Find bearing without using magnetosensitive compnents?

[orienteraren]

Hi!

I´m experiencing a problem. I´m trying to design a circuit where I want to know the bearing. To know the correct angle (i. e. 227 deg) is not of importance but the deviation from the start angle is (i. e. +13 deg) and I cannot use a magnetosensitive component since this application will be used in environments with high precensce of iron. I was thinking of using a gyroscope and from the output there calculate the deviation but I will only record measurements at about 1 measurements every 15 second and I want to know the deviation in bearing I´ve made during this time. To increase the sample rate woud cause me a great hassle. I´ve been looking for components thet deliveres a bearing deviation but to no avail, does anyone know if such a component exists and where to find it or if there is another solution to this problem that I have not considered?

Thanks in advance

Fredrik

 

[karesz]

Hi,
can you use capactive or eddy current sensors pls?
Their are, especially eddy current-sensor, very robust & not so expensive, you can self build it light too...
K.

 

[orienteraren]

Hi! Thanks for you ide but I don´t think that an eddy-sensor would work. What I want to do is to take measurements on a vertikal path and I want to know if my measurement tool is rotating (i. e. positioned along another bearing) so I can compensate for this. A compass would be the given soultion but as mentioned there is a high precensce of iron in the vicinity. Do you still think it will be possible or not?

 

[karesz]

OK,
but especially eddycurrent sensors are made (too) i.e. for sensing/count on rotating cogwheels!
Eddy sensor is excellent to sens the art of metalls & the distances of these metalls too...
K.

Added after 7 minutes:

Mybe you can read some spetial article, apll notes, over subject at: www.micro-epsilon.com?...

 

[dick_freebird]

If it is relative to a starting position, maybe a multi-axis
accelerometer is your ticket. You could get angular info
based on the difference of two matched accelerometers
at opposite ends of a "beam", the difference becomes
either sine or cosine info (you would want two perpendicular
"beams" if you are plane-oriented for sure, three if not).

Difference in 3 dimensions, double integrate the info with no
appreciable error accrual, and bickety-bam, you're done.

Of you might be able to source an inertial measurement unit
already done, but that might be too easy (and expensive).

 

[diemilio]

What's the problem with the gyro? You don't need to increase the sample rate, commercial gyros have great bias stabilities so all you would need to do will be just integrate the output over time to get the angle (the gyro will give you rotation rate). The accelerometer approach seems a little bit too cumbersome to me, plus you would need to integrate TWICE to get the angle each of the accelerometers' outputs, meaning that your drift gets summed up twice! Also, you would need two separate devices to get it done, can't do it with a single multiaxis axl.

Hope this helps,

diemilio

 

[orienteraren]

Hi Diemilio! I thought that I could not use a gyro since the gyro is giving me a rate of angular change in "real time" and I wnat to know the total change in angle between sample one and two and the time between these samples will probably be at least 15 seconds. using a gyro and sampling at the same interval I thought that I could just get the output from the gyro every 15th second meaning that I will lose all of the information about the angular movement in the interval inbetween the sampling. The solution to this was to heavily increase the sampling rate to be able to use the info from the gyro but to that that I cannot use my current datalogger and have to buy a pretty expensive one. Hopefully I´ve got it all wrong and that there still is a solutio for me to use a gyro but I don´t unerstand how. Do you know a feasable setup for this?

Regards

Fredrik

 

[diemilio]

Dear Fredrik,

Let me get this straight...you say you can only sample every 15 seconds, but can't you use an analog gyro with an integrator at its output and then sample this value? By using analog components you will always have the rotation rate at the input of the integrator, hence the integrator won't loose any rate data and will be accumulating continuously.Then all you would have to do is only sample the output of the integrator whenever you want. Now, if you can't use any analog components you have a huge problem because you won't be able to use either a gyro or an axl cause in both cases you would have to integrate (unless the gyro could operate in whole angle mode, but there are no commercial INTEGRATED gyros that can do this). The only solution will be to find a sensor that gives you DISPLACEMENT right away, rather than angular velocity or linear acceleration and, as far as I know you can only do this with magnetic sensors.

Anyway, I hope this information is useful for you. Let me know how it goes.

Regards,

diemilio

 

[orienteraren]

Hi Diemilio!

Thanks for the advices. I´ve been looking around ang found the following nice little board:
https://www.sparkfun.com/commerce/product_info.php?products_id=9431
This board would give me all the dimensions which would make my measurements very accurate if the would be any (accident) tilt in any direction. So far I´m following you but what are your thought about the integrator circuit? The circuits I´ve found are all integrating current and giving a voltage aoutput. The gyros has a voltage output which would require some unnessecary work in converting back and forth I think. Was this what you had in mind or are there circuits that are able to integrate voltage signals in "real time"?

Best Regards

Fredrik

 

[diemilio]

What I'm trying to say is that (as you can see from the board's specs) your Pitch, yaw, and roll sensitivities are given in V/º/s (volts per rotation rate, not volts per rotation angle) meaning that to get the angle of rotation you will need to integrate that ouput over time.

I wasn't sure how you were bulding your system; I though you were putting everything together from scratch!! There should be some boards (like the one you sent me) in the market that give you Volts/angle right away (they do the integration for you). Look for those, that way you won't need anything additional to the board.

Regards,

diemilio