? Did it have to do with having 30 engines and complex plumbing? Or was it about not being able to do proper test fires?
« Last Edit: 10/12/2016 05:50 PM by Pipcard »

Offline Kansan52

  • Full Member
  • ****
  • Posts: 874
  • Hutchinson, KS
  • Liked: 263
  • Likes Given: 311
You'll likely receive posts that each of the items you mention are suspect.

The failure of the same engines on the test stand here and on Orbital's Antares points to the engines (IMHO).

Offline teetlebomb

  • Member
  • Posts: 30
  • Liked: 28
  • Likes Given: 69
Knowing more about the N1 failures is something I would like to know.

In my perspective (as a systems (although not rocketry) programmer), the failure came down to improper or insufficient testing in the KORD control system. From what I understand, KORD could not react properly (or in-time) to handle all the situations that were input to it by the high number of engines.

With modern day rocket control technologies, I don't think controlling 42 engines will be a major problem.

Offline edkyle99

  • Expert
  • Senior Member
  • *****
  • Posts: 12860
    • Space Launch Report
  • Liked: 3610
  • Likes Given: 617
In my mind, the first key problem was that none of the first stages were ever test fired on a test stand!  The Soviets chose to skip that step and it cost them. 

The second key problem was the reliability of the engines, but that wasn't as big a problem as the system engineering problems that caused most of the failures.  Those problems might have been solved during an extensive series of test stand firings.

 - Ed Kyle
« Last Edit: 10/12/2016 06:31 PM by edkyle99 »

Offline Space Ghost 1962

  • Senior Member
  • *****
  • Posts: 2485
  • Whatcha gonna do when the Ghost zaps you?
  • Liked: 2385
  • Likes Given: 1746
N1 flew with NK-15/15V's not NK33/43.  N1-8 had them but was unflown/scrapped. See Soviet Moon rocket plunges to the ground after a minute in flight for the "blow by blow" on this.

NK33 flew as reworked with AJ gimbals as AJ26 flew 4 engines on two flights successfully. Twice with a four nozzled steering engine (NK-33 was fixed) on single engined Soyuz 2.1v.

There have been at least two test stand failures prior to flights. One did extensive damage to the test stand.

It is hard to examine the NK33 results due to the gap in time and also the nature of how they were stored and "re-qualified" for flight - these and the elapsed time (many decades) may have played a significant role in all failures.

Irrespective of design, consolidation of manufacturing decision led to not restarting production, thus discontinuing NK-33 use. So we'll never know.
 
As to the OP, suggest that the question can't be answered as stated, because too much of the program had so many flaws, that attempting to answer it might do a disservice.

A better answer is to examine Falcon 9 (and perhaps eventually Falcon Heavy) because there is more flight history, it is a better run program that has more accountability and transparency, not to mention is also current with current technology.

Know that's not the answer you want, but that's what you have.

IMHO N1 could have been successful given other conditions than it experienced. Saturn V program had vastly more advantages as a program to start with, which allowed it less program risk, which meant that it arrived ahead of schedule with a better measure of confidence. N1 program's lack of those advantages meant that the program risk exceeded it's designers/builders/operators capabilities to "bring in the program" fast enough to garner a success. Which was a "common mode" for other like programs between both.

As an eye to ITS, its "how" you execute on a program as much as the design that matters.

In my mind, the first key problem was that none of the first stages were ever test fired on a test stand!  The Soviets chose to skip that step and it cost them.

Nor did they, like Antares and unlike Saturn S1C, test fire the entire stage independently.

Quote
The second key problem was the reliability of the engines, but that wasn't as big a problem as the system engineering problems that caused most of the failures.  Those problems might have been solved during an extensive series of test stand firings.
The general thought was that they'd accept the fact that the first ones WOULD BLOW UP, and they'd be smart enough to figure out why and fix before they exhausted the program. And that the "speed"/"cost" advantage they'd gain by NOT doing the testing/etc would mean they'd outmaneuver rivals, as sometimes (but not this time) they did.

