IRS Alignment (why 10 mins ?)

[kuwait340]

Hello....

why the IRS's needs approx 10 min's to align...why not less ?

i remember once i read that they need 15 mins at higher latitudes .

what are the factors governing the time ?

thanks

[BOAC]

There are a lot of things the kit has to do, including resolving earth horizon and true north (matching sensed earth rate with position to establish notional axis 'alignment') and so on. Real black box stuff! It takes that long to fine tune all the results to give the required accuracy. It could have a working 'alignment' in a couple of minutes but the platform would not be accurately 'aligned' and the drift would be significant over time. A 'rapid realign' starts from a platform that is 'aligned' and merely zeroes all the accrued position/ground speed errors but any platform 'misalignment' is not removed so the errors will build up more quickly after.

The further north/south you go (nearer the poles) the lower is the earth rate in the roll and pitch sense and it becomes more and more 'yaw' only, reducing the number of factors it can use to 'align' the platform to achieve a heading.

A bit 'simplified' but lots of alignment info via Google if you look.

[spannersatcx]

only takes 7 mins in the UK.

[Short_Circuit]

A tidbit from the 744 maintenance manual about alignment time vs. latitude;

(g) Ten minutes are required to complete the alignment for any
latitude between 70.2 degrees North latitude and 70.2 degrees
South latitude. This is called a regular alignment.

(h) 17 minutes are required to complete the alignment for any
latitude between 70.2 degrees and 78.2 degrees. Since the IRU
will automatically enter the navigation mode after 10 minutes
if the mode select switch is in the NAV position, a high
latitude alignment requires that the mode select switch be left
in the ALIGN position for 17 minutes before being selected to
the NAV position.

[BHenderson]

There are no moving parts in an IRS so why 10 minutes? Surely it's just simple number crunching by the computers?

[Pugilistic Animus]

Bhenderson: remember the first PCs? ---Whew! for them to boot up sometimes took longer than building to stupid thing

-I believe that BOAC has answered la pregunta que tuviste pa arriba---no?

not just a matter of computer 'warm up' but also the cofirmation of position

[Capt H Peacock]

The laser ring gyro is a ‘true gyro’ in that it measures angular displacement (turning to you and me). The laser ring has no idea that it’s on the Earth’s surface, and it simply reports that angle that it has turned through to the IRS systems.

If you tell the IRS where it is on the Earth’s surface, then as the Earth turns, the laser ring senses a turning motion.

The Earth turns 360˚ in 24 hrs, that’s 15˚ an hour.

At the North Pole, the laser ring would sense 15˚ of turn every hour in the normal (horizontal) plane, because the axis of rotation passes through the pole. At the equator, the angular motion is in a ‘tipping’ sense, ie although not rotating in the normal plane, the laser ring ‘tips’ toward the rising Sun – eastwards – at 15˚ per hour.

If you know the position of the laser ring at a given time, at a certain time later, you can say that some of the angular displacement has been from normal rotation, and some from tipping. If you know what your latitude is, then it’s just a matter of doing the sums. In the northern hemisphere for instance, the direction of true North will move westward by 15sin(latitude) per hour, and if your laser ring moved 10˚ in 1 hr, then the latitude would be 42˚N. Hence, if we had input 42˚N into the IRS at ALIGN, then the laser ring could deduce the direction of true North.

In the old days of spinning iron gyroscopes and analogue computers, a platform could take 30 minutes to align, not least because a conventional gyroscope measures angular acceleration which needs to be integrated to provide an angular displacement. A laser ring gyro can accurately sense a 1˚ change in angular displacement directly, and hence 7 minutes at mid-latitudes is sufficient to divine the direction of true North.

Rather fascinatingly, and one of the things that really impresses me about Mother Nature, is that the IRS platform behaves like a huge pendulum with its bob at the centre of the Earth. If the pendulum is displaced, it will swing with a period of about 85 minutes, a phenomenon called the ‘Schuler Loop’. Hence if the initial position input to the IRS is wrong, the heading and attitude outputs from the IRS will wobble about every hour and a half.

Isn’t physics great fun?

[Pugilistic Animus]

Captain H. Peacock ---Who says that Pprune has 'gone down' in quality?

That was a very learned and interesting discourse!!!

there's so much offered on this site its unbelievable!!

