Talk:Accelerometer

What about accelerometers on fan bearings?

Accelerometers are often used on fan bearings to measure vibration. I would think this is one of the most commonplace uses of accelerometers. Is there a separate page dedicated to this? —Preceding unsigned comment added by 12.41.1.54 (talk) 20:11, 17 April 2009 (UTC)

I doubt it. If you honestly think that there is a need for such a page, be BOLD and create one. But if it gets deleted, don't be surprised. I personally don't think we'd need 2 separate pages; I could be wrong. There would be too much repeated material IMHO. 98.202.38.225 (talk) 19:28, 10 June 2009 (UTC)

Accelerometer on fan bearings of computer can be used for hacking PC. — Preceding unsigned comment added by Davidy2001 (talk • contribs) 07:42, 24 November 2018 (UTC)

References

The introduction repeats itself

The introduction repeats itself when it says: "An accelerometer is an instrument for measuring acceleration" and It feels like there are two different explanations of accelerometers --76.22.79.194 00:42, 15 April 2007 (UTC)

This shouldn't be in the physics section, but in the electronics section. A list of low-cost (under $10) devices would be useful for hobbyists and engineers. 67.172.182.35 07:04, 24 January 2007 (UTC)

I think the "types of accelerometers" section needs revision. The list should be split into general types, maybe a MEMS and non-MEMS categorization, probably without vendors listed.

needs some work.

i believe the nike ipod thing is a piezo electric force sensor, not an accelerometer, and the segway's balancing comes from MEMS gyroscopes, not accelerometers... —The preceding unsigned comment was added by 137.189.101.14 (talk) 10:09, 26 March 2007 (UTC).

The iPod + Nike thing uses an piezoelectric accelerometer, see [[1]] Phiren 07:57, 5 September 2007 (UTC)

The iPod + Nike thing is NOT an accelerometer in any way. It is a piezo speaker used in reverse. By Flexing a Piezoelectric device it produces a voltage that the device measures. If anything, this is a force or strain sensor more then an accelerometer. See the article on Piezoelectrics in wikipedia to see a picture of a piezo speaker, then take a look at a disassembled iPod + Nike device at SparkFun.com http://www.sparkfun.com/commerce/present.php?p=Nike_iPod-Internals. The simple test is this: Can you use it to measure gravity? No, you cannot which contradicts the very first paragraph in this article. It only measures a change in applied force by a runner's foot. —Preceding unsigned comment added by 66.238.211.199 (talk) 20:58, 2 January 2008 (UTC)

I would suggest merging the Uses section into the Applications section; if it becomes too long two or three subsections could be added. If no one protests, I will do it in one of the following days. Pietrow (talk) 09:45, 24 August 2008 (UTC)

As noted above, only the Applications section remains. I was able to remove some redundant information, although the Electronic devices subsection is still a bit out of shape. I hope my edit is not too contentious. I also removed the trivia section, because I think applications of the object described is not at all trivia; in fact I would argue it is the second most important section. Pietrow (talk) 10:37, 1 September 2008 (UTC)

gyroscope versus accelerometer

I think this article may confuse gyroscopes with accelerometers in a couple places. The introduction correctly implies that these are different concepts but I would guess image stabilization and segways both use gyroscopes without accelerometers, since I would guess translational shake would be a 2nd order consideration compared to rotational 'shake' in blurring a picture, and I would think the segway would know what its wheels are doing and would only need orientation information (unless it has an anti-slip feature, but even there you'd probably watch sudden changes in torque or rpm instead? You'd probably have to lose balance anyhow at that point...)

If someone can confirm, or can find another example which can be confirmed, then perhaps they could change this.

I agree. I think another instance is the tilt sensing of mobile phones in the consumer electronics section. The articles on gyroscopes (on three axes) and tilt sensors (on two axes) clearly indicate that it's those two that allow this, not the accelerometer -- if the name "tilt sensor" isn't enough proof already. ;-) Allyddin Sane (talk) 19:16, 19 October 2011 (UTC) PS: Just noticed that the same section mentions tilt sensor in the last p, but it still needs some clearing up IMO.

gravitation and vacuum

In the introduction it states: "If dropped in a vacuum it will read zero.". I think the fact that the accelerometer resides in a vacuum does not necessarily imply that there is no gravitational force taking effect. —Preceding unsigned comment added by 213.39.233.138 (talk) 19:37, 18 October 2007 (UTC)

What that editor means is when something is dropped, while it is dropping, it is weightless. Gravity is always there, but what's lacking is the reaction. What keeps you from falling through the floor? Answer: the floor. So to keep you from falling, the floor is working against gravity, pushing you up, supporting your weight. That's the force an accelerometer is actually measuring at 1g. So in a vacuum, with a freefall, there's no air resistance to slow you down, there is no floor to push you up, perhaps the floor is also falling, and the accelerometer is falling. What's stopping you from falling that an accelerometer can measure? What's stopping the accelerometer from falling? Nothing. And that's what the accelerometer measures. Remember, gravity is affecting everything without difference, you, the falling floor, the accelerometer; and a difference of force is what an accelerometer measures. See weightlessness. MMetro 18:05, 28 October 2007 (UTC)

....... —Preceding unsigned comment added by 116.68.69.83 (talk) 15:23, 15 March 2008 (UTC)

About gravity

This text implies that all accelerometers are able of measuring DC acceleration. This is far from real. All vibration monitoring accelerometers usually don't measure anything below 1 Hz. —Preceding unsigned comment added by 201.76.191.59 (talk) 18:17, 20 October 2008 (UTC)

Twisted ribbon FM accelerometer

What is the twisted ribbon FM accelerometer mentioned in this article? __meco (talk) 17:10, 12 December 2008 (UTC)

A comment on this revert war.

Accelerometers are, in general, are based on "effective force" sensing. At this stage, the output will be proportional to (a+g). Then some accelerometers bother removing "g", either through software, or through a voltage offset. Others don't. Stop saying that accelerometers don't feel the 1g. They all feel the 1g. It's just that some choose "1g downwards" to be the reference, and will give you the difference instead, exactly like how you can set a scale to read "−0.454 g" when nothing is on it because the box your packing stuff into has a mass of 0.454 g. Headbomb {^ταλκ_{κοντριβς} – WP Physics} 19:46, 22 January 2009 (UTC)

Nope, accelerometers measure a-g. You can calibrate the g away (if you're stationary or something), but if it changes it changes the opposite way than a- if g goes down, the output goes one way. If a goes down, the output goes the other.- (User) Wolfkeeper (Talk) 20:33, 22 January 2009 (UTC)

Sign conventions are trivials.Headbomb {^ταλκ_{κοντριβς} – WP Physics} 12:39, 23 January 2009 (UTC)

No, not in this case, there's some very fundamental physics involved. Basically in GR gravity is a pseudo-force, and accelerometers can't read pseudo-forces at all of any kind. They *can* read a force that is equal and opposite to a pseudo-force, but there isn't always one. There's an equivalent argument in newtonian mechanics that relates to the fact that gravity is the same for all parts of the accelerometer (apart from tidal forces).- (User) Wolfkeeper (Talk) 13:15, 23 January 2009 (UTC)

Accelerometers are rarely used by people with a focus general relativity, nor special relativity. The people who use accelerometers are more interested in Newtonian mechanics, and are unlikely have heard of "proper time" (nor have any need for it) There is a simple way to define what is measured as the vector a- g, and I think that definition should enter early on the page, instead of "proper acceleration".

The g cannot really be calibrated away if a system is rotating (like in roller coasters and other common accelerometer applications).Ann-Marie Pendrill (talk) 04:14, 11 November 2018 (UTC)

Accelerometers may not always be used by people that understand general relativity, but nevertheless that is the theory that governs their behavior. They're also not exclusively used on earth, so references to "g" in the definition don't make sense (what's "g" for an accelerometer on a spinning satellite in earth orbit?). Waleswatcher (talk) 13:31, 11 November 2018 (UTC)

Come on! What is the problem?? For a spinning satellite, g is obviously the local acceleration of gravity (of -GM/R^2 if you prefer). The spinning Ann-Marie Pendrill (talk) 16:14, 12 November 2018 (UTC)causes an acceleration (in the Newtonian sense) for objects away from the axis of rotation.

AND the equivalence betweeen inertial and gravitational mass (the weak equivalence principle) holds also for Newtonian mechanics. Ann-Marie Pendrill (talk) 16:11, 12 November 2018 (UTC)

How inertial guidance systems work (section copied from the g-force talk page, where accelerometers are mentioned)

A better paper is here: http://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-696.pdf

Many configurations of inertial guidance systems are discussed, but in general they use 3 devices that are gyros of one sort or another, or which correspond to gyroscopes (such as the small MEMS devices, which actually measure angular acceleration using the Coriolis effect with tiny silicon vibrating wieghts). Add to this, 3 accelerometers, one for each orthogonal axis.

The box diagrams for all of them are of interest and should be examined. Uniformly, the gyro ouputs are first used to project the measured accelerations onto 3 axes, THEN gravitational acceleration is subtracted out of the Z component, then everything is double-integrated to get velocity and finally displacement (with, of course, need to add in absolute velocity and position at each point in the integration to get rid of the constants-of-integration and give definite results).