Which is why it's a BAD IDEA(TM) to use N1 as a reference case for "too many engines" ...

Offline Space Ghost 1962

  • Senior Member
  • *****
  • Posts: 2485
  • Whatcha gonna do when the Ghost zaps you?
  • Liked: 2385
  • Likes Given: 1746
Knowing more about the N1 failures is something I would like to know.
See above, read all the flight history.

Quote
In my perspective (as a systems (although not rocketry) programmer), the failure came down to improper or insufficient testing in the KORD control system. From what I understand, KORD could not react properly (or in-time) to handle all the situations that were input to it by the high number of engines.
Not the root cause.

Quote
With modern day rocket control technologies, I don't think controlling 42 engines will be a major problem.

Unproven either way with N1. Nor F9. However, N < 12 is adequately proved, so the need for a single engine per stage (as with Ariane 5/6, Atlas V, DIV, others) is no longer the "gold standard".

The alternative of huge, lower pressure, less efficient combustion chamber engines (as with F1) however is not in doubt. They are obsolete and not coming back.

Offline Hobbes-22

  • Full Member
  • ***
  • Posts: 353
  • Acme Engineering
    • Acme Engineering shop
  • Liked: 101
  • Likes Given: 74
The general thought was that they'd accept the fact that the first ones WOULD BLOW UP, and they'd be smart enough to figure out why and fix before they exhausted the program. And that the "speed"/"cost" advantage they'd gain by NOT doing the testing/etc would mean they'd outmaneuver rivals, as sometimes (but not this time) they did.

Which is why it's a BAD IDEA(TM) to use N1 as a reference case for "too many engines" ...

They were counting on having ~14 test flights before the first manned mission, and losing a large fraction of them. Cost was a major factor, ISTR they skipped the first stage test stand because it'd tie up too much of the available concrete production capacity at the time (but I haven't been able to find a reference for that anywhere).

They ran several hundred engine tests, so there was little doubt about the individual engines.

Offline Michel Van

  • Full Member
  • *
  • Posts: 113
  • Liege, Belgium
  • Liked: 32
  • Likes Given: 70
There Were several reason were N1 failed

1. Sergei Korolev who envision the N1 died during the program in 1966.

2. his successor Vasily Mishin had to make modifications to N1 to increase it payload
installment of 6 additional NK-15 engines in center, increase complexity of feed lines in rocket.
Super cooling of Liquid Oxygen and Kerosine to get more fuel into tanks
i have the suspicion that Mishin not change the steel of feed lines for Liquid oxygen
and that was to brittle for super cool Liquid oxygen


3. the installed NK-15 were not tested, just one of two from production taken and fired
the rest were build into rocket, after N1 destroy it's launch pad
they start to check all NK-15 and find allot bad welding and metal shaving inside the engines
A turbopump sucking in metal shaving can rupture and explode

4. Flight computer KORD was junk
it never work properly on first launches, as they discover that engines produce a vibration
in same frequency as KORD clock frequency !
later version of KORD work on another clock frequency and covert in asbestos protection against fire.

5. a Shoe string Budget
in contrast to NASA, the MoM had Not vast money reserve for Lunar program
so there were no Test Stand for stages of N1 rocket
while NASA tested there Saturn V stages bevor assembly at cape,
hope Vasily Mishin to do 12 to 14 test flight until N1 was ready

6. forget Dallas or Dynasty, the real drama is in Soviet Moon program
the Soviet Space Program is full of Intrigue worthy for a play by  William Shakespeare.
instead to combine there best effort, the major Chief Designers  fight each other in
Chelomei proposed UR-500 and UR-700 rocket with Lk-1 and LK-700 lander
Valentin Glushko refused to build the N1 rocket engines, if not use NTO/UMHD fuel
so the Kuznetsov Design Bureau inexperienced in Rocket engine had to build them.
as in end Valentin Glushko replaced Mishin, instead to get last bugs out the N1.
He order end of program and total destruction of N1 hardware...

and start new program called "Vulcan" what became Energia/Buran in 1976
« Last Edit: 10/12/2016 08:15 PM by Michel Van »

Offline whitelancer64

With the recent reveal of SpaceX's plans for a 42-engine heavy-lift rocket, some people are worried that this would be N1 all over again. But what actually caused its failure? Did it have to do with having 30 engines and complex plumbing? Or was it about not being able to do proper test fires?

