Atlas Chronology

[WallE]

Atlas 114D/Agena 2203 (code name "Ocean Way") launched from PALC 1-1 on the morning of December 22, 1961 on the fifth SAMOS attempt, with so far only SAMOS 2 last January having achieved limited success. The launch went nominally until ground guidance sent the programmed SECO discrete at T+265 seconds, but the programmer did not act upon it. The sustainer engine gradually lost thrust due to LOX depletion and the backup SECO cutoff command was not acted upon either. The programmed signal to disable the Agena ISDS, which normally occurred at SECO, was accomplished by a backup signal at VECO. The satellite vehicle was thus boosted into an overly high 350x125 mile orbit that prevented proper deorbit and recovery of the film capsule, which was believed lost somewhere in Canada. As with SAMOS 4, the Air Force did not say anything in their official statement about SAMOS 5 other than that a satellite had been launched on an Atlas-Agena vehicle, although unlike with 4 the announcement did include the remark that it reached orbit.

The programmer was thought to have an open circuit somewhere that caused the non-response to the SECO discrete, but all other programmer functions were accomplished properly. On subsequent Atlas vehicles a modification was done to provide an independent circuit path for engine cutoff in the event of an open circuit.

Other minor problems on the launch included high thrust section temperatures during booster phase, the cause of which was unknown. The B-2 nacelle door failed to properly close at liftoff and remained so to the limits of camera coverage. The nacelle door remaining open probably did not contribute to the elevated thrust section temperatures, but it could not be entirely ruled out and subsequent flights would have increased temperature measurements in them. There was pronounced in-flight vibration during the last 49 seconds of booster phase which had not occurred on any previous Atlas-Agena launch. Although no detrimental effects resulted, this was corrected by moving the gyro package to a different location on subsequent vehicles.

Side note: At the top of the umbilical tower on PALC 1-1 is a camera designed to get a close-up view of the Agena at liftoff. This camera was not present for the first two SAMOS launches as photos of those show and seems to have been added at some point during the spring or summer of 1961 and PALC 1-2 also sported an umbilical tower camera. It was more likely than not added as a result of the postflight findings from SAMOS 1 so they could get a better look at the Agena umbilicals at liftoff.

[WallE]

This wwas part of a doc explaining modifications to the Atlas for MIDAS/SAMOS, which were relatively few--the vehicle was mostly a stock D-series Atlas with increased LOX tank thickness to support the weight of the Agena, and as with all SLV Atlases only one telemetry package with a relatively small number of booster operating parameters monitored (as well as the range safety system being modified to also destruct the Agena if activated). I would assume NASA Atlas-Agenas were similar aside from having the Mod III-G guidance system used for AMR launches instead of the PMR Mod II-A. The standardized SLV-3 core started being phased in during 1963 but Atlas-Centaur did not start using it until AC-13 in '67.

The part shown here explains in detail how GD/A postflight reports were issued and that they were "stepped", starting with a quick flash report within a few hours hours of the launch and progressing through more detailed ones until the final flight evaluation report issued two weeks after the launch. Lockheed of course would have issued their own entirely separate postflight reports for Agena performance.

Of the flash reports we've seen quite a few of those, there's a large collection on Jonathan's Space Report for VAFB Atlas launches (as well as Thor and Scout launches) but no Cape Canaveral ones and they're one page sheets detailing the basic facts of the launch. There are also a number of 7 day reports on DTIC but only for Cape Canaveral launches and the only two week full flight evaluation report I'd seen was for Atlas 48D. NASA also began drafting their own launch reports in 1964 and NTRS has various of those, which appear to contain very similar data to the flight evaluation reports.

[Jim]

NASA also began drafting their own launch reports in 1964 and NTRS has various of those, which appear to contain very similar data to the flight evaluation reports.

When Goddard to over launch management.