Interestingly, one of the worst causes of error is gyros which don't get the axes precisely right, causing a small component of the subtracted g-acceleration to be subtracted not from the vertical Z-axis (as is proper), but from the X-Y plane, which is where most of the motion of interest is taking place. Since linear accelerations are integrated TWICE to get distance, small errors in integrating incorrect accelerations in the X-Y plane grow rapidly (quadratically with time) and it doesn't take too long for this bias to screw everything up. Here's the direct quote:

[section] 6.2.3 Propagation of Errors

Errors which arise in the accelerometers propagate through the double integration, as described in Section 4.2. This is the obvious cause of drift in the tracked position. Errors in the angular velocity signals also cause drift in the calculated position, since the rotation matrix C obtained from the attitude algorithm is used to project the acceleration signals into global coordinates. An error in orientation causes an incorrect projection of the acceleration signals onto the global axes. This causes several problems. Firstly, the accelerations of the device are integrated in the wrong direction. Secondly, acceleration due to gravity can no longer be correctly removed.

In the strapdown algorithm 1g is subtracted from the (globally) vertical acceleration signal to remove acceleration due to gravity before the signal is integrated. A tilt error θ will cause a component of the acceleration due to gravity with magnitude g ·sin(θ) to be projected onto the horizontal axes. This causes a residual bias due to gravity with magnitude g · sin(θ) to remain in the globally horizontal acceleration signals. There will also be a residual bias of magnitude g · (1 − cos(θ)) in globally vertical axis, however this is much less severe since for small θ we have cos(θ) → 1 and sin(θ) → 0. Hence the error in position caused by a small tilt error will occur mainly in the global xy-plane.

The propagation of gyroscope errors through to the calculated position is the critical error path in nearly all INS [Inertial Navigation system} systems. In most applications the magnitude of g is much greater than the mean absolute acceleration of the IMU itself. In such cases the critical problem is that a component of the acceleration due to gravity is projected onto the globally horizontal axes. As a concrete example consider a tilt error of just 0.05 [degrees]. This error will cause a component of the acceleration due to gravity with magnitude 0.0086 m/s2 to be projected onto the horizontal axes. This residual bias causes an error in the horizontal position which grows quadratically to 7.7 m after only 30 seconds.

Now, I hope I'm not the only one to notice that this perfectly backs up the other academic paper which has been cited, to the effect that nothing in an inertial guidance system, including the accelerometers in it, can measure gravitational g. It must be removed from the accelerometer readings "by hand" without benefit of any instrument which measures it directly. If there existed an instrument (including the accelerometers themselves) which could do this, they simply would add them to inertial guidance systems, and this growing-bias with time which results from bad orientation and incorrect g-removal angle, would not be the major problem for inertial guidance systems that it is. But it IS a problem. We've discussed the various theoretical reasons why this is not just a difficulty with technology, but rather arrises directly from general relativistic considerations: for any g field point there's an inertial frame in which you can make g vanish. And others in which it is larger or smaller. No instrument, even in theory, can tell you for a planetary g-field which is the "real" and "correct" g, at the same time you're moving up-and-down in it! The trickiness of relativity is that when navigating over a lump of inhomogenous rock, the stuff beneith you imposes on your space-time a complex pattern of "inertial acceleration" along an axis pointing vaguely to the center of the Earth, which changes from place to place, and can't be assessed "on the fly" but only if you physically stop (distance, velocity, and acceleration-wise, with regard to the center of the Earth) to measure it indirection, as you would using a simple scale. But there's no getting away from having to do that. And of course, the scale (like the accelerometer and your butt) is not measuring the g-field directly, but the stress it's causing to the supports of a resting test mass.

for any g field point there's an inertial frame in which you can make g vanish

. Inertial??? Matteosistisette (talk) 10:56, 7 April 2012 (UTC)

inertial frame. That reference frame in which things are weightless and all accelerometers read no acceleration (although they might be physically changing velocity due to gravitational effects, as in free-fall). SBHarris 19:00, 7 April 2012 (UTC)

Anyway, we now have TWO academic papers which make it clear that accelerometers can't be used to direcly measure g, versus some company advertising and some unbacked claims by Greg L. to the contrary. I'm sorry, but Greg L's position is "pwned" by the facts.SBHarris 10:36, 23 January 2009 (UTC)

Acceleration and gravity

(moved from my talk page for all to contribute --Sternutator (talk) 15:14, 29 March 2009 (UTC))

Unfortunately you carelessly messed up the article, which I have corrected, and reintroduced the reference. Accelerometers are actually completely insensitive to gravity since it's an inertial acceleration. Accelerometers (in their simplest terms) measure the distortion of a spring. Inertial accelerations are ones that (in the limit case of a small accelerometer, which normal accelerometers approximate quite well in fact) affect every part of an object the same, and hence they can cause no distortion of the spring, and they read zero; e.g. freefall.

Accelerometers can however react to a reaction to gravity, since that's a non inertial, entirely real force. That's why an accelerometer reads upwards 1g when resting on the surface of the Earth, even though it's not accelerating in the lab frame.- (User) Wolfkeeper (Talk) 18:52, 27 March 2009 (UTC)

Wolfkeeper, my edits were anything but "careless". You need to learn more about the Equivalence principle to understand that there is no distinction between "inertial" acceleration and gravity. The physical effects on all parts of the accelerometer are identical between gravitation and any other force. Further reply at User_talk:Wolfkeeper#Accelerometer --Sternutator (talk) 18:15, 28 March 2009 (UTC)

---

Wolfkeeper,

I feel that for Accelerometer both in the introduction and in the section on physical principles, a physics viewpoint is called for, rather than an engineering one, such as evidently espoused by Eschbach. I also suggest we keep the language clear. This means avoidance of adjectives that will confuse an un-initiated reader, such as in your formulation:

accelerometers cannot measure inertial accelerations, but only non inertial ones caused by mechanical forces.

I puzzle what this sentence means. Accelerometers measure any acceleration, gravitational (as in tilt sensors) or otherwise.

Careful here, a proper tilt sensor uses gyroscopes not accelerometers, and requires a known reference to remove drift. Using accelerometers for tilt sensing fails in certain key situations such as in aircraft where the perceived gravity vector can do pretty much whatever it wants.- (User) Wolfkeeper (Talk) 23:50, 28 March 2009 (UTC)

I understand your point. I can certainly fool the sensor in my phone by sliding it horizontally. That said, the accelerometer in smartphones is a simple 3-axis MEMS device, as our article says further down, right to the IC type.

Conceptually, an accelerometer having a test mass uses Newton's law F=m a to link mass and force, regardless of the force's origin. As I wrote, acceleration and gravity are indistinguishable by the equivalence principle, beautifully illustrated in the elevator Gedanken experiment, notwithstanding cranks. An accelerometer can and does measure gravity - cf. gravimeter. I agree with your removal of my remark on vertical integration in an INS, but I feel it does have its place, namely in the navigation section.

You also reverted the text in Accelerometer#Consumer_Electronics to its incoherent flow: first smartphones, the laptops, notebooks (?), and smartphones again. Meandering is a common Wikipedia ailment. I feel less passionate about that section than the one on physics, which is my primary qualification.

I moved your reference to analog.com into the Accelerometer#Structure section. --Sternutator (talk) 17:57, 28 March 2009 (UTC)

OK, so you claim to have a PhD in physics, so you should understand the following, but you may not, depending on what part of physics your PhD was on. It is factually true that accelerometers are completely incapable of measuring inertial accelerations. You will recall that an inertial acceleration is one due to doing physics in a non inertial frame of reference. It is a principle of general relativity that gravity is an inertial acceleration- the geodesic of any object is equivalent to unaccelerated motion; therefore the accelerometer reads zero due to gravity. If you haven't already you need to read the talk page of the accelerometer page and it would be a good idea to read the reference as well; and the talk page for g-force also contains essentially the same discussion only longer.- (User) Wolfkeeper (Talk) 23:45, 28 March 2009 (UTC)

Accelerometers can also be understood in Newtonian mechanics where forces are vectors, accelerometers are linear systems. The force on the spring of an accelerometer due to gravity is always zero; since all parts of the accelerometer fall/accelerate at the same rate due to gravity. When external mechanical forces act as well, such as when the casing is resting on the ground, then the force on the spring due to that acceleration is the sum of that due to gravity (always zero) and the acceleration due to resting on the ground (1 g upwards), giving 1g upwards; which is what an accelerometer reads in that situation.- (User) Wolfkeeper (Talk) 23:45, 28 March 2009 (UTC)

This is unfortunately all rather counterintuitive, and lots of people are getting it wrong and editing on this misunderstanding.- (User) Wolfkeeper (Talk) 23:45, 28 March 2009 (UTC)

Another way (entirely equivalent way) to look at this is that the accelerometer measures the acceleration relative to freefall.- (User) Wolfkeeper (Talk) 23:45, 28 March 2009 (UTC)

It is true that if you put an accelerometer (let's say a test mass on a spring) in space, and a large mass happens to pass by, the whole device will free-fall along an interesting path, all the while reading zero. You could say that one cannot measure gravity, fair enough.

Now put the device on the surface of that large mass, and call the surface reference frame S. You thus apply a force on the casing countering gravity (or inertial force, or fictitious force, ...) and stopping the freefall. Of course, the test mass continues to experience the gravity and is deflected, until countered in turn by the increasing restoring spring force. We measure the displacement d, and knowing the test mass m and force constant k conclude g = k d / m. This is a measurement of weight (gravitational acceleration) by a compensation method. What is being measured and indicated? Why not call this "measure gravity"?

