Re: SiRF module and bluetooth

[Gary E. Miller]

Yo Dave!

On Fri, 26 Jun 2020 19:37:11 -0700
Dave Riesz <dave@riesz.net> wrote:

> Well, I think I may bail on the bluetooth link since it's adding
> complication without giving me much in return.

It happens.

> Part of your response gets to what I'm working on that doesn't deal
> directly with gpsd: signals and noise in the receiver.

Easy to get lost yak shearing.

> A while ago I wanted to try to use a pair of GPS receivers to collect
> some reasonably accurate (or at least consistent) position
> information for my property using one unit as a stationary reference
> and the other to collect locations.

SiRF III does not fall in the "reasonable accurate" bucket.

> The units are identical.  Same
> receiver module, same antenna, same RPi hardware, same software.
> Yet, even with the antennas sitting side-by-side the units would
> report positions that not only were quite different, but also jumped
> around with apparent randomness.

Yup.  This is why gpsprof was invented.  Give it a shot, you will
be surprised by the results.

SiRF claims 2 meter "accuracy".  They define "accuracy" as what you get 50%
of the time.  The rest of the time, anything goes.

https://web.archive.org/web/20070810205750/http://www.sirf.com/products/GSC3LPProductInsert.pdf

> Thus my attention turned to noise
> in the system.  I'd read that the 3.3v pins on the RPi (that I was
> using to power the GPS modules) weren't very clean

Yes, nut not in a way that affects GPS.

> and that the local
> oscillator(s) on a GPS unit could be sensitive to voltage
> fluctuations.

No.  But they are sensitive to temperature.  Once the receiver has
sat lock, that goes away.

> So I tried an RF choke on the power leads and then a
> separate power supply.  The bluetooth idea was to totally isolate the
> receiver from the RPi.

BT is 2.4GHz, so the worst noise source you have.  Not that it matters
much.

> Anyway, if noise from the computer really isn't an issue, what might
> be causing the discrepancy between the receivers?  The location and
> setup of my test units is not ideal and checks all of the boxes you
> mentioned:

If you really want to understand the actual error sources, you need
to devote a lot of time to the Stanford video course on GPS:

https://www.youtube.com/playlist?list=PLGvhNIiu1ubyEOJga50LJMzVXtbUq6CPo

For starters, the GPS ephemeris is only good to about +/- 1.5 m.  Then
add in effects from ionosphere, troposphere, atomic clocks, etc.

> - limited skyview: right next to the exterior house wall
> - multipath reflection: house is wrapped in a steel stucco lath
> - cable loss: cheap mouse antenna probably has a lossy cable
> 
> Not a great setup, but again both units are identical and are seeing
> very different things.

Can you define "very different".  20 meters would not surprise me.
Better yet, run a 2 hour gpsprof plot and send it here.

First thing: get a u-blox 9-series and your life gets much better.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
        gem@rellim.com  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

pgpWaRM1CX_n4.pgp
Description: OpenPGP digital signature