[Jim]

[quote author=WallE link=topic=46681.msg2713665#msg2713665 date=1756572993

Of the 24 hour reports we've seen quite a few of those, there's a large collection online for VAFB Atlas launches but no Cape Canaveral ones. There are also a number of 7 day reports on DTIC but all for Cape Canaveral launches and the only full flight evaluation report I'd seen was for Atlas 48D.
[/quote]

There are Flight Test Reports for the ICBMs.

[WallE]

Those are I think all the Flight Test Working Group reports on DTIC mostly for Atlas D/E/F test flights and also the reports for MA-1 and MA-5. Those would be the 7 day reports mentioned above. Of the full 2 week flight evaluation reports however the only one online is for 48D. Those included among other things photographs and telemetry data charts and the later NASA reports also include most of that data (the NASA report for AC-5 definitely did).

[WallE]

re: Atlas-Agena vehicles. Officially this was known as the LV-3 and as I said was mostly a standard D-series Atlas with thicker LOX tank skin and the Mod III-G/II-A guidance system which was a transistorized version of the guidance system used on Atlas D missiles.

The SLV-3 core would eventually replace this for all SLV Atlases. This started out as the Atlas-Centaur core and it had the enhanced MA-5 engines and thicker LOX and fuel tank skin. SLV-3 vehicles also got rid of the retrorockets at the top of the equipment pods that were used to back the sustainer section away from the upper stage/payload after SECO. An important distinction is that early Atlas-Centaur vehicles used the LV-3C designation and still had the retrorockets as well as the xxxD serial number system instead of the SLV-3's numeric only serial numbers. Beginning with AC-13 they were renamed SLV-3C and got rid of the retrorockets.

The first "proper" SLV-3 with no retrorockets was used on the tenth GAMBIT launch in 1964. NASA first used SLV-3 vehicles for GATV but other NASA Atlas-Agena launches used the older LV-3 core through '65.

It gets more confusing because MA-5 was used as a blanket designation for the engines on space launcher Atlases, even when it wasn't. Mercury-Atlas vehicles were referred to as using MA-5 engines although they just used the MA-2 engines (but with vernier solo mode deleted) and the same applied to LV-3 Atlas-Agenas.

Other than that, the SLV-3 core on Atlas-Agena vehicles differed from Centaur ones mainly in having a tapered forward end and the Mod III-G radio guidance system instead of Centaur's inertial guidance.

[Jim]

re: Atlas-Agena vehicles. Officially this was known as the LV-3 and as I said was mostly a standard D-series Atlas with thicker LOX tank skin and the Mod III-G/II-A guidance system which was a transistorized version of the guidance system used on Atlas D missiles.

Agena LV-3A
Mercury LV-3B
Centaur LV-3C

[Jim]

It gets more confusing because MA-5 was used as a blanket designation for the engines on space launcher Atlases, even when it wasn't. Mercury-Atlas vehicles were referred to as using MA-5 engines although they just used the MA-2 engines (but with vernier solo mode deleted) and the same applied to LV-3 Atlas-Agenas.

Uprated MA-2
Atlas D-Agena A (LV-3A)
Atlas D-Agena B (LV-3A)

MA-5
Atlas D-Agena D (LV-3A)
Atlas D-Mercury (LV-3B)
Atlas D-Centaur (LV-3C)

[WallE]

1:15

Could he be referring to CANYON 4 there? If he started in '64 then surely he can't be talking about MA-3 when he mentions an Atlas that went straight up and didn't pitch over.

The Peter Hunter photo collection had had a document listing all Atlas flight failures and the cause of them, but I suspect now that there was a mistake in there and they got the ones for 95F and CANYON 4 mixed up. It listed 95F as failing to execute its pitch and roll program, but CANYON 4 as "Sustainer Mod 3. turbine failure." Seems the latter was actually the one that didn't pitch over and kept going straight up and 95F had the sustainer gas generator failure.