Inherent in this method is of course the assumption that the device remain stationary in S. If the device is accelerated by acceleration a in S, it will measure a + g, and that's why I formulated "acceleration and gravity". This includes fictitious forces like that in an accelerating car, and centripetal and Coriolis forces in rotating frames. This is indeed all rather deep, but a short version is more to the point.

My Britannica (Copyright 1990, and the same in the current online edition) formulates under accelerometer:

Acceleration cannot be measured directly. An accelerometer, therefore, measures the force exerted by restraints that are placed on a reference mass to hold its position fixed in an accelerating body. Acceleration is computed using the relationship between restraint force and acceleration given by Newton’s second law: force = mass × acceleration.

The engineering reference you cite (analog.com) also states "Accelerometer measures the static gravity field", and also "Changing the tilt (along the sensitive axis) changes acceleration vector".

I undertook to rework the various bits of the article, trying to accommodate your viewpoint. I urge you to review it carefully before doing a kneejerk revert. The effects of acceleration and gravity are identical, by Einstein. You cannot claim to be able to measure one but not the other. I agree that gravity cannot be measured directly, but neither can any other acceleration. Both are measured indirectly, and cannot be distinguished.

No, careful here, gravity and acceleration are not the same, per Einstein, that's the slightly over-simplified version. What he actually said was:

"We shall therefore assume the complete physical equivalence of a gravitational field and a corresponding acceleration of the reference system."

It's the reference system that is accelerated by gravity and that is why an accelerometer is not able to detect it; it's not accelerating!!!! Gravity is an inertial acceleration in General Relativity; bodies follow their geodesic.- (User) Wolfkeeper (Talk) 15:59, 29 March 2009 (UTC)

I'm sorry, you have made the introduction incorrect, and removed a reliable source; therefore your edits are unacceptable. Please do not edit articles where you don't understand the material, and particularly do not remove reliable sources just because they contradict your POV.- (User) Wolfkeeper (Talk) 15:50, 29 March 2009 (UTC)

I suggest to drop Eschbach, as that has been critized elsewhere by User:Greg L. I also argue that the analog.com reference does not belong in the introduction for reasons of style (current practice notwithstanding).

Wolfkeeper, I see this is not the first time you've ruffled feathers. For some reason, first encounters on Wikipedia are always much rougher than a friendly neighbourhood coffee shop ... --Sternutator (talk) 08:15, 29 March 2009 (UTC)

When the dust settled it was generally agreed that Greg_L was wrong on this; and from the wikipedia's point of view, he was completely unable to substantiate his position with reliable sources. Greg_L has since been suspended at least once for these kinds of activities and edit warring; I do not think you should be following his lead.- (User) Wolfkeeper (Talk) 15:50, 29 March 2009 (UTC)

I'm sorry, but you've fallen into the common trap here. Please read the references and check the relevant physics textbooks before 'correcting' this article again.- (User) Wolfkeeper (Talk) 15:50, 29 March 2009 (UTC)

You can measure non inertial accelerations just fine with an accelerometer. You can infer inertial accelerations in special situations, for example, where you know that a non inertial acceleration is equal and opposite to an inertial acceleration, but this cannot always be done.- (User) Wolfkeeper (Talk) 16:05, 29 March 2009 (UTC)

Wolfkeeper, if it is the reference you wish to keep, so be it. We agree that acceleration is not measured directly, only indirectly by means of forces etc. Dare I say that the reference system for an accelerometer is its own casing? The test mass responds to any deviation of this frame from freefall, be it from gravity or another external or fictitious force. That's all there is to it, and there is no need to artificially distinguish the accelerations or introduce additional concepts. You insist on:

An accelerometer is a device that measures non-gravitational accelerations[1]. These are accelerations produced by mechanically accelerating the accelerometer via its casing.

This is incomplete at best. When my phone is sitting on my coffee table it tells me there is a 1 g acceleration upwards, even though it not "mechanically accelerated". I would observe the very same thing if the phone would sit in a rocket that accelerates at 1 g. Think about it. This is why I edited this article in the first place.

Unfortunately, you're still not getting it. For example, a rocket that weighs 1 tonne and is supplying a thrust of 9810N gives the same reading on the accelerometer it carries whether it is standing on its tail next to the Earth, or whether it is thrusting in deep space, far away from any gravity source. In neither case does the gravitational acceleration affect the reading on the accelerometer in any way.- (User) Wolfkeeper (Talk) 18:07, 29 March 2009 (UTC)

Methinks your insistence is due a question of instrument design – at which point is gravity subtracted, and what is reported to the user? This consideration belongs in the Applications section, not in Physical principles. Eshbach and you wish to exclude gravity by calibration and report back only the reduced result; observation prior to reduction tells us otherwise.

I'm sorry that you don't understand the physical principles that permit accelerometers to operate, but I cannot allow you to write your unreferenced, incorrect position in the wikipedia as if it were fact.- (User) Wolfkeeper (Talk) 18:07, 29 March 2009 (UTC)

I made non-trivial contributions, e.g. in the INS section, and corrected typos. You should not revert indiscriminately. Now, please go back and correct the introduction, and merge my edits in Accelerometer#Physical_principles and Accelerometer#Navigation back in. --Sternutator (talk) 17:07, 29 March 2009 (UTC)

Your non-trivial contributions involved removing referenced material, and adding in unreferenced material that happens to be wrong. In the wikipedia unreferenced material can be removed at any time.- (User) Wolfkeeper (Talk) 18:07, 29 March 2009 (UTC)

Wolfkeeper, please stop reciting the rules or and start thinking. Explain the vertical offset of an accelerometer at rest on earth and tell me what is being measured, if not gravitational acceleration? What would be measured in your rocket example (a) at 9810N thrust, (b) at half thrust, and (c) at free fall? --Sternutator (talk) 21:19, 29 March 2009 (UTC)

FFS, for the last time, it is not the gravitational acceleration that is being measured. Consider standing at the top of a tower with an accelerometer. It reads 1g. Now imagine a rocket hovering beside, it also reads 1g. Is this 1g the gravity? No; imagine if you give the rocket a hard shove vertically and it gains altitude; the gravity goes down as it does so, but the accelerometer is constant at each and every point! That proves unequivocally that accelerometers do NOT measure gravity!!!!!!!!!! They measure something else. In the case of them resting on the ground, they measure the reactive force to gravity. In the rocket they measure the thrust of the rocket. In each and every case the accelerometer does not; CANNOT measure gravity.- (User) Wolfkeeper (Talk) 21:53, 29 March 2009 (UTC)

What do you mean by "the gravity goes down as it does so"? During the "shove", the experienced gravity increases (same as in an elevator starting to rise), then goes back to 1g. Never mind. You cannot claim to measure "mechanical" accelerations but not gravity. Both are equivalent and yes, can only be measured via reaction forces. Those forces are trivially translated to acceleration, and so for all practical and measurement theoretical reasons, the accelerometer measures external and gravitational acceleration. Saying they "measure the reactive force to gravity" is decidedly awkward. My phone's accelerometer measures roughly 9.8 m/s² at rest. What is this but a measure of gravity? When I'm in an elevator that's starting to rise, the value briefly increases and then falls back. By rigorous interpretation of the equivalence principle, the gravity that I and the device experience has momentarily increased.

No, the rigorous interpretation of the equivalence principle is that gravity is an acceleration of the reference frame, NOT an acceleration of the object. If you drop an accelerometer, what does it read? Zero. If it read gravity it would be non zero, since no other forces act. This is self-evident, and I don't care how awkward it reads, the wikipedia is supposed to contain referenced opinions, not simplified statements that 'read nice' or something. The wikipedia is verifiability over truth. You do not have a reference that contradicts the Esbach one; and even if you did, you would have to include both POVs.- (User) Wolfkeeper (Talk) 00:50, 30 March 2009 (UTC)

Answer below in #Review of Approach and Sources --Sternutator (talk) 07:33, 31 March 2009 (UTC)

Moreover, if an accelerometer cannot measure gravity, as you claim, why are gravimeters described as "type of accelerometer"? Your insistence is puzzling, and your phrasing and shouting are telling. --Sternutator (talk) 22:43, 29 March 2009 (UTC)

They are; they are a type of accelerometer that when used in a particular way allows you to infer the gravitational field, but they do not directly measure it; and they do not work at all when they are not attached to the ground.- (User) Wolfkeeper (Talk) 00:50, 30 March 2009 (UTC)

Eshbach reference superseded

User:Wolfkeeper insists on Eshbach (Eshbach's Handbook of Engineering Fundamentals By Ovid W. Eshbach, Byron pg 9 - [year and edition missing]) as authority. References can be wrong, be they on the net or in print. Case in point, the most recent 5th edition (2009) contains very little on the subject, at most section "4.8 Moving Axes"; the equivalence principle is not discussed, and does not appear in the index. This makes this reference no longer relevant to this article. Here is a more pertinent reference:

Because of the principle of equivalence, there is no local way of telling whether a gravitational force is acting or whether what 'feels' like a gravitational force may just be the effect of an acceleration. [...] This was Einstein's profoundly novel view: regard the inertial motions as being those motions that particles take when the total of non-gravitational forces acting upon them is zero, so they must be falling freely with the gravitational field (so the effective gravitational force is also reduced to zero). [...] On the other hand, someone just standing on the ground is not executing an inertial motion, in the Einsteinian scheme, because standing still in a gravitational field is not a free-fall motion. To Newton, that would have counted as inertial, because 'the state of rest' must always count as 'inertial' in the Newtonian scheme. The gravitational force acting on the person is compensated by the upward force exerted by the ground, but they are not separately zero as Einstein requires.
– Roger Penrose, "The Road to Reality", Knopf, New York, 2005/Jonathan Cape 2004. Section 17.4, pp. 393-394 in the Knopf edition.

