Accelerometers

"Note that the curvature of travel must not affect the magnitude of EMF generated, only the acceleration of the system will influence the voltage produced."

No. Variations in speed and variations in direction are both forms of acceleration. You could add a third parameter - variations in gravitational pull - and it, too, would be indistinguishable from a changing rate of speed - an acceleration - in the vertical direction.

Without a point of reference outside your system, your accelerometer could not know what, exactly, was changing, as there are arbitrarily many combinations of change in speed and in direction that would produce identical readings from your accelerometer. Infinitely many, actually.

Sorry, Ace.

Thanks for your response, but I respectifully disagree with how you interpreted my question.

Look for MEMS accelerometers (I would recommend Analog Devices). With two or three axes devices you can get all information that you need. Keep in mind that they are gravitation dependent.

How would his system distinguish between its following a shallow, variable curve at a high rate of acceleration, or a tight, variable curve at a low rate of acceleration?

It can't!

A three-axis accelerometer will not give the OP results any better than none at all. Acceleration alone - however measured - is not enough, even in principle, to tell him what he needs to know, i.e., his change in velocity and change in direction, separately and distinct.

Personally, I would not recommend to him that he purchase anything until he proves to himself why this approach is incomplete and then come up with a better plan. One that will actually work. My suggestion to him would be to invest those funds in a nice evening at the theatre. Preferably with a good friend. His subconscious will be quietly working out the particulars of his next idea meanwhile.

Thank you very much for that piece of information. It will go a long way.

While you're at the movies, you might reflect on your question.

Remember, you know what you are trying to accomplish, we do not.

"the varying voltage to speed up or slow down a moving LED light along a path." could mean any number of things to the various engineering geniuses who hang out here. And it means different things to those of us non-geniuses who hang out here, too.

So, are trying to track a large object and relay the information to a tracking board, with lights? What's with the "EMF"? Do you really need a"field"?

Thanks for your response. The word "EMF" is used in place of "voltage". I am working on a "map-like" system to "mimmic" the motion of a large vehicle along a permanent pre-determined tract, that is imaged on a smaller map showing the vehicle's route of travel An LED light will move along a groved path on the surface of the map and will speed up or slow down to emulate the motion of the large vehicle.

To accomplish this, a small voltage will be produced by an accelerometer mounted on the vehicle whenever it speeds up or slows down, and transmits this voltage to the LED which in turn will move at a speed proportional to the magnitude of voltage it receives. Thus, if the large vehicle comes to a stop, so will the LED along the groved surface of the map. When it starts to move,so will the LED.

Ok,

So you are developing an answer to those GPS based systems that already do this?

Why not just wire up the track with some sensors? Is it cost?

That's the problem with coming here. We will want to redesign your system. You have to keep us in check by explaining what exactly you are trying to accomplish.

We don't want to wire the tract, a definite NO! Nor do we want to use GPS technology, for other technical reasons.

But Yhanks for your input.

Is there some reason you cannot tap into the vehicle's 'measurement' of its own speed, or must this device be completely self-contained?

I am very curious as to why you are building a physical map with a moving physical indicator when all this sort of info could be displayed much more elegantly on a computer graphic. This all seems very 70s-ish Retro as compared to modern ways of doing things, IMHO. Is this thing going to be a movie prop or something? A spy movie prop, perhaps, in which the thing, disguised as a cooler full of Guiness is to be spirited aboard the antagonist's own personal Orient Express by Fem-Bots wielding 30 mm nipple-guns?

Will there be a private viewing?

So the path is predetermined then?

This new information changes the solution landscape considerably. Your original post did not prescribe a specific path, known in advance. You simply said, "a path." A path can mean just about any old path. Could be a straight path. Could be a curved path. Could be both. And lacking specifics, certainly an arbitrary path at best.

If it's a Good Answer you want, my friend, there's nothing quite like asking a Good Question, and that means formulating your question with details that allow your potential benefactors to see the problem as you do, ideally.

Fortunately not too many showed up prior to your describing the Actual Problem, which is not the one you described at first. At any rate, you are in good hands here.

Wishing you the best.

Yeah, I can do it. All you need is a linear accelerometer AND an angular accelerometer together. Say, you wouldn't happen to want a bandwidth greater than about 0.2 Hz, would ya?

Thanks for that response.

But are you serous? can your device really produce a small, detectable voltage whenever a system, in which it is mounted, speeds up of slows down? This is the kind of effects I am looking for.