Wikipedia has an excellent N-1 entry, which includes a section with the 4-flight history. I suggest that be read, but to summarize:

Flight 1 - Very early in flight, the rocket's control system shut down two engines in error due to a short circuit. The shutdowns caused POGO oscillations which ruptured a LOX line. Further oscillations ruptured an RP-1 line some 20 seconds later, which resulted in a fire. The fire burned through the wires of the flight control system, which caused it to shut down the entire rocket, including the 2nd and 3rd stages, at 68 seconds into the flight. The rocket subsequently fell to the ground. On the plus side, the launch escape system on the boilerplate spacecraft on top worked perfectly.

Basically the cause of failure was bad wiring of instrumentation and poor programming of the rocket's computers that ran the control system. POGO oscillations and vibrations contributed to the failure as well.

Flight 2 - The most well known failure, a few seconds after liftoff all of the rocket's engines - except for one - shut down and it crashed back on the launch site, exploding and destroying the launch pad. On the plus side, once again, the launch escape system worked perfectly.

Just before liftoff a turbopump on one of the engines exploded, severing nearby fuel lines and starting a fire. The control system shut down two engines, but engineers could not determine why all the other engines were shut down, nor why one continued to operate after that point. Initial causes of the turbopump failure was speculated to be injestion of debris, and subsequently the rocket had fuel and oxidizer filters installed.

Flight 3 - An roll began shortly after liftoff, which quickly exceeded the rocket's ability to control. After the 2nd flight destroyed the launch pad, an interlock was installed to prevent the computer from shutting down the rocket until 50 seconds into flight, so the rocket continued to fly and the roll worsened until structural failure occurred, the first stage twisted apart at the interstage structure and separated from the rest of the rocket, both parts remained intact until impact with the ground. This rocket had a boilerplate spacecraft with no launch escape system.

The roll was later determined to be caused by aerodynamic forces from the rocket exhaust swirling around the base of the rocket.

Flight 4 - This rocket had a much more powerful stabilization system with dedicated roll control rockets and a beefed up flight control system with 13,000 sensors (the previous flight control computers only had 700 sensors). The flight went well until a programmed shut down of 6 engines to reduce aerodynamic loads on the rocket. This induced a shock wave that ruptured fuel and oxidizer lines, starting a fire. The first stage exploded 107 seconds into the flight, which caused the launch escape system to pull the prototype lunar spacecraft away from the rocket, which once again worked perfectly.

Could any of these doom another rocket? Yes, but we've learned a lot about how to fly rockets since then. All of these failure scenarios happened to the early Atlas, Titan, et al. rockets as well.

TL;DR?

Flight 1 - bad flight control computer wiring / programming and POGO - bad flight controls are most likely not a problem for SpaceX and POGO is very well understood now, and is relatively simple to design out of a system.
Flight 2 - turbopump explodes, possibly from injesting debris - Elon Musk has been famously paranoid about this, the Merlins were designed to handle small debris and they do have fuel filters.
Flight 3 - unexpected aerodynamic interactions - much less likely today, we've got fabulous computer simulations for aerodynamics that would help us avoid this.
Flight 4 - unexpected shock waves from engine shut downs - again, something SpaceX most likely has a handle on from previous flight experience.

As an addendum, none of the N-1 failures were because the rocket had too many engines, it had numerous other flaws that would have caused it to fail with fewer engines. Flight 4 came close to succeeding and the redesigned N1-L3 would likely have been successful. Sure, there are more fuel feed lines and turbopumps that could cause an issue, but good design work and computer simulations can help avoid those now.

