Leaky wet compass

[jesse]

Think about the tech in the phone for a sec. There's accelerometers, and a telco-network assisted GPS chipset. AFAIK there is no magnetic compass inside the iPhone. But I could be wrong. Easy to check... grab a magnet... get it close enough to have an effect, but don't fry the phone in the process.

The iPhone 3GS and later have an actual electronic compass - iPhone 3GS uses the AK8973 3-axis electronic compass chip. iPhone 4 uses similar.

How exactly would a magnet fry the iPhone?

Most "Compass" Apps on iPhone merge the GPS data and the accelerometer data themselves. (I don't believe this is in the API, but I'd have to go research... I stopped looking at the Apple API's in iOS 3 days, 'cause I just didn't have time to code up what I wanted to code for the gadget.) Some do a sane job of it, some won't.

Most of them base it off the compass data which is provided in the API. See the CoreLocation framework documentaion, in particular the CLHeading Class:
http://developer.apple.com/library/...Reference.html#//apple_ref/doc/uid/TP40008772

 

[flyingcheesehead]

Think about the tech in the phone for a sec. There's accelerometers, and a telco-network assisted GPS chipset. AFAIK there is no magnetic compass inside the iPhone. But I could be wrong.

You ARE wrong. There's a real magnetic compass in the iPhone.

Easy to check... grab a magnet... get it close enough to have an effect, but don't fry the phone in the process.

Nothing to fry with a magnet - No magnetic storage media, no CRT, etc.

FWIW, the iPhone does recognize when there's a magnet nearby. When I opened the compass app this morning, the iPhone was only about 2-3 inches from the hard drive in my laptop (which, as I'm sure you know, contains a very strong magnet). The compass app threw an error about magnetic interference...

... and then when I moved the phone away it started working - Surprisingly well! And I was standing still, and not moving the phone in a way that the accelerometer would recognize (level plane motion only).

(Of course, an App could offer a choice here... true or magnetic. Actually haven't seen one yet that does, but have only tried four and a couple of "dashboards".)

What are you, an Android user? The stock iPhone compass app offers the true/magnetic choice - Tap on the little i in the lower right and you get:

 

[denverpilot]

The iPhone 3GS and later have an actual electronic compass - iPhone 3GS uses the AK8973 3-axis electronic compass chip. iPhone 4 uses similar.

Ah, got it. Found a datasheet marked "preliminary" with a quick Google search, and noted that the phone/device manufacturer must insert any necessary offset into the chipset for any magnetic devices around the chipset (speaker, microphone) to calibrate it.

Also notes in the datasheet about the maximum magnetic levels the design should subject the chipset to, since speakers vary with the audio level and frequency they're driven at. So sounds coming from the phone during compass readings could have an effect. Same thing with (relatively) high RF levels if the network were to tell the phone, "Can't hear you now!" (Joke. But true.)

http://www.datasheetarchive.com/Indexer/Datasheet-014/DSA00244391.html

How exactly would a magnet fry the iPhone?

Depends on what's inside and how big a magnet someone reading my post might go pick up to try it. :wink2: I was just trying to put a disclaimer on it so someone didn't go grab some giant magnet, slap it up to their phone, and yell at me tomorrow, ya know?

Most of them base it off the compass data which is provided in the API. See the CoreLocation framework documentaion, in particular the CLHeading Class:
http://developer.apple.com/library/...Reference.html#//apple_ref/doc/uid/TP40008772

Yup, which was kinda the jist of my point... you have no idea in most Apps you download what they're using. They could be making it up for all you know. And has the Apple CoreLocation library itself been audited by anyone who knows navigation?

There's also a multiplexer in the device to switch/decide which axis you're using... so CoreLocation must do some guesswork using the accelerometers about which orientation the phone is in, relative to the Earth's center. In an aircraft, 60 degree banked turn that's coordinated... those accelerometers think the floor of the aircraft is the "center of the Earth", I bet.

And a pre-amplifier that can be adjusted by external voltage for sensitivity...