User:Wolfkeeper, please think before you revert (again). If you still feel there is something more to explain, please do so constructively. --Sternutator (talk) 21:19, 29 March 2009 (UTC)

Accelerometers measure the felt acceleration, gravity is NOT ever felt. Never. Ever. What you feel is the reaction to gravity. An accelerometer measures that (sometimes); NOT gravity. If you sit in a chair, you feel the chair pushing you up NOT gravity pushing you down. In a rocket you feel the rocket pushing you up, not gravity pulling you down. The accelerometer measures the self same acceleration you feel; that's why it tells you there is an acceleration UPWARDS when it rests on the ground. If it measured gravity it would read ZERO, since there is both the reaction force and gravity, which cancel. It does NOT read zero. That is why it does not read zero.

The reference that you keep removing says that also, that accelerometers DO NOT READ GRAVITATIONAL ACCELERATIONS.- (User) Wolfkeeper (Talk) 21:53, 29 March 2009 (UTC)

Penrose above does not mention accelerometers, and, frankly, he's dumbing down slightly for a lay audience.- (User) Wolfkeeper (Talk) 21:58, 29 March 2009 (UTC)

Your reference does not hold up - it has been updated by the publisher and can no longer be considered reliable. Yet you keep inserting it back. This is not acceptable. Please, be more constructive. --Sternutator (talk) 22:43, 29 March 2009 (UTC)

It's still a reliable source. It has been published and peer reviewed. You have not found a reliable source that says the opposite.- (User) Wolfkeeper (Talk) 00:26, 30 March 2009 (UTC)

In addition please read this thread Talk:G-force#How_inertial_guidance_systems_work_.28a_better_paper.29 which covers that and several other sources that all say the same thing. The equation for the acceleration read by an accelerometer goes as a-g, where a is the acceleration vector in the lab frame and g is the gravitational acceleration in the lab frame. You will see that this is equivalent to the acceleration relative to free fall, as when the acceleration and the g vector are the same, it gives a zero reading.- (User) Wolfkeeper (Talk) 00:26, 30 March 2009 (UTC)

See: [2] which is another standard textbook, which says the self-same thing.

This is not a reliable source, but it explains it very well: [3]

Review of Approach and Sources

[Outdented and sectioned for easier editing]

Wolfkeeper, you wrote above:

I don't care how awkward it reads, the wikipedia is supposed to contain
referenced opinions, not simplified statements that 'read nice' or something.
The wikipedia is verifiability over truth.

Ouch! I may be new here, but I would expect both good readability and truth from an encyclopedia – and of course, its information should be verifiable.

First, regarding readability: You place heavy emphasis on the distinction between inertial and non-inertial accelerations, as I already pointed out. Your article version states:

However, accelerometers cannot measure inertial accelerations,
but only non inertial ones caused by mechanical forces.

You simply cannot leave out a clear and lucid definition or reference to either term. Doing so hinders understanding and increases wikipedia's bad rep in the sciences field. I suggest to remove or rephrase this.

Second, a side note to your comment on

gravity is an acceleration of the reference frame

That's a bit harsh. The principle of equivalence concerns itself with "the equivalence, in terms of local experiments, of gravitational forces and reactions to an accelerated noninertial frame of reference ... and the equivalence between inertial frames of reference and local freely falling frames of reference" (Encyclopedia Britannica (C) 1990, vol. 26, p 534, "Relativity").

It's not harsh, it's the same thing. That this is an accelerated reference frame means that gravity is non detectable, because according to GR the accelerometer isn't 'really' accelerating under gravity, it's the reference frame accelerating. Incidentally, the same is true for other pseudoforces like coriolis acceleration and centrifugal acceleration and euler acceleration; accelerometers don't register them at all. (Tidal forces are detectable though.)- (User) Wolfkeeper (Talk) 15:04, 31 March 2009 (UTC)

Third, here is what Eshbach actually says (and what redeems this source):

... an acceleromer measures the nongravitational specific force, output = f = a – g, and cannot distinguish between the inertial acceleration, a, and gravitational acceleration, g.
– 4th edition, 1990, "Accelerometers", page 7-29.

What Eshbach/Byron says here is more than that an accelerometer does not measure gravitational acceleration [at all?]. Such a limited view is my key criticism. Further, he calls f (the quantity measured by an accelerometer) "nongravitational" acceleration, not "non inertial ones caused by mechanical forces" as you state further down in Physical principles (inconsistent terminology is another major Wikipedia ailment).

Finally, to work this into our text here: of course specific force is antiquated even by Wikipedia's standards, so, in more appropriate language, I would suggest the following:

... measures inertial acceleration minus the local gravitational acceleration, where inertial acceleration is understood in the Newtonian sense of acceleration with respect to a fixed reference frame.

That is, in the earth frame, with a z-axis vertically up and g straight down, g = – g e_z, then, in Eshbach's notation, we are interested in a, but measure f = a – g. This comes out correctly to:

• f = +g e_z (straight up) for the accelerometer at rest (a = 0 )
• f = 0 for free fall (a = g ).

The sublime subtlety here is that the definitions of inertial frame of reference in the Newtonian sense (as stated) and Einsteinian sense (free fall) differ crucially in the aspect of which acceleration we are interested in on earth, especially since the earth's surface is often considered a reasonable approximation of an inertial frame. --Sternutator (talk) 07:33, 31 March 2009 (UTC)

Now you're more or less getting it, in both cases, accelerometers fail to measure gravity- that's a reasonably deep reason why you're subtracting it off the actual acceleration in the above equation; although the GR position that it's not actually accelerating due to gravity is even deeper ;-)- (User) Wolfkeeper (Talk) 15:04, 31 March 2009 (UTC)

We should also note that when the article referenced above says the surface of the Earth is "almost an inertial frame" it is only talking about behavior in 2-D (like the surface of an air hockey table). In the third dimension, there is no inertial behavior at all-- it takes a force just to keep things still on Earth's surface. That's an accelerated frame, not an inertial frame. In truly inertial frames, also called "free fall frames" things are weightless, and they follow Newton's laws in all axes/dimensions. It's the inside of a falling elevator or cabin of an orbiting space shuttle that is "almost an inertial frame," certainly not the surface of the Earth! So that very statement fouls up the meaning we're getting at, with accelerometers. I wish the author hadn't written it, because he was only thinking about behavior on a surface.

Accelerometers, in the end, measure "weight per unit test mass." Also called the specific force or proper acceleration. That is all. Relativity need not be involved, for the preceding statement is Newtonian also. But notice that this means accelerometers don't (cannot!) measure the gravitational field per se. They don't "see" it, and humans don't "feel" it. All accelerometers "see" is proper accelerations away from the local inertial frame, which they sense by means of mechanical forces (due to weight), or from physical deviation of masses from an inertial path (due to proper acceleration of the accelerometer housing, again due to a mechanical force). The only time accelerometers are aware of gravity per se, is when they are allowed to test the difference between the "inertiality" of frames at different places in space, so that they can test for gradients in space, in what counts as an "inertial" frame for each point. That is a gradiometer-- it consists of a bunch of accelerometers all separated by little distances from each other (see GOCE for an example). However, a gradiometer made from accelerometers only measures tides (the gravitational gradient), not the absolute gravity value anywhere. The absolute gravity at a point in an inertial frame is unmeasurable. Newton would say that's because there's no way to do it. Einstein would say the REASON there's no way to to it, is because in that frame gravity actually doesn't exist! Which is to say, that at a carefully chosen single point in an inertial frame constructed just for that point, space-time is flat. That's what is meant by "locally flat." Right at that point (this doesn't apply to "near the point"), there is no gravitation field and no proper acceleration, so there's nothing for an accelerometer to measure: you can't measure an absolute gravitational field at a point, anymore than its possible to measure an absolute velocity. If you feel nothing and measure nothing, you're quite free to say there IS nothing, and challenge your opponents to prove you wrong (just as with velocity). That's the equivalence principle. SBHarris 21:47, 16 January 2011 (UTC)

acceleration ??

An accelerometer doesn't measure acceleration in the sense that most people think about it (in respect to earths reference frame). I edited the acceleration to read proper acceleration hopefully this will lead to less confusion. Cheers! —Preceding unsigned comment added by 194.171.252.108 (talk) 10:28, 26 October 2009 (UTC)

I really think that the use of "proper acceleration" is confusing for most people interested in accelerometers, who tend to work in situations where Newtonian mechanics is perfectly adequate.Ann-Marie Pendrill (talk) 09:43, 11 November 2018 (UTC)

physical principles

This section needs to be re-worded. "inertial acceleration" is not a standard term in physics and the reference given is to an engineering handbook, which does not suffice and is not appropriate in a section on fundamental physics. More seriously, the description as written is imprecise to the point of being wrong. There is no such thing as a "fixed reference frame". The correct way to talk about this is to refer to a locally inertial frame. Such frames are guaranteed to exist by the equivalence principle, and an accelerometer measures acceleration with respect to that frame (i.e., proper acceleration). The earth's surface is undergoing approximately constant acceleration upwards in a locally inertial frame, which is why an accelerometer measures it. If you want to know the acceleration with respect to that frame then you do indeed need to subtract that contribution, but the result cannot be thought of as acceleration "due to motion alone", period. Such statements are meaningless without specifying what the motion is respect to (in this case it's the earth's surface).