[kuwait340]

Guys....this is a great effort to explain the theory behind the IRS alignment.

Thank you Very Very much.

[mono]

Captain H. Peacock ---Who says that Pprune has 'gone down' in quality?
That was a very learned and interesting discourse!!!
there's so much offered on this site its unbelievable!!

Most of it however was wrong!!
I will try to deal with them item by item.

The laser ring gyro is a ‘true gyro’ in that it measures angular displacement (turning to you and me).

Two errors here. First. The RLG is in fact not a gyro at all - it has none of the properties of a gyro (precession and rigidity). It is in fact a laser interferometer. Secondly. The RLG is a rate device, it measures the rate of angular displacement. Once the angle has changed and the heading/attitude/etc is constant there is no RLG output (dither oscillation excluded). Rate outputs provide much higher accuracy because they can be tuned such that small rates of change give large outputs.

At the equator, the angular motion is in a ‘tipping’ sense, ie although not rotating in the normal plane, the laser ring ‘tips’ toward the rising Sun – eastwards – at 15˚ per hour.

The maths is correct but The RLG can only sense rate of rotation in one plane. For that reason there are in fact 3 RLG's in a typical inertial reference computer. These will be orientated such that they sense movement about the a/c lateral, longitudinal and normal axis.

In the northern hemisphere for instance, the direction of true North will move westward by 15sin(latitude) per hour, and if your laser ring moved 10˚ in 1 hr, then the latitude would be 42˚N. Hence, if we had input 42˚N into the IRS at ALIGN, then the laser ring could deduce the direction of true North.

An IRS can quite happily calculate latitude and the true North-South axis without any operator input. What it can't resolve is the ambiguity. Is it north or south of the equator and which direction on the axis is north? For that it needs to know its longitude and either N or S wrt latitude.

In the old days of spinning iron gyroscopes and analogue computers, a platform could take 30 minutes to align, not least because a conventional gyroscope measures angular acceleration which needs to be integrated to provide an angular displacement.

As previously stated the RLG is a rate device. It measures angular velocity. Integration of its output is still required to calculate angular displacement.

A laser ring gyro can accurately sense a 1˚ change in angular displacement directly, and hence 7 minutes at mid-latitudes is sufficient to divine the direction of true North.

Try 0.01 degrees. These things are very sensitive.

Rather fascinatingly, and one of the things that really impresses me about Mother Nature, is that the IRS platform behaves like a huge pendulum with its bob at the centre of the Earth. If the pendulum is displaced, it will swing with a period of about 85 minutes, a phenomenon called the ‘Schuler Loop’.

Not so much a mother nature phenomenon as a man made one. In order to be accurate IRS (and INS) platforms need to be "Schuler tuned". This is so that they ignore angular movements due to the curvature of the earth (in effect they are what we call "tied" to the earth). Older gymbal based platforms were tied mechanically via pendulous sensors that precessed the gyros and kept the platform level. Modern strapdown systems are tied electronically with speed, altitude and other parameters used to calculate the expected drift and then cancel it out. This Schuler tuning gives rise to the schuler period phenomenon. Failure of the schuler circuits can often lead to schuler period oscillations.

[Dream Land]

mono, thanks for the lesson!

[Pugilistic Animus]

I'm still thankful as the original discourse did indeed introduce some of the specific terminology. This was used to thoroughly explain the whole deal, perhaps some technical details were a bit 'over simplified', I've learned a good deal---more than I ever knew about this topic, so I'm thankful to both posters for the lesson---that empahsizes the quality of the technical experiences of the posters on this site---wonderful indeed!

[BOAC]

My advice to anyone seeking to understand 'fixed' ie laser-ring platforms is to start with the old-fashioned gyro-based platforms and understand how they aligned. That way you will learn about earth rate and schuler etc. It makes it much easier to get to grips with something that not only does not move, but does not hum or even glow red

For BHenderson - the 'time' is basically needed because the blackboxery has to see some 'changes' before it can carry out its accurate alignment.

[kuwait340]

Gents Thanks very very much indeed.

[fantom]

If Kuwait340 had asked the question in 1994, he could have come to Fintas Towers and met the man who invented the system. No joke; he was there, and I met him. His son was a First Officer on the 300/310 at the time.