I think the final take-away is this: no matter what, always have a launch escape system on your rocket.
"One bit of advice: it is important to view knowledge as sort of a semantic tree -- make sure you understand the fundamental principles, ie the trunk and big branches, before you get into the leaves/details or there is nothing for them to hang on to." - Elon Musk
"There are lies, damned lies, and launch schedules." - Larry J

Offline Toast

Flight 1 - bad flight control computer wiring / programming and POGO - bad flight controls are most likely not a problem for SpaceX and POGO is very well understood now, and is relatively simple to design out of a system.
Flight 2 - turbopump explodes, possibly from injesting debris - Elon Musk has been famously paranoid about this, the Merlins were designed to handle small debris and they do have fuel filters.
Flight 3 - unexpected aerodynamic interactions - much less likely today, we've got fabulous computer simulations for aerodynamics that would help us avoid this.
Flight 4 - unexpected shock waves from engine shut downs - again, something SpaceX most likely has a handle on from previous flight experience.

Going back to what edkyle said:

And to
In my mind, the first key problem was that none of the first stages were ever test fired on a test stand!  The Soviets chose to skip that step and it cost them. 

The third failure was the only one that likely could not have been caught by more extensive testing of the rocket after assembly. The component testing was flawed, but there was pretty much no integration testing at all.

Offline whitelancer64


Going back to what edkyle said:

And to
In my mind, the first key problem was that none of the first stages were ever test fired on a test stand!  The Soviets chose to skip that step and it cost them. 

The third failure was the only one that likely could not have been caught by more extensive testing of the rocket after assembly. The component testing was flawed, but there was pretty much no integration testing at all.

Very true. One of the reasons we won the moon race was extensive pre-flight testing.

Though it is worth noting that Apollo 4 and 6 were all-up flight tests. Apollo 6 suffered from severe POGO oscillations and the 2nd stage controller was cross-wired such that it shut down a good engine when another one failed, nearly caused the failure of the entire flight.
"One bit of advice: it is important to view knowledge as sort of a semantic tree -- make sure you understand the fundamental principles, ie the trunk and big branches, before you get into the leaves/details or there is nothing for them to hang on to." - Elon Musk
"There are lies, damned lies, and launch schedules." - Larry J

Offline savuporo

  • Senior Member
  • *****
  • Posts: 5155
  • Liked: 975
  • Likes Given: 345
Unproven either way with N1. Nor F9. However, N < 12 is adequately proved, so the need for a single engine per stage (as with Ariane 5/6, Atlas V, DIV, others) is no longer the "gold standard".
R-7 and Soyuz-U specifically is the single most launched carrier rocket ever, and kind of defines the gold standard with its reliability. Definitively not a single engine per stage.
Orion - the first and only manned not-too-deep-space craft

Offline Oli

  • Senior Member
  • *****
  • Posts: 2124
  • Liked: 379
  • Likes Given: 55
Unproven either way with N1. Nor F9. However, N < 12 is adequately proved, so the need for a single engine per stage (as with Ariane 5/6, Atlas V, DIV, others) is no longer the "gold standard".
R-7 and Soyuz-U specifically is the single most launched carrier rocket ever, and kind of defines the gold standard with its reliability. Definitively not a single engine per stage.

Definitively a single engine per stage. Unless you count the 4 boosters as one stage.
« Last Edit: 10/13/2016 06:40 AM by Oli »

Offline savuporo

  • Senior Member
  • *****
  • Posts: 5155
  • Liked: 975
  • Likes Given: 345
Unproven either way with N1. Nor F9. However, N < 12 is adequately proved, so the need for a single engine per stage (as with Ariane 5/6, Atlas V, DIV, others) is no longer the "gold standard".
R-7 and Soyuz-U specifically is the single most launched carrier rocket ever, and kind of defines the gold standard with its reliability. Definitively not a single engine per stage.