I'll leave this comment up for a bit and then edit if there are no objections. I'm going to remove that reference and all statements about "inertial acceleration" and "fixed reference frames", and replace them with standard physics terminology. Waleswatcher (talk) 14:42, 29 November 2009 (UTC)

As you can see from prior discussion up here this is sensitive matter, and perceptions on physics vs. engineering differ. Note that the term inertial acceleration is referenced as such in the scientific literature. The crux of this article is a clash between Newtonian and relativistic definitions of "inertial reference frame". I agree that a better formulation can be found, but the sentence

Put another way, at any point in spacetime the equivalence principle guarantees the existence of a locally inertial frame, and an accelerometer measures the acceleration relative to that frame.

is not very helpful to the uninitiated reader. --Sternutator (talk) 06:20, 8 December 2009 (UTC)

Inertial acceleration is not a standard term in physics. The link you gave is the first time I've ever seen it used in a technical context, and I've read thousands of papers on general relativity and gravity. Proper acceleration is the only standard term. I do agree with you that that sentence I added isn't very clear to non-experts, but the stuff about "Newtonian reference frame" or "fixed reference frame' is just not correct at all. It's not the right way to think about this, as we've known since the time of Einstein.

So I'm going to have to revert this. I suggest we work together to explain the reading of the accelerometer on earth's surface in the correct way as understood by Einstein and discussed in other wiki articles. That way is to find an inertial frame in which some point on the earth's surface is instantaneously at rest, and then notice that something at rest on the surface at that point is accelerating upwards at g. Waleswatcher (talk) 14:46, 10 December 2009 (UTC)

By the way, if we were discussing, say, the technical operation of some specific accelerometer design, an engineering reference would be completely appropriate. But when discussing the fundamental physics issue of why properly functioning accelerometers read what they read in free fall versus on earth, the only appropriate references are mainstream physics books and papers discussing the equivalence principle, because that is our best modern understanding of that question. It is one of the most important and fundamental principles in modern physics. There are tons of good physics references, I'll dig one up and add it when I have time.Waleswatcher (talk) 15:05, 10 December 2009 (UTC)

Accelerometers are rarely used by people with a focus general relativity, nor special relativity. The people who use accelerometers are more interested in Newtonian mechanics, may never have heard of "proper acceleration" (nor have any need for it) There is a simple way to define it as the vector a-g, and I think that definition should enter early on the page, possibly with a note linking it to "proper acceleration". I also have not encountered "inertial acceleration" in the context of people who actually USE accelerometers. (I have written many papers on interpreting accelerometer data from roller coasters and other amusement rides, where relativity is not the prime concern. I have also authored many papers in atomic physics, with relativistic many-body calculations)Ann-Marie Pendrill (talk) 04:21, 11 November 2018 (UTC)

Relativity digressions are better than crack

Is relativity really necessary to describe a ball on a spring, or are we just showing off? Honestly, if our lives depended on people being able to understand how accelerometers work, would any of this be here? This article was written for wikipedia editors, not wikipedia readers.

The word "spacetime" is jargon and should be avoided. The validity of the concept was challenged last month on the front page of Scientific American splitting-time-from-space. At any rate, the operation of an accelerometer does not hinge on the existence of a local Minkowski space.

The validity of the equivalence principal is determined by experiment, it is falsifiable. There are many viable theories in which the equivalence principal is false. Currently tens of millions of dollars are spent searching for violations. [4]

"the reason for the appearance of a gravitational offset is Einstein's equivalence principle"
Falsifiable theories are NOT the cause of physically measurable phenomena.

"at any point in spacetime the equivalence principle guarantees the existence of a local inertial frame"
Falsifiable theories about falsifiable concepts cannot provide ontological certainties.

Who challenged the existence of inertial frames anyway? I recognize this cliche as standard relativity boilerplate:

"in Newtonian mechanics, the first law guarantees the existence of a inertial frame"
"the definition of E3 (r) guarantees the existence of inertial frames"
"the ricci tensor guarantees the existence of an instantaneously inertial frame"
"both theories [GR & Newtonian} assume the existence of inertial frames"
"one ingredient of special relativity is the existence of inertial frames"
"special relativity presupposes the existence of inertial frames"

The operation of an accelerometer does not depend on the validity of any scientific theory. Why not avoid all the drama and describe them in simple terms claiming only experimental facts?

---

Now I want to get in on the flame war! New SQUID gravity gradiometers can easily distinguish acceleration from gravity. The equivalence principal, like Newton's law of gravitation, only applies to point masses. All real world objects have spatial extent and therefore will be influenced by tidal forces. Since uniform gravitational fields do not exist in reality, there will always be some measurable tidal influence for a sufficiently sensitive measuring device.

Accelerometers measure specific force, not acceleration. Forces and specific forces are the same in all frames, accelerations are frame dependent. --Norbeck (talk) 02:25, 28 January 2010 (UTC)

Sigh. G-meter accelerometers are calibrated in 'g's. A g is 9.8066 m/s^2 not 9.8066 N/kg. Note the difference in units. Accelerometers do not measure specific force, they measure 'proper acceleration', which is the 'felt' acceleration. I don't particularly care whether you do your calculations in Newtonian or GR, but GR is simpler; and you definitely can't mix them; the frames of reference are differently defined.- Wolfkeeper 02:48, 28 January 2010 (UTC)

G-meter accelerometers are calibrated in 'g's. A g is 9.8066 m/s^2 not 9.8066 N/kg. Well N=kg*m/s^2, hence 9.8066m/s^2=9.8066N/kg Matteosistisette (talk) 19:27, 14 January 2011 (UTC)

Yep. These are acceleration units, and thus a "specific force" (full name: mass-specific force = force/mass) is actually an acceleration. But it's a proper acceleration, not a coordinate acceleration. Which is why it's perfectly-well and correctly measured in "g"s. It is a "felt-acceleration," and (when multiplied by mass) it gives the magnitude of the weight (which also, is obviously "felt" mechanically). Read the wiki on specific force. It is an older term but an exact synonym for proper acceleration. SBHarris 02:35, 15 January 2011 (UTC)

It's not possible even in principle to directly measure gravity with an accelerometer; you can however, measure the reactive acceleration, and in some circumstances that allows you to infer what the gravitational acceleration must be.- Wolfkeeper 02:48, 28 January 2010 (UTC)

"It's the reference system that is accelerated by gravity", Gravity can not accelerate abstractions in our minds. Besides if the equivalence principal were to be falsified we wouldn't be guaranteed these references systems even exist!

Others have choosen, however unwisely, to follow Ockham's razor and avoid presuming the existence of things that do not. However, this is probably the best opportunity we'll ever have to compel strangers to visualize a book on a table as geodesic Fermi-Walker triad in instantly comoving frame.

The situation is grim:

Results 1 - 10 of about 1,450 for accelerometer "proper acceleration" -wikipedia.

Results 1 - 10 of about 38,400 for accelerometer "specific force" -wikipedia

Better than 26 to 1.

Results 1 - 10 of about 112 for accelerometer "proper acceleration" journal physics

Results 1 - 9 of about 723 for accelerometer "specific force" journal physics

Still 7 to 1. Sigh. --Norbeck (talk) 10:53, 29 January 2010 (UTC)

I don't see your point. The specific force is what causes the proper acceleration of the test mass. Measure the one and you have the other: a(proper) = F/m(test) (definition of specific force). They are literally the same thing. SBHarris 11:30, 31 January 2010 (UTC)

Non-gravitational forces on proof masses

"Although the proof masses are shielded from non-gravitational forces by the spacecraft, cosmic rays and solar flare particles can significantly charge them, leading to electrostatic forces that can corrupt the measurements by the inertial sensor"

An accelerometer in free fall near an electrostatically charged object will accelerate towards the object due to electrostatic induction. What the accelerometer reads during this acceleration depends on the electrostatic induction coefficients of the proof mass and the casing. With proper shielding the accelerometer will approximately measure the proper acceleration. In practice, the proof mass is acted upon by several forces.

• Gravitational attraction between the proof mass and the casing
• Tidal forces
• Magnetic torques
• Electrostatic gradients caused by the photoelectric effect
• Convection due to thermal gradients (the proof mass is not in a pure vacuum)
• Interal radiation pressure
• Forces due radioactive decay
• Forces due to collisions with cosmic rays
• Electric charging due to cosmic rays

The proof mass does not follow a geodesic. Accelerometers do not measure proper acceleration.

There are two types of proof-mass accelerometers. Free floating and fixed. For free floating proof masses (used in zero gee environments) the measured quantity is the relative position between the proof mass and the casing. For fixed proof masses, the measured quantity is the force between the proof mass and the casing. The measured force is then divided by the known value of the proof mass. The terms "force intensity" and "specific force" are DEFINED as quantities of this type.

There is a reason why 26 out of 27 people say accelerometers measure specific force, and it's not because they don't know relativity. To the contrary, it is because they do.