Cool to see that Apple's using a typical I2C bus in there somewhere. That's some interesting stuff, and fun to "bit-bang" on microcontrollers if you have a lot of time on your hands. My microcontroller days are long gone unless I can find a whole lot of hours left in a week to dig out all that stuff from the basement. I was kinda partial to the Microchip PIC, but wanted to try some coding for the Atmel AVR line also and never got around to it.

You probably see what I'm getting at here... Compass Apps on the phone, are probably best thought of as toys... right up until you've done a whole lot of your own testing against a known good compass.

Then you know that at least *your* phone is fairly accurate... but swap hardware, you better test it again, if you're going to rely on it for anything other than playin' around.

You've given me interesting ideas though... like playing with Compass Apps to see if they follow ANDS like a real compass would... or that question above... what does this thing do in steep turns? Would be fun to get a safety pilot and go play, and point a video camera at the phone while doing it for later analysis. If nothing else, it'd be an excuse to go flying!

(And if they don't follow ANDS or do the things "normal" compasses do... then you'd scratch your head and wonder "Why?"... Did the developer integrate the GPS information? Is the data mathematically smoothed over time by CoreLocation? How laggy is the algorithm?)

All sorts of fun tech-support engineer-geek implementation questions! I love finding other engineer's bad assumptions and patterns of mistakes in final products... it's what I do for a living! ("Mr. Customer, I've confirmed the behavior you saw, and I think I understand exactly the thought process that got us here. I've escalated a ticket with a good description of the problem, and we'll have to see what Engineering says.")

It's all relative, when it comes to compasses. If someone goes to the effort to create a compass card that accounts for errors introduced by the aircraft for a hard mount-point I wanna see it!

Even the good ol' FAA tells me on this study question for the written for my IFR ticket that I'm only supposed to make 2-3 degree corrections when on the localizer... and if they have some folks with sharp enough eyes to be able to see and fly a 2 degrees difference on my Skylane's DG in turbulence, they're bloody amazing pilots. I want to learn from them! Maybe I'll get that good someday, but that seems more like an arbitrary number taken while the pilot scanned the CDI and made the corrections that worked... hard to say. Not enough Instrument time under my belt. I could see it in nice stable day, but bumps? Just fly the needle... put the aircraft where you want it to be.

So yeah, my posting was a bit tongue-in-cheek, just to remind everyone that... "It's designed to be a phone. Not a navigation device."

CoreLocation was probably written by a guy/gal in a cushy office chair who's never needed a bearing to be accurate within 2 degrees or they're going to be miles off course, ya know?

This mixed hardware/software engineering stuff is HARD! Well, detailed and time-consuming anyway, at the very least.

Ever seen this on the Apple Support site about "Compass Calibration Mode"?

http://support.apple.com/kb/TS2767

Ever waved your phone in a figure eight? Were you supposed to lay it flat for that figure eight, or facing you, or some obtuse angle in-between? Portrait or Landscape orientation? Apple left out some details there.

Some fun discussion about it, and YouTube links of people's opinions in this posting:

http://www.iphonelife.com/blog/2440/iphone-3gs-compass-calibration

Here's a guy who's name in this forum is "B737" and he's "in a field where he can check the compass on airplanes" (paraphrased)... he says 45-90 degrees off is the worst he's seen, with it usually being about 25 degrees.

http://forums.macrumors.com/showthread.php?t=790895

Guess we know what that 737 pilot is doing in cruise!

 

[denverpilot]

What are you, an Android user? The stock iPhone compass app offers the true/magnetic choice...

Bwahaha... nice.

Nah, just not all that interested in the Phone as a Compass, other than as an Engineering exercise.

After playing with that App once or twice, I haven't opened it since then. The GPS is far more useful... as long as I can walk three or four paces then read it.

I forgot it had a setting under the "i"nfo button.

 

[flyingcheesehead]

Ever waved your phone in a figure eight? Were you supposed to lay it flat for that figure eight, or facing you, or some obtuse angle in-between? Portrait or Landscape orientation? Apple left out some details there.