Definitively a single engine per stage. Unless you count the 4 boosters as one stage.
Uh, Soyuz has 32 liquid engine thrust chambers going off at launch, lit by a giant matches made out of birch wood. It's not single engine by any stretch of imagination.
Orion - the first and only manned not-too-deep-space craft

Offline rocx

  • Full Member
  • ***
  • Posts: 384
  • NL
  • Liked: 264
  • Likes Given: 145
Unproven either way with N1. Nor F9. However, N < 12 is adequately proved, so the need for a single engine per stage (as with Ariane 5/6, Atlas V, DIV, others) is no longer the "gold standard".
R-7 and Soyuz-U specifically is the single most launched carrier rocket ever, and kind of defines the gold standard with its reliability. Definitively not a single engine per stage.

Definitively a single engine per stage. Unless you count the 4 boosters as one stage.
Uh, Soyuz has 32 liquid engine thrust chambers going off at launch, lit by a giant matches made out of birch wood. It's not single engine by any stretch of imagination.
Yet most rocket engineers qualify the RD-107 (or RD-108), which the core and each of the four boosters have one of, as one engine with four nozzles, because of its shared turbopump and other parts.
Any day with a rocket landing is a fantastic day.

Offline edkyle99

  • Expert
  • Senior Member
  • *****
  • Posts: 12860
    • Space Launch Report
  • Liked: 3610
  • Likes Given: 617
Uh, Soyuz has 32 liquid engine thrust chambers going off at launch, lit by a giant matches made out of birch wood. It's not single engine by any stretch of imagination.
It is five engines at liftoff, one in the center "second" stage and one each on the four booster "first" stages.  Each engine has one turbopump feeding six to eight thrust chambers.  (The turbopump is really the "engine" part of the machine.)  The design was clever, because use of multiple chambers allowed each chamber to be shaped for higher ISP while still combining to produce the needed thrust.

 - Ed Kyle
« Last Edit: 10/13/2016 03:16 PM by edkyle99 »

Offline ncb1397

  • Full Member
  • ****
  • Posts: 745
  • Liked: 324
  • Likes Given: 3
The general thought was that they'd accept the fact that the first ones WOULD BLOW UP, and they'd be smart enough to figure out why and fix before they exhausted the program. And that the "speed"/"cost" advantage they'd gain by NOT doing the testing/etc would mean they'd outmaneuver rivals, as sometimes (but not this time) they did.

Which is why it's a BAD IDEA(TM) to use N1 as a reference case for "too many engines" ...

SpaceX will be using the same N-1 philosophy of having to accept initial failures if Falcon is anything to go by. So far there has been 3 Falcon 1 launch failures, Falcon-9 Dev-1, a F9 pad failure, 5 Falcon 9 landing failures and CRS-7. Even assuming the booster or ship only costs ~$200 million each, we are talking a couple billion dollars for a similar number of failures.

Offline edkyle99

  • Expert
  • Senior Member
  • *****
  • Posts: 12860
    • Space Launch Report
  • Liked: 3610
  • Likes Given: 617
SpaceX will be using the same N-1 philosophy of having to accept initial failures if Falcon is anything to go by.
I don't see that myself.  Unlike the N-1 planners, SpaceX has spent a lot of money and effort ground testing its stages, both during the development period and continuing now with acceptance testing.  It has suffered launch and pad failures, but I think that its failure rate is generally consistent with most "brand new" launch vehicle families.  You can compare with Ariane 5, and H-2, both of which suffered frustrating failures in their early years.   Ariane 5 had four failures in its first 14 flights.  H-2(A) had three failures in its first 13 flights.  Falcon 9 has had three failures in its first 29 launch campaigns if you count AMOS 6 and CRS-1 as failures.  Only one, the CRS-1 launch, suffered a problem with the 9-engine first stage (CRS-1 made orbit, but the secondary payload was lost).