By the way, proper acceleration is not the acceleration relative to a "local inertial frame". It is relative to an "instantaneously co-moving inertial frame". The word local is redundant in this context. Stick to cut & paste, it will never do you wrong.

LISA is a gravity wave detector: The Laser Interferometer Space Antenna, LISA ... the detection band is limited by the accelerometer (specific force) noise of the proof masses. Here is a powerpoint presentation about Gravity Probe B. see page 19. Both articles correctly use the term specific force, just as we were taught in college, just as the dictionary says, just as the manufacturers say, just as the relativists say, just as engineers say, just as this article used to say before it was hijacked and injected with confabulated unsourced pseudorelativity doubletalk.

Wikipedia The Free Encyclopedia

An accelerometer a device for measuring specific external force. Specific external force is the sum total of external forces acting on an object divided by the mass.

Websters Dictionary:

weigh : to experience a specific force due to gravity —Preceding unsigned comment added by Norbeck (talk • contribs) 14:50, 30 January 2010 (UTC)

Neither of those systems are accelerometers. And even if they were, I'm not saying that describing an acceleration as a specific force is necessarily incorrect, I'm saying that that's not what accelerometers are usually said to measure, and that's not the units are used either.- Wolfkeeper 15:15, 30 January 2010 (UTC)

Are you claiming that LISA and Gravity Probe B do not have accelerometers? [y/n] --Norbeck (talk) 10:49, 31 January 2010 (UTC)

They doubtless do. But that's faulty logic. You have a skeleton. That doesn't mean you are a skeleton. You're confusing the different levels of the system. Just because Gravity Probe B has an accelerometer or two, doesn't mean it is an accelerometer.- Wolfkeeper 14:25, 31 January 2010 (UTC)

Since proof masses can never truly be in free fall, they cannot be used to determine the proper acceleration. The truth is there are not many people who claim that accelerometers measure proper acceleration. There is only a small minority of smug and condescending pseudorelativists who forgot to consider non-gravitational forces before they hijacked the definition.

Besides non-gravitational pertubations, there is another really important reason why there are no NASA citations to support the claim that accelerometers measure proper acceleration, even in the idealized situation where there are no non-gravitational forces on the proof mass. There is a very specific reason why professional relativists who work on LISA and other relativity programs do not refer to the output of an accelerometer as proper acceleration. You're gonna kick yourself. --Norbeck (talk) 15:33, 31 January 2010 (UTC)

Ideal versus real accelerometers

In reality, accelerometers would barely behave in the way described in the "Physical properties" section. A parabolic flight is a transient event and the signal of an actual accelerometer would then oscillate or slowly converge towards 0g, without ever truly getting there. Also, many real accelerometers behave as high-pass filters and thus are unable to capture the earth acceleration (which is essentially a static event) unless they experience a change in orientation. In other words, the main purpose of the "accelerometers" presented in this section seems to enable some thought experiments (which may be worthwhile to get a better grasp of the laws of dynamics), but they do little to convey the notion that accelerometers are measuring instruments with specific physical limitations. I submit that more focus should be placed on the working principles of accelerometers as mechanisms, and less emphasis be given to the meanings of gravitation and acceleration as physical concepts (there should be plenty of room in other Wikipedia entries to address these topics). After all, it be would strange if an article on watches were to expand on the relativistic theory of time. Decan tulahi (talk) 08:32, 3 June 2010 (UTC)

The slow convergence bit is wrong- the convergence is typically an exponential decay towards the correct value, nearly always with a low time constant. Convergence happens very rapidly, just 10 time constants gets you within a few ten millionths of the right answer, and another ten within a couple of billionths.- Wolfkeeper 16:29, 3 June 2010 (UTC)

If an accelerometer were to have a time constant of 1 second (due to a large proof mass and weak spring element), it would still miss the mark for the first few seconds of the flight. Now, I do agree that the rate of convergence of a standard accelerometer would be more than sufficient in that case. Rather, the problem I see with the examples is the fact that they mislead the reader into thinking that accelerometers must have a large bandwidth (unlike accelerometers used in seismology) and that they can always detect a static acceleration (unlike piezoelectric accelerometers). Decan tulahi (talk) 22:04, 3 June 2010 (UTC)

Another confusing statement on the first line of the "Physical properties" section: "Proper acceleration ... is the acceleration felt by people and objects". Yet, the human body is not ideal either and is basically insensitive to accelerations above a few hundred Hertz (see ISO standard 2631-1). Decan tulahi (talk) 01:23, 4 June 2010 (UTC)

I've added some language in the lede that notes that real accelerometers are not perfectly sensitive to changes in proper acceleration on small time scales. And of course, neither is the human body. You'll have to add some amplification of that for real instruments, in the article body. But at least the qualifier is now in. SBHarris 19:42, 17 January 2011 (UTC)

Consumer electronics: position sensing = accelerometry?

I actually came to this page trying to find out about position sensors in consumer electronics, like smartphones. Why is this an accelerometer? What acceleration does it measure? Is a mercury switch, e.g., an accelerometer? If I turn off my iPad (so its internal circuitry is incapable of measuring anything at all), rotate it 90% in the vertical plane, and turn it back on again to find the display has been reoriented, was acceleration measured? My ThinkPad has an accelerometer to decide if it's falling off of a table. If the computer's turned off, this accelerometer does nothing if I drop the computer, and in this sense the notion of measuring acceleration makes sense to me. In terms of determining the position of a display, not so much.

I don't mean to sound obtuse, but I think that it is in this context that "accelerometer" has entered the vernacular. Maybe somebody who knows something could explain it.

Thank you. —Preceding unsigned comment added by 24.87.65.232 (talk) 03:14, 26 August 2010 (UTC)

Accelerometers don't measure dv/dt, the definition of acceleration we all learned in physics. That should help with your conundrum. Accelerometers measure proper acceleration, which is associated with whatever causes weight in a test mass. So they see orientation, since that changes direction of weight-force. They similarly measure g-forces and changes in g-forces and (of course) directions of g-forces. You can see that all those things change when you change orientation of your device, although you generally have to have several acceleromaters looking at several physical axes, to tell directions of the "weight-forces" that are sensed by the test masses in the device. When something is dropped, the device sees all its test masses go weightless, and knows that it's falling. So long as it has a memory, it can tell differences in weights, and knows that since these are changing, that either it's being physically accelerated, or else somebody has turned off, or changed, the gravity. It assumes the latter. If you turned it off and took it out between stars and turned it on again, it would say "OMG, I'm falling! (and so would your inner ears). But in both case, the "device" would be wrong.SBHarris 19:40, 17 January 2011 (UTC)

I would suggest that the orientation can be established easily since an accelerometer always produces a DC output voltage; change the orientation relative to the axis of acceleration and the DC voltage will also change. Most MEMS accels in phones,computers,etc.. are tri-axial so change the orientation and you change the relationship between the DC outputs of the 3 axes. The relationship (as opposed to the absolute values) between the DC output voltages can be used to establish the orientation. In industrial measurement applications, the DC is usually ignored since we're only interested in the AC signal (ie. freq above 0Hz). The signal from most accelerometers is conditioned and filtered somewhere in the downstream signal chain (typically above 0.1Hz).

AcoustiMax (talk) 21:27, 5 February 2011 (UTC)

Why not only Newtonian mechanics terms?

I think this article should use Newtonian Mechanics and references to specific force. Why should a discussion on accelerometers be considered in the fundamental physics category. Newtonian Mechanics is perfectly able to explain the basic operation of the accelerometer to the level that most people need. I know that in satellite systems that relativistic effects are accounted for in the inertial navigation system.Skimaniac (talk) 02:57, 26 September 2010 (UTC)

Specific force is just a fancy name for "weight per unit mass." G-force (also in units of acceleration) is exactly the same concept, but with yet another name. An accelerometer does measure g-force, specific force, and weight-per-test-mass. These should all be mentioned in the lede. This type of acceleration is called proper acceleration in relativity theory, to distinguish it from coordinate acceleration (dv/dt) that we all think about when we see the word "acceleration". Accelerometers cannot, and do not, directly measure dv/dt (although they CAN in the very special case where gravitation is absent). That causes a LOT of confusion. SBHarris 19:38, 17 January 2011 (UTC)

This confusion can be resolved in purely Newtonican mechanics by introducing the vector a-g where g is the local acceleration of gravity. The use of relativity jargon is completely misplaced here. (The weak equivalence principle does not need relativistic considerations.)Ann-Marie Pendrill (talk) 09:37, 11 November 2018 (UTC)

Move applications to a separate page?