Never had seen that error before, but when I got it this morning, I moved my phone away from the laptop and did a single figure eight, vertically, in portrait orientation.

BTW, the iPhone 4 actually has a gyroscope in it as well as the accelerometer - So, it should be possible to make a true backup AI these days.

 

[jesse]

There's also a multiplexer in the device to switch/decide which axis you're using... so CoreLocation must do some guesswork using the accelerometers about which orientation the phone is in, relative to the Earth's center. In an aircraft, 60 degree banked turn that's coordinated... those accelerometers think the floor of the aircraft is the "center of the Earth", I bet.

Actually the iPhone 4 has an actual electronic solid-state 3-axis gyro on top of the accelerometer.

It uses this:
www.st.com/stonline/books/pdf/docs/17116.pdf

That said - I'd have to do a hell of a lot of flight testing before I would ever try to seriously use the gyro data from an iPhone for real life aircraft handling. As you suggest there are a lot of unknowns with what the frameworks us developers are given access to are actually doing on the back-end.

 

[dmccormack]

When DTK and TRK don't match, it's simply because the plane isn't traveling parallel to the current leg of the flight plan (or direct-to leg). Adding a crosswind from either direction won't change that as long as you maintain the same track.

Exactly.

DTK = 030
TRK = 042

Guess what my heading is?

 

[John Collins]

Exactly.

DTK = 030
TRK = 042

Guess what my heading is?

Let me guess, somewhere between 0 and 359 degrees.

 

[HPNPilot1200]

Let me guess, somewhere between 0 and 359 degrees.

How'd you do that?!

 

[kgruber]

Once again there is no way to determine a heading. Please tell me how you determine a heading with a 430. You can determine the course - that is it.
)

You can't determine heading from a whiskey compass either, at least not accurately. Face it, most in GA are pieces of junk, calibrated on the ground with only certain equipment on.

Compass is just a guess of heading. One can make the case that a just as accurate determination of heading is from GPS track and guess the WCA.

 

[Everskyward]

This is not an endorsement to fly without a magnetic direction indicator but I think the vectors you get from ATC are at best an educated guess. They can't know exactly what the wind is doing at all levels and I don't think they get out their protractor. Many times I've had them make fine tune adjustments after the initial vector, "ten more left" etc. Also there have been times when I can tell the vector to final will not ever intercept final.

 

[dmccormack]

This is not an endorsement to fly without a magnetic direction indicator but I think the vectors you get from ATC are at best an educated guess. They can't know exactly what the wind is doing at all levels and I don't think they get out their protractor. Many times I've had them make fine tune adjustments after the initial vector, "ten more left" etc. Also there have been times when I can tell the vector to final will not ever intercept final.

Precisely!!!

 

[Ted]

This is not an endorsement to fly without a magnetic direction indicator but I think the vectors you get from ATC are at best an educated guess. They can't know exactly what the wind is doing at all levels and I don't think they get out their protractor. Many times I've had them make fine tune adjustments after the initial vector, "ten more left" etc. Also there have been times when I can tell the vector to final will not ever intercept final.

+1. I wouldn't make a big deal of not having a magnetic compass for ATC vectors. If you want to make a big deal of not having it for other reasons, then that's another matter. Partial panel/in case of some other failure are the biggest reasons I can come up with.

And it makes me happy that I have as many ways to do something in my plane as Lance does in his Baron.

 

[flyingcheesehead]

Exactly.

DTK = 030
TRK = 042

Guess what my heading is?

12 degrees right of what it should be - And still completely unknown. If you're in the Chief at cruise speed, chances are good that your heading is somewhere between 360 and 090, and if you're flying in the midwest and have ground contact, you should be able to get it within 10 degrees - Somebody was nice enough to print us a nice grid on the ground.

However, if you're flying over trees, rocks, etc. with few or no manmade references, or if you're flying over a solid undercast, you still don't know what your heading is.

