Compass errors

graham

New member
For my sins I seem to spend quite a lot of time playing with other people's survey data as well as my own.

Over the weekend I have been having an interesting exchange about compass errors. We found a discrepancy between two surveys of the same cave, done thirty years apart, of up to 14 degrees. Now, it is probable that 4 degrees of that is down to magnetic deviation, but that might possibly leave a compass error of ten degrees.

This was no surprise to me, as much of the data I have used has had anything up to 12 degrees of correction needed, though much of that was Mag Dev, in Ireland where it's bigger than round here, but was found to be alarming by others.

So my question is: what sort of compass errors have people had to deal with?
 

Peter Burgess

New member
I have found places with anything up to 20 degrees variance, but these were not in comparison with other surveys, just errors that messed up loop closure, which were easily detected by comparing fore and back bearings. The cause was almost always hidden steelwork, or live HT cables close by! That said, I have found significant differences between two people measuring the same bearing, possibly caused by things like steel spectacle frames or steelwork in or around the helmet. That could explain why one of 'your' surveys is consistently at variance with another if one was made by someone with a steel plate in his head.
 

graham

New member
I'm not talking about unusual situations or blunders, Pete, but about systematic errors on a compass.
 

Ed W

Member
Graham,

I use fore and backsights for a lot of my surveying, and this usually means callibrating two person/instrument combinations.  I don't have any data in front of me, but differences of 1-2 degrees between people with different instruments seem to be the norm.  I would guess that this is part person related and part instrument related.  I have seen cases of up to 3.5 degrees.  In most cases these differences are repeatable with the same person/instrument combination. 

I have never seen a systematic error of 10 degrees, though I guess it is possible.  However, if the difference in bearing is consistent with historical data, then I don't see why you shouldn't just rotate the old data to fit, if someone wants to argue with you they can always measure it themselves!

Hope this is useful,

Ed
 

Peter Burgess

New member
How can you tell? Anyway you asked "what sort of compass errors have people had to deal with?" and I provided an answer. I won't bother next time. I have seen two surveys, both seemingly consistent in layout but about 30 degrees out. I have no idea why this was so, but I suspect it was a draughting error not a surveying error.

 

graham

New member
Peter Burgess said:
How can you tell? Anyway you asked "what sort of compass errors have people had to deal with?" and I provided an answer. I won't bother next time. I have seen two surveys, both seemingly consistent in layout but about 30 degrees out. I have no idea why this was so, but I suspect it was a draughting error not a surveying error.

Oh for christ's sake get down off your high horse. I'm sorry if I wasn't clear enough first time round.  ::)

In this case it is most certainly not a draughting error as I have recompiled the data from the original in both cases.
 

footleg

New member
When I organised a surveying workshop a few years ago, and we had several sets of instruments to play with, we discovered that one compass read around 10 degrees out compared to all the others. This was spotted during instrument calibration training.

I also suspect that some people have lamps which affect the compass reading but do not realise. People are often quite blas? about this sort of thing. I think they do not want to admit that their much loved lamp is not suitable for the job. Of course in this case you would expect to see some variation in the error from leg to leg. If the error between the surveys appears consistent over different part of the cave then an instrument calibration error would be the cause I think. 
 

martinr

Active member
graham said:
... a discrepancy between two surveys of the same cave, done thirty years apart, of up to 14 degrees. ...........

Would one of these be an MCG survey of Longwood by any chance?
 

graham

New member
martinr said:
graham said:
... a discrepancy between two surveys of the same cave, done thirty years apart, of up to 14 degrees. ...........

Would one of these be an MCG survey of Longwood by any chance?

;) No, I'm afraid that I have yet to see the 1963 data, if it still exists. Fortunately a very clever chap in the BEC has digitised the plan and kindly let me have a copy of his data. I am now only lacking 3 or 4 bits of the puzzle from ST474 to ST499 (both Eastings) and from ST563 to ST552 (Northings). It's a fascinating area, might have some interest to some of your guys, Martin, who I believe have found a bit of cave to the east of here. It clearly shows where the best places to dig are, I think.

I am afraid I cannot upload the model for a while, yet, though as much of the data is not mine and its owners want to finish a few tasks off, first.
 

graham

New member
Footleg, what sort of instrument errors were there on the other compasses? I routinely see 3-5 degrees on the data, even not allowing for Mag Dev.
 

footleg

New member
graham said:
Footleg, what sort of instrument errors were there on the other compasses? I routinely see 3-5 degrees on the data, even not allowing for Mag Dev.

We only saw +/- 1 degree at most in sighting errors between different people across all the other instruments against the calibration targets. But then they were learning to read them, so they kept checking fore and back sights until they were happy they were getting the right sort of readings.

These were all Suunto compasses. Maybe one of your sets of data was from an inferior compass design? Or surveyed during a period of intense solar activity (which I'm told can deflect the magnetic field by a few degrees at times)? You could rule that out if the data was collected over a longer time period rather than all on one date.
 

martinr

Active member
graham said:
martinr said:
graham said:
... a discrepancy between two surveys of the same cave, done thirty years apart, of up to 14 degrees. ...........

Would one of these be an MCG survey of Longwood by any chance?

;) No, I'm afraid that I have yet to see the 1963 data, if it still exists. Fortunately a very clever chap in the BEC has digitised the plan and kindly let me have a copy of his data. I am now only lacking 3 or 4 bits of the puzzle from ST474 to ST499 (both Eastings) and from ST563 to ST552 (Northings). It's a fascinating area, might have some interest to some of your guys, Martin, who I believe have found a bit of cave to the east of here. It clearly shows where the best places to dig are, I think.

I am afraid I cannot upload the model for a while, yet, though as much of the data is not mine and its owners want to finish a few tasks off, first.

Do you mean you digitised the MCG survey? I ask beacause I believe there was an error in the north point on the survey!
 

graham

New member
martinr said:
Do you mean you digitised the MCG survey? I ask because I believe there was an error in the north point on the survey!

I didn't, but it has been done. I know about that North point error. I think it was corrected on the last iteration, the one that has the Moody's Reynold's Passage extension on it.
 

graham

New member
footleg said:
graham said:
Footleg, what sort of instrument errors were there on the other compasses? I routinely see 3-5 degrees on the data, even not allowing for Mag Dev.

We only saw +/- 1 degree at most in sighting errors between different people across all the other instruments against the calibration targets. But then they were learning to read them, so they kept checking fore and back sights until they were happy they were getting the right sort of readings.

These were all Suunto compasses. Maybe one of your sets of data was from an inferior compass design? Or surveyed during a period of intense solar activity (which I'm told can deflect the magnetic field by a few degrees at times)? You could rule that out if the data was collected over a longer time period rather than all on one date.

The 1970's survey was done in two trips over 3 months, the 2003/4 one in three trips over 19 months.

However, I have now cracked the problem which was done to an incorrectly entered calibration!

Ho Hum  :-[
 

SamT

Moderator
you two a like a right pair of old women today, Anyone would have thought you two had pmt

sheesh  ::)
 

SamT

Moderator
we're laaaadies

little-britain-cover.jpg
 
Top