From the video it seems a tech made an ill-conceived adjustment that apparently did something to the booster engine acutators and prevented them from being able to gimbal the engines. In that case the result would be no pitchover maneuver. If so then it would also explain why NASA launched AC-26 with no serious delay, if it was established that the failure on CANYON 4 was just a tech's careless mistake and there was nothing to be concerned about.

[WallE]

Atlas-Centaur AC-3 was the third Centaur R&D test flight, launched June 30, 1964. Atlas 135D was erected on LC-36A January 27 and Centaur 1C February 20. An initial launch attempt June 26 was called off due to guidance system problems. At 9:00 AM on the morning of June 30, AC-3 was launched. Booster performance was almost entirely normal throughout Atlas phase except for a malfunction of the propellant utilization system--for the first 70 seconds of the launch the sustainer engine ran slightly LOX rich. At 70 seconds the PU system abruptly returned the propellant mix to normal. Sustainer thrust however remained at essentially normal levels throughout powered flight. An electrical short was believed to have caused the PU malfunction and had it continued an early SECO from LOX depletion could have resulted. BECO occurred T+147 seconds and SECO T+226 seconds. The Azusa tracking system ceased operating about 30 seconds into the launch; this was due to the addition of an RF shield between the tracking beacon and its power supply which apparently contacted a terminal and caused a short.

The Centaur insulation panels were jettisoned properly at T+177 seconds, thus accomplishing one of the main goals of the launch and vindicating the hard-won effort over the last two years to design a workable separation mechanism. A detachable payload fairing was carried for the first time and it was jettisoned successfully at T+202 seconds. The fairing separation aroused some concern due to momentary high vibration that affected the guidance system. Separation of the Centaur from the Atlas was also successfully accomplished and did not repeat an incident that occurred during staging on AC-2 when Centaur attitude control was lost for a few moments due to an improper umbilical disconnect. Engine start began T+242 seconds.

Five seconds into Centaur burn, hydraulic pressure for the C-2 engine abruptly fell to zero and gimbaling control was lost. This left the C-1 engine to perform attitude control. Pitch and yaw stability were maintained but roll control was lost. The Centaur developed an initially relatively low roll rate of 15 degrees per second but this gradually increased to as much as 40 degrees per second. A programmed pitch-up command at 432 seconds removed any further ability of the C-1 to control the roll rate and it reached 310 degrees per second by 496 seconds. The centrifugal force of the roll forced the LOX in the main tank away from the pump inlets and resulted in engine shutdown from LOX starvation. The programmed MECO discrete at 619 seconds closed the propellant valves and halted residual thrust from venting gases. Attempts at test-operating the boost pumps were mostly unsuccessful due to vehicle tumbling. Operation of the small attitude control engines was successful in arresting the roll rate. Impact of the Centaur in the South Atlantic occurred 12 minutes after liftoff.

The hydraulic system failure was believed due to a break in the connection between the turbopump gearbox and the hydraulic pump. Despite the failure of the Centaur to attain orbit, the launch was considered relatively successful and several important goals were accomplished, in particular verifying the operability of the insulation panel jettison system.

[Jim]

NASA also used mostly Agena B boosters through 1965 but the Mars '64 probes flew on Agena D as the mission requirements needed enhanced performance.

no, the B was not available for the Mars '64 contract

They were also forced to launch Mariner 4 from Pad 12 due to scheduling conflicts although that was normally set up for Agena B launches.

No, two pads were required to launch two probes in the short launch period

Wonder if there were any issues with the flame bucket on Pad 12 not being equipped for the more powerful MA-5 engine thrust?

it was equipped. Many other MA-5 launches occurred from there.