 

[dmccormack]

12 degrees right of what it should be - And still completely unknown. If you're in the Chief at cruise speed, chances are good that your heading is somewhere between 360 and 090, and if you're flying in the midwest and have ground contact, you should be able to get it within 10 degrees - Somebody was nice enough to print us a nice grid on the ground.

However, if you're flying over trees, rocks, etc. with few or no manmade references, or if you're flying over a solid undercast, you still don't know what your heading is.

Over a solid undercast in the Chief would be dumb. I've flown over some fog, but always had gliding distance of non-fog areas.

In addition, a glance at a sectional should provide enough data as to where I am and what I'm pointing at.

I don't have a G430 in the Chief (no electric and all that) but in one of the 430-equipped airplanes that I fly I can know reasonably well what my heading is from track data -- I haven't had too many days of 300 knot crosswinds.

 

[jesse]

12 degrees right of what it should be - And still completely unknown. If you're in the Chief at cruise speed, chances are good that your heading is somewhere between 360 and 090, and if you're flying in the midwest and have ground contact, you should be able to get it within 10 degrees - Somebody was nice enough to print us a nice grid on the ground.

However, if you're flying over trees, rocks, etc. with few or no manmade references, or if you're flying over a solid undercast, you still don't know what your heading is.

Exactly. I'm not sure why there is any discussion of this. You can't determine heading from a 430. You can guess at it all you want but you're not determining heading.

Can you substitute course for heading? You can if you really want to but you should understand the limitations of what you're doing and that it is not heading.

 

[jesse]

You can't determine heading from a whiskey compass either, at least not accurately. Face it, most in GA are pieces of junk, calibrated on the ground with only certain equipment on.

Compass is just a guess of heading. One can make the case that a just as accurate determination of heading is from GPS track and guess the WCA.

The compass is an indication of heading with potential variables causing errors. The GPS track is not a indication of heading but you could make a guess at it and you're going to be by far less accurate then the compass.

Actually the magnetic compass in the air, properly calibrated, without a bunch of turbulence is pretty damn accurate. Way more accurate then you'd be making guesses using the 430, especially if you were in IMC.

The magnetic compass is a damn useful tool in a full electrical failure. I always know where the nearest VMC is when I'm flying IFR. The compass will get me there. Hopefully I won't resort to that if my handheld GPS is still working but otherwise that simple little compass could easily save my life. There is a reason it's required equipment.

 

[dmccormack]

Exactly. I'm not sure why there is any discussion of this. You can't determine heading from a 430. You can guess at it all you want but you're not determining heading.

Can you substitute course for heading? You can if you really want to but you should understand the limitations of what you're doing and that it is not heading.

::

Seriously -- here's the situation -- you're in Class Bravo, it's VFR, you're compass juice is leaking.

ATC tells you "flying heaidng 300 for traffic."

Are you seriously saying you would declare "Unable!" and then request vectors for an emergency landing?

 

[jesse]

::

Seriously -- here's the situation -- you're in Class Bravo, it's VFR, you're compass juice is leaking.

ATC tells you "flying heaidng 300 for traffic."

Are you seriously saying you would declare "Unable!" and then request vectors for an emergency landing?

Most likely I have a DG that is accuretly set and I'll be using that. If it is VMC I can set heading using all the north south roads since I know the deviation in the area. If my DG weren't set or had failed and my compass had failed I would inform ATC that I've lost heading information and will be flying courses from the GPS. No big deal but might as well tell them.

I'm not saying you can't make a GPS course work. I'm saying it is not heading and there is no way to determine heading and one should understand the limitations of navigating with a course instead of heading. Fixing the leaky compass certainly is something you should be doing after it fails.

 

[Steve]

The real source of error in magnetic compasses

http://www.youtube.com/watch?v=DpXhlGI64UM

 

[flyingcheesehead]

I don't have a G430 in the Chief (no electric and all that) but in one of the 430-equipped airplanes that I fly I can know reasonably well what my heading is from track data -- I haven't had too many days of 300 knot crosswinds.

At 100 knots, it only takes about a 35-knot wind - Pretty common - to be 20 degrees off.

