757 Magnetic Heading Error

[BL]

I have noticed that the 757s that I fly all seem to have a 3 degree error in magnetic heading. (Published ILS course is ALWAYS 3 degrees less than magnetic track flown on ILS or the track maintained during takeoff on that runway)

Magnetic variation on the maps/charts/aerads is approx 3W in western Europe. Switching from Magnetic to True changes the heading or track by 6 degrees lower. eg mag hdg 090, true hdg 084.

This means that our aircraft think that the mag variation in is 6W, and not the published 3W.

Speculation time:
Does the mag variation that the 757 uses come from the IRS computers or the FMC database?

If it comes from the IRS computers, could it be that our 10-15 yr old 757s are still using the mag variation from 10-15yrs ago, hence the discrepancy?

Any thoughts/info?

[Hew Jampton]

The magnetic variation data are stored in the FMC (I think, can't remember for sure). Unfortunately the information is hard wired into the system during construction so is not capable of rapid update like the monthly AIRAC NOTAM data changes, for example. It requires a hardware mod, presumably a new chip, to change from the variation map used at the time of construction to the later one, at a cost of some thousands or even tens of thousands of dollars. Great planning by somebody. There was an article in Boeing's Airliner a few years ago; apparently there are two variation maps, one is about twenty years old (the one in most aircraft), the other is about five years old and is presumably only in aircraft built since then.

[QAVION]

My sources tell me that there is both a magvar table in the IRS and the FMC. The IRS table is used for current position and the FMC table is used for displaying flightplan hdg data (further down the track). If this is correct, then, in your case, it sounds like the IRS's need updating.

Hope this helps.

Rgds.
Q.

[ICT_SLB]

May depend, as well, on the maker of the IRUs fitted. Litton use a look-up table model derived, I was told, from a Swissair database. Honeywell use an algorithm.