6
$\begingroup$

The three-pointer dial altimeter (with a shortest pointer indicating tens of kilofeet, a short pointer indicating kilofeet, and a long pointer indicating hectofeet), the most instantly-recognisable style of analog barometric altimeter, is also, apparently, the type most vulnerable to misreading, with a number of accidents either definitively attributed to, or suspected to have been caused by, pilots misreading three-pointer altimeters (frequently in the form of believing one’s altitude to be ten kilofeet higher than one it actually is).1

Examples include:

  • A BOAC Bristol Britannia on a test flight crashed on 24 December 1958 while on approach to land; testimony from surviving crewmembers definitively identified a misread altimeter as the cause of the crash.
  • United Airlines Flight 389 crashed into Lake Michigan on 16 August 1965 after the flightcrew failed to arrest its descent at the altitude to which they had been cleared; the cause of the continued descent into the water was never conclusively determined, but a misread altimeter is strongly suspected.
  • Iberia Airlines Flight 62 descended below its clearance limit on 4 November 1967 and crashed into a hill in West Sussex; the cause of the crash is not known with certainty, but one of the possibilities is a misread altimeter.

Further, to quote the NTSB report on the crash of UA389:

As a part of this investigation a review of research on altimeter reading problems was conducted. There are some fifty studies relating to the various altimeter systems of presenting information to pilots which have been conducted since 1947. One of the most recent studies completed by the Naval Research Laboratory, Washington, D. C. 16/ was conducted to determine the relative effectiveness of various altimeter displays. These investigations, consisted [sic] of laboratory research at NRL and flight test [sic] at a nearby naval air facility, were in support of a Department of Defense program. Four types of altimeter presentations, the counter-pointer (CP), counter-drum-pointer (CDP), drum-pointer (DP) and three pointer (3P) were compared in a series of laboratory experiments measuring reading time and accuracy with both pilots and non-flying enlisted men. The same instruments were also compared by the pilots in flight, using the questionnaire technique to elicit pilot judgments. The results of the various phases of laboratory research showed that the three-pointer altimeter consistently ranked poorest. The pilot preference in the flight tests was overwhelmingly in favor of one of the other types of altimeter. In all objective measurements taken, the three-pointer altimeter ranked the poorest among the instruments tested. The finding that the three-pointer was poorest proved to be statistically reliable in every study in which tests of statistical significance were applied. In a preceding part of this study 17/ the mean exposure time and number of errors on each altimeter was recorded for the two test groups, one consisting of pilots, and the other of non-pilot enlisted men. In 1080 readings of the four altimeters by 18 pilots the mean exposure time in seconds was 0.80 for the CP, 0.84 for the CDP, 1.38 for the DP, and 2.28 for the 3P. The number of reading errors was 7 for the CP, 14 for the CDP, 26 for the DP, and 80 for the 3P. The same data collected during 420 reading trials using 7 non-pilot enlisted men resulted in mean exposure times of 0.85 for the CP, 0.86 for the CDP, 1.50 for the DP, and 2.93 for the 3P. The number of errors by this group were 5 for the CP, 2 for the CDP, 10 for the DP, and 52 for the 3P. The mean number of errors made on each altimeter for pilots was 0.39 CP, 0.78 CDP, 1.44 DP, and 4.44 3P, and for enlisted men 0.71 CP, 0.29 CDP, 1.43 DP, and 7.43 3P. [Pages 24-25 (paper report)/28-29 (PDF).]

This is quite surprising, given that the three-pointer altimeter display works on the same principle as an analog clock face, and even nowadays (with the proliferation of digital clock displays), the vast majority of people can rapidly and accurately read the time displayed on an analog clock; back in the 1960s (before the digital clock was a consumer item), practically every literate adult would have been able to do this. It shouldn’t be hard to translate those skills into reading altitude rather than time, so that, instead of (say) 3 hours (shortest pointer) 25 minutes (short pointer) 32 seconds (long pointer) past midnight, you have (say) twenty- (shortest pointer) -five thousand (short pointer) seven hundred (long pointer) feet above mean sea level.

What gives?


1: It should be noted, however, that other altimeter styles are not immune to being misread; see, for instance, AA320 and SA228, both involving the misreading of drum-type altimeters.

$\endgroup$
  • $\begingroup$ Related $\endgroup$ – Pondlife May 26 at 17:58
  • 1
    $\begingroup$ Perhaps the claim that people can "rapidly and accurately read the time" is rather optimistic, esp. when seconds must be read too (3 hands). I wonder if there was a proper research into that. $\endgroup$ – Zeus May 27 at 5:02
  • $\begingroup$ Actually, in altimeter, the pointers are the long hand for hundreds, short somewhat trapezoidal hand for thousands and the triangle or point running on the outside for tens of thousands. And the triangle is of course often at the end of a very thin hand, so it makes it the longest, but it is the highest order, not lowest. $\endgroup$ – Jan Hudec May 27 at 20:27
  • 1
    $\begingroup$ "the vast majority of people can rapidly and accurately read the time displayed on an analog clock". You obviously do not spend much time around many US school children. It seems very few of them can read an analog clock. So few that, instead of teaching them how, many US schools are removing them from their buildings. (Seems bassakwards to me, but that's what they seem to be doing.) $\endgroup$ – FreeMan May 28 at 12:47
2
$\begingroup$

enter image description here

The altitude is quarter to 10,000. No wait, one minute before 8,000. Let's see, 5 divisions between 7 & 8 so altitude is 7,800 and zero hundreds.

No it isn’t. From Jan Hudec’s comment:

The thin hand is a bit strange—normally it is just the triangle running on the outside or the hand itself is black—but the triangle means it is the 10,000 one, so it should be 780, not 7,800. The long straight hand is always hundreds, the short trapezoidal is always thousands and the triangle on the outside is always tens of thousands. Still clearly demonstrates just how confusing it can get indeed.

It is a design issue: it looks like an analogue clock but works differently. The long thin one denotes the major units, not the minor one of the second hand. Units can be 10,000, 1,000 or 100 foot. The thousands are probably most intuitive, the 10,000s and the 100 are not.

Perhaps a different story in China, Japan and Korea, where there is a dedicated word for 10,000

$\endgroup$
  • 3
    $\begingroup$ On top of this, the cockpit is not always an easy environment - you're bouncing in turbulence, avoiding deadly weather, managing a low fuel state... mistakes are easier to make when you've got all that too. $\endgroup$ – Ben May 26 at 8:54
  • $\begingroup$ The link is dead... $\endgroup$ – sweber May 26 at 17:26
  • 2
    $\begingroup$ The thin hand is a bit strange—normally it is just the triangle running on the outside or the hand itself is black—but the triangle means it is the 10,000 one, so it should be 780, not 7,800. The long straight hand is always hundreds, the short trapezoidal is always thousands and the triangle on the outside is always tens of thousands. Still clearly demonstrates just how confusing it can get indeed. $\endgroup$ – Jan Hudec May 27 at 20:31
  • 1
    $\begingroup$ @Sean, I haven't seen that many different real instruments, but judging from images and simulators the triangle on the outside seems to be most common for tens of thousands. There is not a complete consistency though, so some instruments may use very short cross-shaped hand or other shapes. The other hands are almost standard, but there are probably some exceptions too. $\endgroup$ – Jan Hudec May 28 at 5:16
  • 3
    $\begingroup$ I think the discussion in the comments is the answer! It demonstrates just how confusing this is! $\endgroup$ – FreeMan May 28 at 12:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.