Seriously -- here's the situation -- you're in Class Bravo, it's VFR, you're compass juice is leaking.

ATC tells you "flying heaidng 300 for traffic."

Are you seriously saying you would declare "Unable!" and then request vectors for an emergency landing?

No - But if I fly a track of 300 - And the other guy was assigned a heading of 300 - And we end up tracking 20-30 degrees towards each other and violating that controller's separation - I would expect to be given a phone number to call, and I would hate to have to tell the controller that the reason he's gonna get called on the carpet is because I took off with a known deficient compass in clear violation of 91.205(b)(3). At that point, I would expect a certificate suspension. Not worth it.

Now, if the compass failed, and I reported same to ATC when it happened as Jesse suggested and informed them that I'd be flying courses, not headings - Okay. But taking off with the compass already bad and heading towards busy airspace and pretending everything is OK is just a bad idea.

 

[dmccormack]

At 100 knots, it only takes about a 35-knot wind - Pretty common - to be 20 degrees off.

Sure is -- and would be real easy to figure out from your actual versus desired track now related to the nose of the airplane, wouldn't it?

No - But if I fly a track of 300 - And the other guy was assigned a heading of 300 - And we end up tracking 20-30 degrees towards each other and violating that controller's separation - I would expect to be given a phone number to call, and I would hate to have to tell the controller that the reason he's gonna get called on the carpet is because I took off with a known deficient compass in clear violation of 91.205(b)(3). At that point, I would expect a certificate suspension. Not worth it.

Now, if the compass failed, and I reported same to ATC when it happened as Jesse suggested and informed them that I'd be flying courses, not headings - Okay. But taking off with the compass already bad and heading towards busy airspace and pretending everything is OK is just a bad idea.

Let's clear this up -- I'm not arguing for takeoff with inop required equipment -- ok?

I'm merely suggesting that panic over a compass going south is a but overboard.

Telling ATC "We may have a compass issue -- can you confirm our heading?" ain't hard. But overwrought hand wringing "Gee -- am I going East??!" should not be a pilot's reaction to a bubble in the fluid.

 

[AuntPeggy]

::

Seriously -- here's the situation -- you're in Class Bravo, it's VFR, you're compass juice is leaking.

ATC tells you "flying heaidng 300 for traffic."

Are you seriously saying you would declare "Unable!" and then request vectors for an emergency landing?

Interesting redefinition of the situation.

It was daylight, we were heading into mountainous regions and cloudy skies where we would need to file IFR to continue.

What we did was find a nearby airport with mechanic and wait two days for replacement gaskets. We did not declare an emergency or ask for vectors.

Vectors, of course, are the problem. If I did need to ask for vectors, it would be after specifying relative vectors, such as "Turn left 20 degrees." instead of "Turn left to 20 degrees".

 

[AuntPeggy]

Sure is -- and would be real easy to figure out from your actual versus desired track now related to the nose of the airplane, wouldn't it?



Let's clear this up -- I'm not arguing for takeoff with inop required equipment -- ok?

I'm merely suggesting that panic over a compass going south is a but overboard.

Telling ATC "We may have a compass issue -- can you confirm our heading?" ain't hard. But overwrought hand wringing "Gee -- am I going East??!" should not be a pilot's reaction to a bubble in the fluid.

Another interesting redefinition. The "bubble" in the compass occupied about half the container.

Not a silly question. The compass was low on fluid, but still rotating. However, since the top of the fluid line intersected the middle of the markings, it was difficult to read the compass. There is a question also on the accuracy of the compass. I have no idea whether its movement was impaired due to the low fluid level and chose to believe it was.

 

[gismo]

Exactly.

DTK = 030
TRK = 042

Guess what my heading is?

Your heading could be anywhere between about 20 degrees and 60. It cannot be deduced from DTK and TRK. Even with no wind, you could be tracking 42 degrees (on a 42 degree heading) with a DTK of 222. You might as well try to compute heading by subtracting your MP in inHg from your RPM/100.