Most of the article is just a very long list of applications. I suggest that this is put into a separate article and just a reasonably short summary included here. Furthermore, I came to page to find out what kind of sensors are used in consumer electronics but I could not find the answer by skimming the text. Quite possibly more emphasis should be put on that. Tronic2 (talk) 15:46, 20 June 2012 (UTC)

I don't have any problem listing applications in a separate list article, and moving specific material from here to there. Another possibility is to leave them here, with summaries and collapse boxes. SBHarris 00:10, 21 June 2012 (UTC)

Collapse boxes are no good for articles as you expect to read it without clicking. If there s so much text that it has to be collapsed, then another article is the way to go, with a summary in the parent article. Graeme Bartlett (talk) 11:11, 21 June 2012 (UTC)

Collapse boxes are handy when you have >10-20 items (too many to put into a list without wasting space, but < 100-200, items so not really enough to make for a separate list article. You just need to be judicious about it. And have some respect for the reader. He or she will click if they have to, and are warned. Example: [5]. SBHarris 20:58, 21 June 2012 (UTC)

I think that this article should consider who would be interested in accelerometers. It is a bit ridiculous as it stands starting with jargon (i.e. jargon as seen from people who use accelerometers, which have their main application in Newtonian mechanics). I suggest that the relativity considerations be moved to a separate section further down (or a page of its own) Ann-Marie Pendrill (talk) 09:33, 11 November 2018 (UTC)

Lead

A rambling relativistic discussion of acceleration is not necessary twice in the article, and certainly not in the lead. Might it not sbe more encyclopediac to refer to whatever artcile we have on [[acceleration] and leave this article to talk about its nominal subject, namely, accelerometers? --Wtshymanski (talk) 22:00, 31 October 2012 (UTC)

You cannot discuss accelerometers without noting that they actually do not measure acceleration, but rather weight per mass. Relativity doesn't need to be discussed, and actually isn't in the lede. But it's necessary to discuss the fact that gravity is not felt by either people or accelerometers. If you're falling down an elevator shaft in vacuum, you'll be weightless. You won't feel gravity even though it is acting on you, and your accelerometer will read "zero" even though you're clearly accelerating due to gravity. See the point? This is not a minor detail. It is explained by the equivalence principle, but that can be left until later. The present lede sentence: "An accelerometer is a device that measures proper acceleration, the combined effect of gravity and change of speed," is not informative. Proper acceleration is not a "combined effect of gravity and change of speed" unless you explain that they can "combine" in tricky ways to cancel each other out. And that takes a lot of explaining. It's easier just to say that accelerometers measure the types of acceleration that produce weight. SBHarris 23:14, 31 October 2012 (UTC)

So how does an inertial navigation system work? Missiles are in free fall for much of their trajectory...surely the INS needs to find Moscow accurately? The current version of the lead is much better and I don't think the lack of mention of "four-acceleration" hurts it at all; after all, physics grad students aren't going to be looking up accelerometers on the Wikipedia and the rest of us have no idea what a "four-acceleration" is. --Wtshymanski (talk) 13:53, 1 November 2012 (UTC)

Inertial nav systems use gyroscopes to measure angular motion and the accelerometers measure "weight per mass" (which has the unfortunate name of "specific force"-- this actually means "mass-specific force" which means F/m, which has units of acceleration). In order to go from weight/mass to coordinate acceleration (what you really want to integrate, in order to get velocity and position) you need to add or subtract gravitational acceleration from "weight/mass" acceleration, and coordinate acceleration, dv/dt, is what's left. How do you know gravitational acceleration? You have to calculate it from your position and some kind of gravity model, since you can't actually measure it. The gravity model tells you the value of "g" for altitude and latitude and it would be different for every planet. We use Earth. The position is your initial reference position plus the change in distance you calculated from your inertial nav system, so it's a bootstrap thing. But it's a quantity in the system that needs to be compensated for, wherever gravitation is an issue. This can be done in realtime if you're on a slow machine like a submarine. For an ICBM which has no guidance, all the calculations for gravity compensation over the path it's supposed to take, are done on the ground, and fed into the missile as a sort of set of parameters that it should be seeing from its accelerometers and gyros. The ballistic missile then performs a few course corrections to try to minimize deviations from this, in a process called Q-guidance. The process ends up essentially doing the integrations described (although much of the problem has been "pre-computed"). For (quite different) guided missiles and cruise missiles, of course, there are external corrections that can be used to keep the thing on track. Ballistic missiles go by dead-reckoning, but they use a gravity model of SOME sort, as pre-crunched part of input. If you put a giant mountain in their path that wasn't part of the gravity model they use, they wouldn't know about it, and would miss their targets. A new mountain would not show up in pre-computed parameters that come from from prior assumptions about the gravity field of the Earth along the ballistic path.

By the way, some of the arguments above have convinced me not to mention proper acceleration in the lead, or at least qualify it. Specific force (weight per mass) is very close to proper acceleration for most circumstances, but technically is not quite the same if there are influences on the proof mass which affect its motion, but aren't macro-gravitational, and at the same time, don't show up as force between mass and case. Natually these are small, but a list of them is given above, and they aren't quite zero. SBHarris 00:06, 3 November 2012 (UTC)

I re-wrote the first paragraph. I think this is confusing enough to most people that we need specific examples. Also, I didn't like the previous language because it referred to weight in a way that isn't really consistent with the wiki article on weight (which defines it as due to gravity). As for the second paragraph, it really doesn't seem important enough to be there, and it makes the lead too long - so I deleted it. Waleswatcher (talk) 16:01, 4 November 2012 (UTC)

Consumer Products Section

In the Consumer Products Section within Applications, the following sentence appears: . '.... It is also used in some data loggers to monitor handling operations for shipping containers. ...' . Am I missing out on something, do all the other good consumers have their own shipping containers? Where are they hiding these? What do they do with/in the containers? Why the effort to conceal this from me (and probably others)? . Maybe someone would be kind enough to move this to the Industrial section, which will allow me to pretend like I'm not missing out on some cool consumer application of shipping containers. . 70.185.109.98 (talk) 23:10, 22 February 2013 (UTC) BGRIFFIN

Consumer Products Section

In the Consumer Products Section within Applications, the following sentence appears: . '.... It is also used in some data loggers to monitor handling operations for shipping containers. ...' . Am I missing out on something, do all the other good consumers have their own shipping containers? Where are they hiding these? What do they do with/in the containers? Why the effort to conceal this from me (and probably others)? . Maybe someone would be kind enough to move this to the Industrial section, which will allow me to pretend like I'm not missing out on some cool consumer application of shipping containers. . 70.185.109.98 (talk) 23:10, 22 February 2013 (UTC) BGRIFFIN

External links modified

Hello fellow Wikipedians,

I have just modified 3 external links on Accelerometer. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Replaced archive link http://iespuigcastellar.xeill.net/Members/vcarceler/articulos/jugando-con-el-wiimote-y-gnu-linux/sensor971.pdf with https://web.archive.org/web/20081213001559/http://www.analog.com:80/en/technical-library/faqs/design-center/faqs/CU_faq_MEMs/resources/fca.html on http://www.analog.com/en/technical-library/faqs/design-center/faqs/CU_faq_MEMs/resources/fca.html
• Corrected formatting/usage for http://www.wilcoxon.com/knowdesk/Guidance%20for%20mounting%204-20mA%20sensors%20on%20fans.pdf
• Added archive https://web.archive.org/web/20110604063917/http://www.memagazine.org/backissues/membersonly/june98/features/tilting/tilting.html to http://www.memagazine.org/backissues/membersonly/june98/features/tilting/tilting.html

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 18 January 2022).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 06:41, 3 October 2016 (UTC)

Improving the lead

On old version of the lead (one I wrote back in 2012) had this first paragraph:

An accelerometer is a device that measures proper acceleration. The proper acceleration measured by an accelerometer is not necessarily the coordinate acceleration (rate of change of velocity). For example, an accelerometer at rest of the surface of the earth will measure an acceleration g≈ 9.81 m/s2 straight upwards, due to its weight. By contrast, accelerometers in free fall or at rest in outer space will measure zero. Another term for the type of acceleration that accelerometers can measure is g-force acceleration.

I think that's probably an improvement on the current lead paragraph, although some kind of merge might be best. Thoughts? Waleswatcher (talk) 16:48, 11 November 2018 (UTC)

I think that the introduction to a page about Accelerometer should consider who is likely to be needing that information.

First a bit about my own background, and why I think that my voice should carry some weight:

My own relation to accelerometers is to measure g-forces in roller coasters and interpret the data coming out from 3D accelerometers, sometimes with gyro measurements. I also have a background in relativistic calculations in atomic physics, and did a post-doc at UW, Seattle, and have followed the work of the Eöt-Wash group search for possible violations of the equivalence principle.

And I very much appreciate the relation between the equivalence principle and what an accelerometer measures - and know that it is a counterintuitive concept. I have also argued in some of my publihsed work that the equivalence principle should be taught much earlier in school. No relativity is needed to do that! For many years I have taught introductory engineering courses, where students have been using accelerometers and interpreting the data (and also have a large web site supporting teachers usings accelerometers in amusement parks).

In my papers about forces in roller coasters and other rides, I always define an expression based on the vector a-g, or some variation of it. I also have contacts in the standards community for Biomechanical effects on amusement ride passengers.

The terminology is tricky, yes.

BUT

I do not think that talking about "proper acceleration" is going to be any help for anyone seriously looking for information about accelerometers. Nor do I think that the community who prefers the term "proper acceleration" is, in general, particularly interested in accelerometers.

Looking through old talk sections for the page, I noticed a number of different expressions (including g-force, normalized force and a few more expressions - but I have never used the term "coordinate acceleration", which I think may lead to other conceptual problems). One possible way to open the page for many different types of audiences might be to include all possible expressions, possibly with reference to the groups using them.

I also think that it might be worth dividing the page into subpages for different communities.

It might also be worth adding a section about accelerometers in education.

Best regards from a place where g≈9.82 m/s^2.

