View Single Post
  #5   Report Post  
Lloyd Sumpter
 
Posts: n/a
Default Cheap and Nasty Knotmeter

On Wed, 21 Jan 2004 19:10:31 +0000, Glen "Wiley" Wilson wrote:

On Wed, 21 Jan 2004 09:18:16 -0800, "Lloyd Sumpter" wrote:

Can I get a new sender, with proper documentation so I know how it
works, and hook it into my laptop?


I'm afraid I don't have anything like an answer to your question, though I would
have thought that the sender wouuld output pulses rather than a variable
voltage, leaving the instrument to count the pulses. You can see some readymade
data acquisition solutions at B&B electronics http://www.bb-elec.com/ but I
suspect the cost would qiuickly approach that of a new instrument, if cost is
your major objective.

If it is pulse data, you might be able to attach (properly buffered/amped) it to
something like the RTS line of the serial port. You can definitely capture RTS
events if you take direct control of the port, and probably even if you use the
standard MSComm activeX object. Since Windows is kind of crappy when it comes
to doing anything in real time, you'd likely lose pulses occasionally though. I
think you'd be better off to count them with a circuit and just pass along the
count. Don't know nuthin bout no parallel ports, though.

I'm actually more interested in your thought process. I have something of a
professional interest in displaying this sort of thing on PCs. To date, I've
worked with NMEA data, but I've been asked to look into direct acquisition of
certain data, so I'm wondering what leads you to want to do this. Of course, if
it's just cost, that's pretty self-explanatory.


Thanks

Doing a bit more research, I've found the OLD Signet meters sent out an analog
voltage (they were basically small generators), but the newer ones, notably
Airmar (www.airmar.com) send out pulses.

Airmar even has some samples of how to interface with CMOS and TTL. I thought of
using one of the RS232 signals, but think I'll use the parallel port instead.
I'm using Linux, so I'll attach the pulse signal to the interrupt and use Unix
signals to minimize CPU time. (Airmar says they output pulses about 4.8/sec at 1
knot, so max is around 100pps, or 10ms between rising edges).

I think I'll be interfacing the GPS as well, so I can compare readings. I may
get energetic and write a "calibration program" that would average the two
readings and adjust the "fudge factor" for the knotmeter (assuming you push the
"calibrate" button, then run a specified distance, then back)

Only questions now a 1. will an Airmar sender fit in my existing thru-hull,
or do I have to buy a new one (apparently they're not expensive), and 2. sample
programs in C showing how to access parallel port in Linux.

Lloyd Sumpter
"Far Cove" Catalina 36