Well, yeah, except I don't have such a device. I could build one, at least in principle, but I don't have one.

What kind of specs do you have? What are you making?

In principle, if the orientation of the vehicle stays the same with respect to the track, an accelerometer oriented in the direction of motion could provide an output which, when integrated once would yield velocity, and twice, position -- provided that the track is level so that gravity does not get into the picture. This is the principle behind inertial navigation. In practice, you might have a problem with the accuracy of the accelerometer which would be magnified by the integration giving you increasing errors over time.

Thank you for that response, espectially the part about integrating the output (acceleration) twice. to yield both the velocity and position. Makes sense. You mentioned "accuracy" of the Accelerometers. Operationally, we definitely want to mimimize the errors that would result.

Iwill explore "inertial navigation" to develope this system further.

Thank you.

Braun Cameron

But, why don't you just use interference or even Doppler shift? That seems so much easier.

If by "system" you effectively mean (say) a "stable car", then a single accelerometer measuring longitudinal acceleration will do the job. Just make sure that accelerations not in the direction of interest do not introduce error.

Three axis devices et. are only needed where "the system" has to keep track of which way it is pointing and to output a velocity,displacement or acceleration that stay valid in terms of some outside fixed reference system.

If your "system" is more like a "spinning car" where the way it points is not necessarily the way it is travelling, then you'd need at leasts two axis of accelerometer to resolve what was required (if everything was constrained to a single "horizontal" plane). If not so constrained then you'd need three axis to do the job.

Thanks for the response.

The first system you described is the correct one. Thanks for the advise.

Put a rotary encoder with a follower wheel that runs along the track. Transmit the signal through the track to wherever you want to read it out. This will require a continuous track, perhaps welded, or maybe jumpers between each section. Just an idea.

Hi,

the devices you are searching for are "Geo-phones" used by thousands in the oil exploration activities and use a moving coil (with a housing-fixed magnet) to detect seismic waves.

You can make one quite simply by getting a good and rugged loudspeaker and glue some additional mass to the center. This is not really good as a transducer as the coil is optimised to generate force (low count of windings to operate at low voltage) and your application will be better suited by a coil of very many windings.

The generated voltage divided by the coil velocity is called the system constant k and is identical to the ratio of force to current! (Friction neglected).

If this approach is not generating enough voltage, then think about series connection of more than 1.

If still too low the get a better amplifier, if still too low, think about rewinding the coil.

Be cautious with data interpretation as the direction of your "car" is changing and so the direction of your sensitive axis.

We have used these "pendulum devices" for a series of accelerometers: metallic pendulum, ceramic housing, metallised surfaces, capacitive readout, pulse width modulating capacitance to voltage electronics.

Have success!

RHABE

Thank you for that rewsponse. Will the accelerometer continue to produce an output voltage even when the acceleration is at or near zero (constant velocity)? do you have the data to answer the question: Which type of acceoerometer is the least sensitive; or which of these continues to provide voltage when the velocity is constant in the direction of motion? I'm are looking for a system that will do just that.

Thank you very much.

Hi,

so you need a low-pass behaviour: sensitive from true DC to an upper limit that is set by construction and electronics and your needs.

So you have to use a mass-and-pendulum accelerometer, that is equipped with either capacitive or inductive or optical readout.

If you really want to have the least sensitive device there are some used in artillery - look at Schaevitz.

A cheap one is available at Analog Devices.

I got some good and cheap ones at ebay.

If in contrary you want to have a good sensitivity, good accuracy and good bandwidth (0..2KHz) then look for the Sundstrand Q-Flex (the cheapest was 2000$s each).

If the velocity is constant there is no acceleration so output-voltage is zero (or should be, an error is unavoidable.).

So look into high quality electronic integration with instrumentation amplifier and maybe autozero possibility.

Be very cautious with inertial navigation, first try to read the old book from Savet: The gyroscope and its applications. This task is a big one and is requiring usually a group of 10 to 20 engineers for 2 to 4 years to get the system running.

I wish you success

RHABE

For goodness sakes!

I'm still not sure what your application is, but you're not helping by holding back on details. You are, in essence, talking about running a class yard with a control room display like this

Now, your application may not have anything to do with trains, but the gist of the problem is the same.

Seeing this puts a sharp focus on my answer to you. Dead band will absolutely kill your application. If you want location, measure location, not one of its derivatives.