Subject:
|
Re: Dinsmore Compass interface
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Sun, 2 Apr 2000 10:06:37 GMT
|
Viewed:
|
1454 times
|
| |
|
|
In lugnet.robotics, Wes Matchett writes:
> I've read a lot of discussion about compass sensors ranging over almost two
> years and as of yet have heard a success story.
I have a working interface for the RCX. The 'problem' with it, is that it is
all surface mount, making construction a 'bit of a pig' for anybody who is not
experienced with this type of construction. It returns 0-96 'round the
circle', seen as a normal 'light' sensor, and is pretty good. At 'north', it
will sometimes see 96, and sometimes 0 (depending on which direction you have
arrived from). The Dinsmore sensor is to my mind 'overdamped' in this
application, I had to write my code so the unit stops, takes a reading, then
corrects it's course, rather than working 'on the fly'. You also have to be
very careful of magnetic interference from the motors...
> I've reached the point I'd like my RCX to be able to accurately do two things:
> 1) be able to follow a sequence direction/distance commands.
> 2) be able to go someplace, recording its movement, and be able to retrace • it
> path.
>
> Since the granularity of non-military grade GPS accuracy is expressed in terms
> of yards .vs. inches - it seems that compass based guidance is the best
> alternative.
>
> I am about to purchase a Dinsmore compass (model 1655) for RCX orientation.
> Before making the purchase, I'd like to hear from anyone with experience with
> either of the analog Dinsmore compasses (1525 or 1655).
>
> Since they produce an analog voltage in the 0-5v range I'm expecting the
> interface described by Michael at http://www.alynk.com/usr/gasperi/gpa.htm to
> work just fine.
You have to be more complex. The compass returns two voltages, one a sin
curve, and the other the cosine. The interface has to read both voltages, and
where they 'cross', and then convert these to generate the 'angle'. I use
linear interpolation between the crossing points, then a 'look up' table to
actually do the required conversion - this is done by a PIC12C671, inside the
actual sensor...
> My hope is that the cost involved will be minimal above the compass - $35, and
> that the interface is quite simple; simple enough that anyone willing to do a
> little soldering can build it. If this works, we should have a well priced
> direction sensor.
The 'cost' is about right, but as I said, the complexity a little more. I
could 'rebuild' my unit in non surface mount, without it getting too large (at
present it fits in a 4*4 'brick', by the height of the Dinsmore sensor (which
is unfortunately rather tall). I have since been trying other approaches, such
as very sensitive direct sensors rather than the Dinsmore unit, to try to
reduce the cost, and size...
Best Wishes
|
|
Message has 1 Reply:
 | | Re: Dinsmore Compass interface
|
| Roger, I'm interested in your interface, do you have schematics available? Or is this a package you sell? I've done a good deal of electronics work, so I'm ok with converting it to a non surface mount device. I have also just found a compass device (...) (25 years ago, 3-Apr-00, to lugnet.robotics)
|
Message is in Reply To:
| | Dinsmore Compass interface
|
| I've read a lot of discussion about compass sensors ranging over almost two years and as of yet have heard a success story. I've reached the point I'd like my RCX to be able to accurately do two things: 1) be able to follow a sequence (...) (25 years ago, 2-Apr-00, to lugnet.robotics)
|
4 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
Active threads in Robotics
|
|
|
|
|
|