[WallE]

Atlas 7112 ("White Pine") launched from PALC 1-2 on July 12, 1965 to launch the 20th GAMBIT satellite (the Atlas S/N and the launch date are an amusing coincidence). The launch went entirely according to plan until BECO at T+131 seconds at which point the programmer sent simultaneous BECO and SECO commands and advanced its timer to T+400 seconds, terminating thrust. This triggered a premature Agena separation and engine start. Unfortunately the altitude and velocity were too low for orbital insertion to be possible. The Agena burned all the way to propellant depletion but it was hopeless and the stage and satellite vehicle impacted in the Pacific Ocean northwest of Samoa, about 4,700 miles downrange. Postflight investigation believed that vibration caused by BECO had triggered a switch in the programmer that generated an erroneous SECO. Several components were redesigned afterward and dipped conformal coated PCBs would be abandoned.

The transistorized "square" Atlas programmer was a tough beast to tame especially considering the old electromechanical "round" autopilot had been super reliable and as we see here was still causing problems as late as 1965.

[WallE]

If poor  aerodynamics were the suspected cause of the MA-1 failure, then why did they Big Joe capsule, essentially the same shape as the MA-1 capsule (though lighter) make it through Max-Q? I understand they reshaped the trajectory after MA-1 but weren't the aerodynamic forces on the Big Joe shot essentially the same as the ones on the MA-1 shot. I don't understand how weight would affect the aerodynamics.  I understand the shape of the Mercury Capsule was different than the "standard" Atlas warhead, but I'm assuming that the weight of the Mercury capsule was LESS than the Atlas warhead, given the velocity requirements for an orbital vs a ballistic shot.

I found out the answer. Turns out that in fact Big Joe (and its backup vehicle Atlas 20D) had both thicker fuel and LOX tank skin and then on MA-1 they inexplicably went back to using the thinner Atlas ICBM LOX tank skin. In addition, 67D, which launched MA-2, and 77D, which would have launched MA-3, were built this way. They put a reinforcing band on 67D and 77D was recalled and not flown.

[WallE]

Guess there were no changes to the LV-3A propulsion system after all and it was just the standard Atlas D ICBM engines. This also describes vehicles 29D, 45D, 75D, and 84D as assigned for the MIDAS program. Obviously MIDAS did get moved to VAFB after the second launch and most likely because Pad 14 was completely turned over to the Mercury program. 75D did end up launching MIDAS 9 long afterwards in July '63 and it seems a bit strange that the vehicle sat in storage for so long while MIDAS 3-8 used higher numbered boosters built later. 84D ended up not a space booster at all but an on-duty missile at Offutt AFB and launched September '63 in a very late Atlas operational test from VAFB.

Second MIDAS 3 launch attempt: An ominous sign of things to come.

This was the first Atlas-Agena B launch and booster performance was normal despite the satellite failing not long after reaching orbit. Although not mentioned here, the programmer reset itself after BECO which had no real effect on anything but NASA asked the Air Force to look into it just in case. It was apparently an open circuit somewhere and judged a random Q/C fault that the more tightly controlled Mercury program should not have to worry about.

NASA were also a little unhappy that it had taken so long to fly this LV combination as they assumed the Air Force would fly it first and obtain data for them on its performance. Instead, Ranger 1 went up a few weeks later with just one previous Atlas-Agena B launch--the Ranger program had ended up progressing quickly while MIDAS/SAMOS were beset with endless delays which was why it took so long.

[WallE]

Atlas 12D (Tall Fellow) was the first West Coast Atlas launch. It lifted from Pad A-2 on the morning of September 9, 1959, six hours after Big Joe's launch at Cape Canaveral. The flight was an essentially flawless one and impact of the blunt-nosed R/V in the South Pacific occurred nearly on target. However, T/M data was lost from T+112 to 194 seconds due to an apparent short in a wiring harness, preventing evaluation of missile performance during that portion of the flight. The Atlas 9C disaster on September 24 notwithstanding, the second half of 1959 and the first two months of 1960 saw an overall excellent run of Atlas flights with only minor hardware glitches. One other launch occurred in September '59, 17D which launched from Pad 13 at the Cape on the 17th. This one was also trouble free except for another recurrence of telemetry issues during portions of the flight, in this case due to a power supply problem.

[Blackstar]

It would be great if we had a source for all these assertions about why launches failed.