 

[Everskyward]

The solution was obviously what Peggy's relative suggested. Buy a Boy Scout compass.

 

[flyingcheesehead]

Sure is -- and would be real easy to figure out from your actual versus desired track now related to the nose of the airplane, wouldn't it?

How do you know exactly what the winds aloft are if you don't have a compass? They're often significantly different than forecast.

Let's clear this up -- I'm not arguing for takeoff with inop required equipment -- ok?

I'm merely suggesting that panic over a compass going south is a but overboard.

Who's panicing? We were simply reaffirming Peggy's excellent decision to get the compass fixed prior to taking off into an IFR situation.

Telling ATC "We may have a compass issue -- can you confirm our heading?" ain't hard. But overwrought hand wringing "Gee -- am I going East??!" should not be a pilot's reaction to a bubble in the fluid.

Nor was that anyone's reaction.

 

[Fearless Tower]

Vectors, of course, are the problem. If I did need to ask for vectors, it would be after specifying relative vectors, such as "Turn left 20 degrees." instead of "Turn left to 20 degrees".

But you have a DG right? You would be flying the vectors off the DG not the whiskey.

The purpose of the whiskey compass (or vertical card for that matter) is to set the DG initially, and then periodically use it during flight to verify that the DG hasn't drifted (some drift more than others) and also use it as a backup if the DG fails.

IF (because of the loss of fluid) you aren't comfortable using the compass to set the DG on the ground and in straight, level flight, it is difficult to read, then I could understand wanting to get it fixed before committing to an IFR flight.

But, if you are only concerned about the event that you might need to fly vectors using the whiskey, then the decision seems a little extreme.

What Dan is saying is that in an emergency (both DG and whiskey are OOC), you could get by using the GPS. In fact, some GPS units have a similated 6-pack (based all off of gps data) that in an emergency can keep you alive and get you back on the ground in the event of a vacuum failure.

I'm not second guessing your decision to want to get it repaired, just wondering about the reasoning.

 

[John Collins]

But you have a DG right? You would be flying the vectors off the DG not the whiskey.

The purpose of the whiskey compass (or vertical card for that matter) is to set the DG initially, and then periodically use it during flight to verify that the DG hasn't drifted (some drift more than others) and also use it as a backup if the DG fails.

IF (because of the loss of fluid) you aren't comfortable using the compass to set the DG on the ground and in straight, level flight, it is difficult to read, then I could understand wanting to get it fixed before committing to an IFR flight.

But, if you are only concerned about the event that you might need to fly vectors using the whiskey, then the decision seems a little extreme.

What Dan is saying is that in an emergency (both DG and whiskey are OOC), you could get by using the GPS. In fact, some GPS units have a similated 6-pack (based all off of gps data) that in an emergency can keep you alive and get you back on the ground in the event of a vacuum failure.

I'm not second guessing your decision to want to get it repaired, just wondering about the reasoning.

The OP had the compass repaired because she anticipated needing to fly IFR. It is not extreme to want to meet the applicable regulations. If she departed with an aircraft that was unairworthy and did not meet the requirements of 91.205 for the type of flight she was undertaking, it would be a violation of the FAR's. An emergency would not absolve her from departing IFR while not meeting 91.205 because it would be an emergency of her own making. In my opinion, she clearly did the right thing and sounds like the kind of pilot with good judgement that I would be comfortable flying with.

 

[Fearless Tower]

If she departed with an aircraft that was unairworthy and did not meet the requirements of 91.205 for the type of flight she was undertaking, it would be a violation of the FAR's.

Which is why I asked in an earlier post what the condition of the compass was - leaky but functional or questionable. And as I stated in response, if I had the same concerns over the ability to read the compass or questioned it's accuracy on the ground, then I too wouldn't have initiated such a flight.

My most recent post was more a question of the inflight significance of the compass.

If you know that your instrument is OOC before you launch, then the decision is pretty clear from a FAR standpoint. If the instrument is still functional and you are just concerned about the 'what if's' of that one instrument when you have other backups, then I would personally think the decision not to go would be a little extreme.