Ann-Marie Pendrill (talk) 19:32, 11 November 2018 (UTC)

I also thought for a while about why I don't like the phrase "coordinate acceleration" (which I have not come across before). I think it actually feeds a common misconception that acceleration is relative, whereas in Newtonian mechanics acceleration is absolute while velocity is relative.

I also think that the expression "falling ... at a rate" signals speed, not acceleration (as it is used in the introduction),

but I do not expect any misconceptions to arise from that expression. — Preceding unsigned comment added by Ann-Marie Pendrill (talk • contribs) 13:25, 12 November 2018 (UTC)

See the literature:

Google	Scholar	Books
"coordinate acceleration"	1190	607
"coordinate acceleration" "proper acceleration" accelerometer	460	6

- DVdm (talk) 13:35, 12 November 2018 (UTC)

It doesn't follow that it doesn't lead to conceptual problems.

I think the way you are running this page makes it fairly useless for those who try to learn about the physical background behind accelerometers - I mean: learn and understand, rather than echoing jargon, that is not needed for this understanding. This is not qualified refereeing as I have come to expect from serious publications.

The page reads like a student report, where students have brought together scraps from miscellaneous sources, without any attempt to show the relations between the different ways people view and describe a phenomenon, depending on their background and context.

It also seems like this page has been abandoned for a long time by people who have tried to make serious contributions and aimed for serious interaction.

I don't notice any serious intellectual engagement in the discussions from your or your anonymuos coeditors, so I am afraid that I have to leave this messy page in its sad stage.

And I will be even harsher than before when students make wikipedia-references for their work.

Ann-Marie Pendrill (talk) 16:32, 12 November 2018 (UTC)

Hi Ann-Marie Pendrill, I don't think we can write a lead for this article that doesn't say that accelerometers measure proper acceleration, because that is simply the correct physical term for what they measure. We also can't write one that will perfectly satisfy the needs of all possible readers. In the end wikipedia is for everyone, and articles have to follow the basic guidelines (mainly, being based on reliable sources). With that said, the use of one technical term shouldn't prevent someone unfamiliar with it from understanding the lead. So I ask again - what precisely would you like to see changed or improved? Do you think the old wording I posted above is (a) worse than, (b) better than, or (c) equally bad as the current wording? Do you have other specific suggestions? Do you think we should write an equation involving a-g in the lead? If so, can you propose some text, and give a reliable source? Thanks. Waleswatcher (talk) 18:25, 12 November 2018 (UTC)

This is a response to your comment above - I'm moving it here since it's hard to carry on a discussion in two sections at once. Above you wrote: "Come on! What is the problem?? For a spinning satellite, g is obviously the local acceleration of gravity (of -GM/R^2 if you prefer)." That is not only not obvious, it is not true. An accelerometer on a non-spinning satellite will read 0 (so the "g" in your "a-g" formula should be zero), irrespective of the radius R the satellite is orbiting at and the mass M of the earth. On a spinning satellite it will read an amount that depends on the spin and its distance from the spin axis, so there is again no notion of non-zero "g". So even someone with experience in this topic (such as you) can make elementary mistakes, which makes it all the more important to write the article carefully and correctly. Waleswatcher (talk) 18:45, 12 November 2018 (UTC)

YOU are wrong! In a spinning Satellite, there are two accelerations to consider. The acceleration toward Earth (or whereever) leads, indeed, to an acceleration of the centre of gravity, and will not show up on accelerometer placed at the centre of gravity of the satellite.

However, in addition, there is an acceleration toward the centre of mass due to the spinning needs a centripetal force, and will give a reading corresponding to this acceleration (unless the accelerometer is at the centre of mass or at the axis of rotation for this motion).

In a satellite around Earth there is a slight difference in perceived gravitational field in different parts of the satellite, for two reasons:

1: The inhomogeneity of the gravitational field around the Earth

2: The difference in centripetal acceleration in the orbital motion.

These effects add up to what is called "microgravity".

(Note that the centre of gravity and centre of mass are not identical in an inhomogeneous gravitational field). — Preceding unsigned comment added by Ann-Marie Pendrill (talk • contribs) 07:03, 13 November 2018 (UTC) Ann-Marie Pendrill (talk) 07:19, 13 November 2018 (UTC)

It seems like we are also in conflict with what we denote by "g". To me "g" is always the gravitational force on an object divided by its mass and an acceleration is always an acceleration. And an accelerometer measures neither a nor g. Possibly you could say that it measures a "perceived acceleration of gravity" for an object at rest in a system.

I notice that I am not the first one to suggest a-g to define what accelerometers measure.

There are also some version of it, with the normal force divided by N/m which, of course is also

(F-mg)/m or a-g.

Concerning references ... this is bascially introductory textbook stuff, so I don't think anyone really deserves citation credit for this.

I think this page needs much more work than just a few sentences in the introduction, but one way to be for "all" would be to include a-g and versions, together with "g force" and whatever other definitions have been suggested.

Ann-Marie Pendrill (talk) 07:33, 13 November 2018 (UTC)

I saw that you had also inserted "g force acceleration" in the introduction. I think that "g-force" with quotation marks is better. — Preceding unsigned comment added by Ann-Marie Pendrill (talk • contribs) 07:34, 13 November 2018 (UTC)

Please sign all your talk page messages with four tildes (~~~~) and indent the messages as outlined in wp:THREAD and wp:INDENT — See Help:Using talk pages. Thanks.

That is why Wikipedia needs wp:reliable sources: to avoid just this: "You are wrong. No, YOU are wrong! Nonono, it is YOU who is wrong." We do not discuss the subject here. We discuss content additions, changes and removals here, and when challenged, we must bring reliable, agreed-upon sources. See wp:Talk page guidelines. The a-g thing will need a proper book source. Without that, it is wp:original research. - DVdm (talk) 07:55, 13 November 2018 (UTC)

To Anne-Marie - I'm not sure where you think I'm wrong. But indeed, perhaps for you the accelerometer on a (non-spinning) satellite reads a-g=0, whereas for me it simply reads a_proper = 0. The point, fundamentally, is that you cannot distinguish "the gravitational force on an object divided by its mass" from "an acceleration" - they are totally indistinguishable, and that is why no definition like "a-g" truly makes sense. Anyway, as DVdm says, whatever is written in the article requires reliable sources per wiki policy. If you can find some and propose some specific changes, then we can discuss whether they are an improvement. Thanks. Waleswatcher (talk) 01:27, 14 November 2018 (UTC)

Hi, Waleswatcher. We may both be right, but your claim that I am wrong is still incorrect :-) I think we first need to sort out what we agree on and where we disagree.

We both agree that from inside a black box without external references (like Einstein's elevator) there is no way to distinguish gravity from acceleration (in the opposite direction). We agree that the equivalence principle is a beautiful fundamental principle. We agree that the equivalence principle is also fundamental to the working of accelerometer. We agree about what an accelerometer would read in different situations. We agree that what an accelerometer measures can be confusing is often misunderstood by people who haven't reflected about it. We probably agree that the equivalence principle deserves a more prominent place in education and in an understanding of the world.

You may agree that different people can use different terminology to describe the same things. E.g. a roller coaster designer would happily refer to a-g as "acceleration", whereas earth scientists are more likely to refer to the combination of Earths gravitational field and the centripetal acceleration due to the rotation of the Earth as "gravity" (which is what is measured by "gravimeters").

You seem to DISAGREE that "acceleration of gravity" is a well-defined and meaningful concept, in spite of it being part of any high-school and undergraduate physics curriculum - and follows from Newton's law of gravitation.

However, this is part of a more fundamental disagreement, about the need to use a relativistic framework to discuss accelerometers. (Remember, Einstein started from the equivalence principle to get to general relativity - not the other way round.) This disagreement is so fundamental for the choice of presentation, that I see no way for them to coexist throughout the text. For my taste, the reference to proper acceleration should be a footnote to a page about accelerometers, not the main story.

I leave you with a quote from part 1 of Feynman's Lectures on Physics, section 1-1 (talking about velocity dependent mass, but relevant even here)

"Now, what should we teach first? Should we teach the correct but unfamiliar law with its strange and difficult comceptual ideas, e.g. the theory of relativity, four-dimensional space-time and so on? Or should we first teach the simple "constant-mass" law, which is only approximate, but doesn't involve such difficult ideas? The first is more exciting, more wonderful, more fun, but the second is easier to get at first, and is a first step to real understanding of the second idea. This point arises again and again in teaching physics. At different times we have to resolve it in different ways, but at each stage it is worth learning what is now known, how accurate it is, how it fits into everything else, and how it may be changed when we learn more."

Ann-Marie Pendrill (talk) 07:02, 16 November 2018 (UTC)

Accelerometer ‎ use for sound recording

There is nothing about promotion or advertisement in fact that Accelerometer can be used and is actually used for sound recording. This fact was published by IEEE in 2012 , later investigated by University of Michigan and others. User:Davidy2001 —Preceding undated comment added 07:38, 24 November 2018 (UTC)

Please sign all your talk page messages with four tildes (~~~~) — See Help:Using talk pages. Thanks.

I had reverted your edit because the first source is clearly promotional. If we agree here that it is not, then there's probably no problem. Comments from others welcome. DVdm (talk) 09:50, 24 November 2018 (UTC)

As is it's so poorly written as to be incomprehensible. If that material is going to stay, it needs a rewrite. Waleswatcher (talk) 12:22, 25 November 2018 (UTC)