[Jim]

It would be great if we had a source for all these assertions about why launches failed.

Like the like full documents of  accident report or flight test report attached to each post vs clips?

[Blackstar]

It would be great if we had a source for all these assertions about why launches failed.

Like the like full documents of  accident report or flight test report attached to each post vs clips?

Anything. Anything at all.

[WallE]

Oh, this was from the Atlas instrumentation difficulties doc which you can get from DTIC.

[WallE]

Atlas-Able 2 was launched on Atlas 80D from Pad 12 on September 25, 1960, a year and a day since the 9C disaster demolished the pad and put it out of use for eight months. Aside from restoring Pad 12, now with only a small umbilical mast instead of the large tower that got knocked over on 9/24/59, it had taken a long time before another Atlas vehicle was available to fly. 80D was sixty vehicle #s ahead of 20D which had launched the first Able attempt the previous November and unlike 20D, a recycled Mercury booster, it was a mostly stock D-series Atlas with vernier solo mode and only a few small changes to support the use of upper stages. The lunar probe was mostly similar to the Able 1 probe.

The Atlas began its journey from the GD/A plant in San Diego to Florida August 12, the same day that Pad 12 hosted the mostly successful R&D flight of Atlas 66D. It was erected on the pad September 2. The probe arrived September 6. The approach of Hurricane Donna on the 10th necessitated removing the booster from Pad 12 and Missiles 76D, 79D, and 3E on adjacent pads. The Atlas was re-stacked on Pad 12 and prelaunch preparations continued. The probe was heat sterilized on the 22nd to prevent possible contamination of the Moon with microbes. There was no PFRF for this vehicle like the one that led to the loss of 9C; PFRFs had mostly been discontinued on D-series vehicles for a year now.

At 10:13 AM on the morning of the 25th, the Atlas was launched. Engine start and vehicle rise-off were normal and it began steering downrange. The booster was visible until disappearing behind clouds at 90 seconds. BECO occurred at 134 seconds, payload fairing jettison at 170 seconds, and SECO at 271 seconds. At about 14 minutes after liftoff, all tracking and telemetry data was lost. Something had gone very wrong. An initial analysis of telemetry found that the malfunction had occurred entirely in the Able second stage, apparently the AJ10 engine and interstage section. Atlas performance during powered flight had been normal and uneventful, except that the booster failed to act on the programmed VECO discrete; vernier solo mode was only planned for five seconds on this flight but ended up burning until propellant depletion nearly 20 seconds later. This added 100 extra miles of trajectory, but that alone was not enough to be detrimental to the flight.

Due to the late VECO, staging was initiated by a backup signal from an oxidizer valve switch in the second stage, which was set to activate when the valve was 80% open. The intention was that the Able engine would ignite and separation from the Atlas occur when it reached 60% thrust level. The late VECO caused the engine to be at 100% thrust when staging occurred. This caused higher than nominal pressure in the interstage section. As soon as the separation signal was received, the AJ10 engine began to experience thrust decay and pitch control was lost due to an apparent broken wire in the potentiometer. Oxidizer began to leak due to possible damage to the thrust chamber. The combination of the oxidizer leak and loss of pitch control caused vehicle tumbling. Thrust was at essentially zero by 352 seconds. The manual fuel cutoff command was received from the ground at 379 seconds and the solid third stage ignited at 382 seconds and completed its 40 second burn as planned. However, the third stage and probe were now pointed downward and back towards Earth. Final loss of signal and telemetry occurred at 14 minutes.

The Atlas's non-response to the VECO discrete was believed due to a short in the engine relay box but the precise cause of the propulsion system failure in the Able was not determined. The oxidizer leak would have caused the thrust decay, but it was unclear why the gimbaling failed or an oxidizer leak developed and the unintended 100% thrust at separation was not thought to be in of itself enough to damage the vehicle.

Complete flight report can be found on the STL archives on sdfo.org