Forgot the Earth's rotation

[Jairo]

I heard about a Gemini mission where they made a wrong maneuver because they forgot to consider the Earth's rotation. Which Gemini was that?

A HB insinuates that this is a mistake too basic for a country that reached the Moon years later. I know this doesn't follow, for even recently the Beagle 2 was lost due to a unit conversion error. But I am missing any important context there? Was that error in the project since the beginning, or was it in a last minute in-orbit calculation?

[blackstar]

I think it was the Mars Climate Orbiter that suffered the unit conversion error, Beagle just slammed into Mars somewhere. And of course it goes without saying that even with Apollo behind them NASA still made spectacular and tragic errors with their manned flight program.
Also roughly at the same time as Gemini the Soviets, whom the HB's always cite as being technologically asuperior to the US, had some serious issues with their program, try this site for an overview:

www.strangehorizons.com/2004/20040503/shadows.shtml

[laurel]

It might have been Gemini 5.

"The crew had to use the re-entry thrusters to orient the spacecraft due to OAMS system failures. The retrofire and re-entry were conducted in darkness by the spacecraft computer. However the computer had been misprogrammed with an erroneous rotation rate of the earth (390 degrees per day instead of 360.98 degrees per day). Cooper's efforts compensated for what he recognized as an erroneous reading and brought the capsule down closer to the ship than they would otherwise have been."
www.astronautix.com/flights/gemini5.htm

[homobibiens]

Jul 28, 2009 15:34:44 GMT -4 Jairo said:

I heard about a Gemini mission where they made a wrong maneuver because they forgot to consider the Earth's rotation. Which Gemini was that?

I was just reading about this somewhere. Let me see if I can find it. The mistake was a one earth rotation per 24 hours vs. one earth rotation per 23 hours 56 minutes error. This resulted in the craft landing way of course; I guess the landing is the only part of the mission where this type of error would matter, unless the mission involved observing the earth.

Edited to add - oops, looks like I'm in second place on this one

[homobibiens]

Jul 28, 2009 17:17:05 GMT -4 laurel said:

It might have been Gemini 5.

Looks like it is. There is some information here under the near-fatalities section.

en.wikipedia.org/wiki/Space_accidents_and_incidents

[Ginnie]

Gordon Cooper was one of the best pilots in the world. On his Mercury flight, his navigational instruments gave out, the automatic control system died...
"Well, it looks like we've got a few little washouts here. I've lost all electrical power. Carbon dioxide levels are above maximum limits, and cabin and suit temperatures are climbing. Looks like we'll have to fly this thing ourselves. Other than that, things are fine, " he said.
...Faith Seven came out of the sky, rolling steadily, the Oklahoma farm boy flying with a precision that controllers mumbled was tighter than the autopilot or computers had ever delivered...*
He landed the capsule within a "stones throw" from his recovery ship.

*Italicised from "Live From Cape Canaveral" by Jay Barbree

[ajv]

Obviously the 390 on the astronautix.com page should be a 360.

But the summary doesn't quite agree with the details from the Mission Report (18M):

6.2.2.2.1

It was discovered postreentry that a part of the preretrofire update calculation was in error by approximately 7.89°. This quantity, which references the longitudinal (Xe) axis of the spacecraft coordinate system to Greenwich, was computed incorrectly in the real-time program at MCC-H. The net result of this error was that at retrofire the spacecraft computer was instructed that the spacecraft was at 187.44° west from Greenwich when, in fact, it was at 195.33° west. Thus, to the onboard computer, the spacecraft appeared to be overshooting the target, and the computer displayed corrections for the situation, resulting in a zero-lift indication and an actual undershoot. The MCC-H computer program was correct as written. An earth's rotation rate of 360.98° per day is used in the program which requires that the total elapsed time from G.m.t. midnight, prior to launch, be inserted to derive the spacecraft present position relative to the earth. The 7.89° error resulted from an omission of the elapsed number of days in the G.m.t. of retrofire term.

Now misprogrammed with an erroneous rotation rate of the earth is a quite a different description of the error than the mission report's omission of the elapsed number of days. But the effect is the same: 360° vs 360.98°.

However, I'm having difficulty getting the figures to add up. But it does work if we take the omission of the elapsed number of days literally - i.e. number of days excluding hours:minutes:seconds.

