Need Precise Electronic Compass

bobledoux · 2006-08-10 23:09

I'm working on an application that requires a compass. I need to read within 0.5 degrees. The output needs to be linear and repeatable. By repeatable I mean, if I sweep 30 degrees and then return to my starting point, the result will be read as my original heading +/- 0.5 degrees.

The electronic units I've been looking at don't appear to meet these requirements. I'm wondering about mounting a supermagnet on the end of a 10 bit encoder. If encoder friction is low enough the super magnet may follow magnetic North.

I can design my application so the compass apparatus is always parallel to the ground, so tilt isn't a problem.

I can also live with an absolute error or 1 or 2 degrees, provided the readings are consistently linear, or an algorithm can be created to adjust for the error to a final value of +/-0.5 degrees.

The apparatus can be placed to eliminate the impact of ferrous objects. But it needs to be moveable, typically within +/- one degree of longitude or latitude.

Any ideas out there?

Franklin · 2006-08-11 22:09

How about this?

http://www.ssec.honeywell.com/magnetic/datasheets/HMR3500.pdf

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen

John R. · 2006-08-12 17:18

From references on the same site, see also the 3100 (at only $100.00 vs. $675.00)

http://shop.ssec.honeywell.com/shopdisplayproducts.asp?id=6&subcat=12&cat=HMR3xxx+-+Products+and+Accessories

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
John R.

8 + 8 = 10

LoopyByteloose · 2006-08-12 20:27

Having lived in Oregon for 15 years, I learned a lot from woodsmen. In particular, I used a compass to hike a straight line of 15 miles for timber cruising [noparse][[/noparse]inventorying the trees in a forest].

A compass is a fundamental tool for them, but the tool has its limits. Anything ferrous that is nearby can throw them off by quite a bit. For example, if you have a roving robot and it passes a metal table leg, you are not going to get a good reading.

When hiking in the woods, metal buttons on a Levi jacket can make for lost woodsman. Or a metal watchband. A good woodsman will set the compass down on a rock and look at it from a bit of distance.

So think about it. Nickel is ferrous [noparse][[/noparse]as in NmH batteries], and their case is Iron [noparse][[/noparse]also ferrous].
If you look at ships and sail boats, even they have compensation for the ferrous metals around them.

That seems to be 'the way of the compass'.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"If you want more fiber, eat the package.· Not enough?· Eat the manual."········

···················· Tropical regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

bobledoux · 2006-08-12 21:58

My application calls for me to set up my compass unit. I then make readings on a distant moving object. This process will be repeated on a later date but at a slightly different longitude and latitude.

The comment about moving a degree of latitude or longitude was in recognition that electronic earth field sensors require new calibration with significant longitude changes.

TimC · 2006-08-14 15:14

Try:
www.robot-electronics.co.uk/

or maybe in the states at:
www.acroname.com/index.html

Should be around $50-$60

The software side is easy, here is a full workout of what the unit does:

' {$STAMP BS2px}
' {$PBASIC 2.5}
'***********************************************************
'**                                                       **
'**   CMPS03 Demonstration Software for the Basic Stamp   **
'**                                                       **
'**                                                       **
'***********************************************************
' Working on July 31 2006 Tim

SDA        CON 8      ' Define Data pin
SCL        CON 9      ' Define Clk pin
Compass    CON $C0     ' Compass I2C Address
CmdReg     CON 0       ' Command register
BearingReg CON 2       ' dir register

Bearing    VAR Word    ' 16 bit variable for Bearing
Version    VAR Byte    ' should return low digit number like 9
LowRes     VAR Byte    ' 8 bit Bearing

   DEBUG "*** Compass Test Program Ver1 ***", CR
    I2COUT SDA, Compass, [noparse][[/noparse]0]
    I2CIN SDA, Compass, [noparse][[/noparse]Version]
   DEBUG "CMPS03 Version: ", DEC version, CR

Main:
   ' get High Res 0-360.0 scale
   I2COUT SDA, Compass, [noparse][[/noparse]BearingReg]
    I2CIN SDA, Compass, [noparse][[/noparse]Bearing.HIGHBYTE, Bearing.LOWBYTE]

   ' get data on 0-255 scale
   I2COUT SDA, Compass, 
    I2CIN SDA, Compass, [noparse][[/noparse]LowRes]

    'DEBUG "Compass: ", DEC Bearing/10, " Degrees", CR ' Use terminal window
   DEBUG DEC Bearing, " Degreesx10, OR 0-255: ", DEC LowRes, CR ' Use terminal window

    PAUSE 1000 ' slow down
GOTO main

Regards
Tim

Need Precise Electronic Compass

Comments