 

[AuntPeggy]

The solution was obviously what Peggy's relative suggested. Buy a Boy Scout compass.

Imagine the savings...
http://www.amazon.com/M-Tech-Boy-Sc...IVFQ/ref=sr_1_9?ie=UTF8&qid=1294679008&sr=8-9

 

[flyingcheesehead]

If you know that your instrument is OOC before you launch, then the decision is pretty clear from a FAR standpoint. If the instrument is still functional and you are just concerned about the 'what if's' of that one instrument when you have other backups, then I would personally think the decision not to go would be a little extreme.

And in this case, the situation was the former... Seems some people want it to be the latter so they can tell someone else what a silly, over-cautious pilot she is.

 

[Matthew]

At what point does a leaking whiskey compass become inop?

On a solo training flight, maybe even an XC, I was taking off and the seal on the compass let go. A lot of fluid did leak out, but there was enough fluid inside that the compass was still able to float. I was headed back home anyway so I just squawked it when I got back.

Even though the fluid level was low, it at least 'seemed' to be working.

 

[dmccormack]

Seems some people want it to be the latter so they can tell someone else what a silly, over-cautious pilot she is.

Really? Who said that?

 

[Capt. Geoffrey Thorpe]

At what point does a leaking whiskey compass become inop?

On a solo training flight, maybe even an XC, I was taking off and the seal on the compass let go. A lot of fluid did leak out, but there was enough fluid inside that the compass was still able to float. I was headed back home anyway so I just squawked it when I got back.

Even though the fluid level was low, it at least 'seemed' to be working.

They don't really float. The card rides on a bearing. The fluid is there to keep it from moving too fast.

So, it will still "work" even when it is empty.

 

[David]

They don't really float. The card rides on a bearing. The fluid is there to keep it from moving too fast.

So, it will still "work" even when it is empty.

I was wondering about that. My plane has the old style RAF compasses. The one for the rear seat was completely dry but worked just fine indicating north. With no fluid to dampen it's movement it was a bit twitchy but usable. I guess they hold around 6oz of alcohol each and weigh about 2lbs 6oz each. That's more almost 5 lbs of compasses in a plane with a gross weight of 2100lbs .

 

[Ted]

Perhaps one of our ATC guys can chime in with some experience. It seems to me that the ATC folks guess on the headings they give you. They expect some level of consistency, but I've had plenty of times that I turn to a heading, then they say "Turn [left/right] an additional 10 degrees." Clearly their heading that they figured I would be on was not what they guessed. Either I had it off a bit, my compass is off a bit, or both.

Like most other gauges, it's an indicator that has a certain amount of error. In practical instrument flying, being on to the degree isn't important for en-route sections, only approaches. Once you're on an approach, though, you've got other indicators you're following.

 

[Fearless Tower]

It seems to me that the ATC folks guess on the headings they give you. They expect some level of consistency, but I've had plenty of times that I turn to a heading, then they say "Turn [left/right] an additional 10 degrees." Clearly their heading that they figured I would be on was not what they guessed. Either I had it off a bit, my compass is off a bit, or both.

I see the same thing alot with SoCal approach and it isn't my DG or compass (verified the accuracy both before takeoff and post-flight).

 

[jesse]

I see the same thing alot with SoCal approach and it isn't my DG or compass (verified the accuracy both before takeoff and post-flight).

Well that's because there are differing winds aloft and the conditions are always changing. All they really can do is take a guess. They will develop a feel for the "winds" that day at altitude and end up doing a really good job most of the time. That isn't a reason to fly course instead of heading though because they're expecting you to fly heading - which is why they say "left heading 320"

 

[gprellwitz]

(Of course, an App could offer a choice here... true or magnetic. Actually haven't seen one yet that does, but have only tried four and a couple of "dashboards".)

What are you, an Android user? The stock iPhone compass app offers the true/magnetic choice - Tap on the little i in the lower right and you get:

Actually, the compass application I use for the Android has that option too! (And the Motorola Droid has a magnetometer/compass, too.)