Launch occurred at 1965-08-21 14:00:00 GMT (rounded to the nearest second). Retro-fire occurred at 190:27:43 GET. So retro-fire occurred at 204:27:43 after midnight on 1965-08-21. i.e. 8 days + 12:27:43. The actual solar/siderial ratio (86400/86164) is 360.986°. And 8 * (360.986° - 360°) gives us the reported 7.888° figure.

I'm not very happy about the elapsed number of days being taken so literally but we don't actually know how the computer was programmed. It would certainly be possible for the program to separate the days calculation from the HH:MM:SS part.

[homobibiens]

Jul 28, 2009 20:40:29 GMT -4 ajv said:

Now misprogrammed with an erroneous rotation rate of the earth is a quite a different description of the error than the mission report's omission of the elapsed number of days. But the effect is the same: 360° vs 360.98°.

However, I'm having difficulty getting the figures to add up. But it does work if we take the omission of the elapsed number of days literally - i.e. number of days excluding hours:minutes:seconds.

Launch occurred at 1965-08-21 14:00:00 GMT (rounded to the nearest second). Retro-fire occurred at 190:27:43 GET. So retro-fire occurred at 204:27:43 after midnight on 1965-08-21. i.e. 8 days + 12:27:43. The actual solar/siderial ratio (86400/86164) is 360.986°. And 8 * (360.986° - 360°) gives us the reported 7.888° figure.

I'm not very happy about the elapsed number of days being taken so literally but we don't actually know how the computer was programmed. It would certainly be possible for the program to separate the days calculation from the HH:MM:SS part.

I'm not entirely sure I understand exactly how this works, but could it have been a data entry error, in which someone misspecified the dd part of dd:hh:mm:ss, which then may or may not have been converted to some other format? I don't know if this was a quantity that someone would have entered somewhere manually, or if it would have been read automatically from an on-board (or ground-based) system.

[scooter]

Jul 28, 2009 17:17:05 GMT -4 laurel said:

It might have been Gemini 5.

"The crew had to use the re-entry thrusters to orient the spacecraft due to OAMS system failures. The retrofire and re-entry were conducted in darkness by the spacecraft computer. However the computer had been misprogrammed with an erroneous rotation rate of the earth (390 degrees per day instead of 360.98 degrees per day). Cooper's efforts compensated for what he recognized as an erroneous reading and brought the capsule down closer to the ship than they would otherwise have been."
www.astronautix.com/flights/gemini5.htm

Wasn't that the one where Armstrong made the dubious claim of the most distant landing (from planned landing point) in NASA histoy?

[laurel]

Armstrong was on Gemini 8 with Dave Scott.

[ajv]

I'm not entirely sure I understand exactly how this works

Well, I'm not sure either. Here are some more thoughts.

The calculation was performed (and the error was made) on Earth (MCC-H).

At 189:58:46 GET the Capcom reads up a backup guidance quantity preview for the retro burn containing a time of 12:27:43 (Actually the technical transcript has :53 but that's probably a typo in that transcript - the PAO transcript has :43). That's the retro burn time in GMT without the days component so such times without the days like that are passed around.

If I were to write the calculation these days (and assuming I wasn't writing a man-rated system) I'd just jam the DDD HH:MM:SS figure into a double precision variable and trust that the trig functions would work. In the days when you had a limited amount of precision it would make sense to rewrite the calculations.

We're trying to calculate the longitude (or the location in space based on a longitude) so it we don't need to handle multiples of 360°. (BTW I'm not saying the computer calculations were done in degrees - e.g. the AGC used radians - but we can stick with degrees for the moment).

The Earth longitude will be (longitude in Fixed Earth coordinate + Earth rotational coordinate correction) where the Earth rotational correction will be a computation of the sidereal rate * days since the Fixed Earth coordinate system was defined

360.986° * ddd.ddddddd
= (360.986° * 0.ddddddd) + (360.986° * ddd)
= (360.986° * 0.ddddddd) + (0.986° * ddd) + (360° * ddd)
= (360.986° * 0.ddddddd) + (0.986° * ddd)

because the (360° * ddd) ends up as 0°

So it might make sense to split up the calculation where we treat the days separately from the HH:MM:SS. The (0.986° * ddd) is the value that ends up as 0.986° * 8 = 7.888°.

The omission of the elapsed number of days in the G.m.t. of retrofire term could mean that either the ddd figure was fed in as zero (as a data input error) or that the whole term was missing from the computer program (perhaps it had been dropped in the original program specification). The comment that The MCC-H computer program was correct as written could mean either.