 - Ed Kyle
« Last Edit: 10/14/2016 03:48 AM by edkyle99 »

Offline georgegassaway

  • Full Member
  • *
  • Posts: 174
    • George's Rockets
  • Liked: 181
  • Likes Given: 65
While lack of static test firing of the stages was a critical problem, there was also a lack of dynamic (vibration) testing.  Here’s a Wiki link about the stand at MSFC used for the Saturn-V, and later also used by the shuttle (First time a shuttle was stacked, was Enterprise inside of this building).

https://en.wikipedia.org/wiki/Saturn_V_Dynamic_Test_Stand



Since some of the N1 failures were related to vibration issues, lack of dynamic testing was a major issue.  Can’t find out everything by just a static test firing, that would not simulate the vibrations that shook some of the N1 components apart.

IIRC, maybe it was in “Stages to Saturn”, where it was stated that the Russians budgeted their Lunar program (or perhaps it was the N1 specifically), at about 20% of the cost of Apollo (or Saturn-V).  As such, they cut so many things to the bone, that they effectively cut off some appendages too by not doing static firings  and not doing vibration testing.

Now, in the case of SpaceX, modern computer programming and perhaps some small scale dynamic testing (models) may have allowed them  to get by without doing dynamic testing of a fully stacked Falcon, or perhaps they did do full size vibration testing (I am unfamiliar with the dynamic vibration testing done with any launch vehicles of today).  Certainly they static test fire the “heck” out of the stages.

Also as stated, the “KORD” control system was really screwy. 

So as regards FH with 27 engines vs N1 with 30, Apples and Bricks comparison.  Statistically, FH should have no more odds of one out of 27 engines having a problem, compared to any three Falcon-9 launches.  At least as regards the total of 27 engines and the three boosters that each 9 are mounted in, not counting any potential unique FH issues.

 If they lose one engine on say the left booster (without a “bang” to cause a RUD), they can throttle down the 9 on the right side to about 91% throttle, albeit with some performance hit to the orbit.   But so far they’ve only had one Falcon-9 engine fail during a launch and that was a long time ago.
« Last Edit: 10/14/2016 04:58 AM by georgegassaway »

Offline Arch Admiral

  • Member
  • Posts: 98
  • 14th Naval District
  • Liked: 66
  • Likes Given: 0
The real problem with the NK-15 engines is that the actual flight engines could not be static tested. Many valves were opened and closed with pyros instead of hydraulic or pneumatic power. After a test run the valves were welded shut and could not be opened. I think this technique was developed for the GR-1 FOBS system, and retained for N-1 as a weight-saving device.

The test policy was to accept engines from the factory in lots of 6. 2 were sent to the Kuznetsov Bureau's test facility and tested, then scrapped. If these two completed one full burn, the other 4 were sent to Baikonur and installed in the stages without any testing. Thus it is not surprising that out of 120 flight engine runs, there were two turbopump explosions and one hot gas leak (the actual cause of the first flight failure). Every Saturn V engine had made three full-duration test burns, one of which was in the flight stages at Stennis.

The official causes of the 4 N-1 flight failures as given in Wikipedia are not accurate. The Soviets had the pernicious practice of having the Chief Designer of the suspected system lead the failure board. Of course Kuznetsov and his staff went to great lengths to exonerate his engines and blame everything on other designer's systems. These other designers pushed back and the final reports were political compromises. There was no evidence that debris in the propellant tanks caused any of the three engine failures. The most convincing account of the N-1 failures is in Vol. 4 of Boris Chertok's memoirs at the NASA History website.

The idea that N-1 could have been a success with the improved NK-33 engines is disproved by the terrible record of these engines in the Antares program. That failure board tore down many NK-33s in stock at Aerojet and found that about 1/3 had metal missing from the turbopump shaft. This was probably a machining error by the night shift at Kuznetsov's factory that was missed by quality control. (How was this missed by Aerojet when they rebuilt the engines????)

So N-1 failed because

A) there were too many engines
B) they were packed too close together
C) individually they had terrible quality control and nonexistent testing

plus many structural, aerodynamic, and electrical defects too numerous to mention.


Tags: