DIY GPS?

[bruce varley]

I don't want to do this myself, but the question came up at work today and I
wondered what the answer is: What would be involved in building ones own GPS
receiver, from discretes and MSI, plus maybe a micro? I'd guess that one
would need RF skills (frequency in the GHz?), including resonator design,
and a fair amount of signal processing stuff like balanced modulators, PLLs
and the like. Would it be at all a realistic home project?

 

[Mike Harding]

On Wed, 3 Dec 2003 21:54:39 +0800, "bruce varley"
<bxvarley@weqstnet.com.au> wrote:

I don't want to do this myself, but the question came up at work today and I
wondered what the answer is: What would be involved in building ones own GPS
receiver, from discretes and MSI, plus maybe a micro? I'd guess that one
would need RF skills (frequency in the GHz?), including resonator design,
and a fair amount of signal processing stuff like balanced modulators, PLLs
and the like. Would it be at all a realistic home project?

Not a chance. Just the timings involved in GPS processing
are amazing. My Garmin GPS12XL measures it's internal
temperature so that it can correct, in software, errors caused
by oscillator drift. The sats. send timing data derived from
onboard atomic clocks and the receiver has to be able to
resolve that sort of accuracy to be able to calculate position.
I've never looked at the software algorithms to decode and
derive positional data from the raw info. but I imagine they
would be a tad on the complex side too. I seem to recall
reading somewhere that it took hundreds of hours of time
on a supercomputer to design the original Rockwell chipset
which did most of the above. Try something simple like an
MRI scanner

Mike Harding

 

[Gregory Toomey]

It was a dark and stormy night, and bruce varley managed to scribble:

I don't want to do this myself, but the question came up at work today and
I wondered what the answer is: What would be involved in building ones own
GPS receiver, from discretes and MSI, plus maybe a micro? I'd guess that
one would need RF skills (frequency in the GHz?), including resonator
design, and a fair amount of signal processing stuff like balanced
modulators, PLLs and the like. Would it be at all a realistic home
project?

I worked with someone who was commuting between Australia & England every 6 weeks. In England he was designing GPS software.

You get signals from 3-6 satellites with (atomic) clocks. You then need to do triangulation & apply Einsteins relativity equations to get you location.
You can also get a velocity vector by looking at doppler shift.

Doesn't sound like a weekend project to me.

gtoomey

 

[Ken Taylor]

"Gregory Toomey" <nospam@bigpond.com> wrote in message
news:1313267.pAx9qKsfty@gregs-web-hosting-and-pickle-farming...

It was a dark and stormy night, and bruce varley managed to scribble:

I don't want to do this myself, but the question came up at work today
and
I wondered what the answer is: What would be involved in building ones
own
GPS receiver, from discretes and MSI, plus maybe a micro? I'd guess that
one would need RF skills (frequency in the GHz?), including resonator
design, and a fair amount of signal processing stuff like balanced
modulators, PLLs and the like. Would it be at all a realistic home
project?

I worked with someone who was commuting between Australia & England every
6 weeks. In England he was designing GPS software.

You get signals from 3-6 satellites with (atomic) clocks. You then need to
do triangulation & apply Einsteins relativity equations to get you location.
You can also get a velocity vector by looking at doppler shift.

Doesn't sound like a weekend project to me.

gtoomey

Actually it's trilateration, as you need the extra axis, but you're correct,

particularly in the conclusion. Just take a peek at some of the old GPS
units, pre-single and two-chip chipsets! Our GPS's are currently tracking 7
satellites - it's 6 or 7, unless we have some sort of problem or an
obstruction. Below 5 (IIRC) you no longer get altitude.

Ken

 

[gcd]

Hi,
the gps system began devlopment in the 1970s and the 1st satellites were up
in the late 70s ( 1978 if im not mistaken).

Given the military back then were about 5 - 10yr ahead of commercial
elctronics, technology available to you in the mid to late 80s would be
sufficient to do the job.

A GPS receiver is no different to a direct sequence spread spectrum
receiver, because that's what they are.

So if you are comfortable dabbling with spread spectrum, then you can dabble
with gps as well.

The complexity is more in the math rather than the hardware (imo).

All the satellites are on the same frequency and each has a different psuedo
noise gold code so that each satellites ephemeris data can be extracted.
After you've got code lock, you then extract the data and use that
information to determine your position.

You need to deal with orbit determination equations and use the data from
the satellite to ammend those equations and then do some kinematics.You then
compensate that location fix with models for errors such as inosphere,
troposphere etc.

The whole idea of using more than 3 satellites to create a fix is so you
don't have to use a very accurate clock on the ground. You create pseudo
ranges based on your inaccurate clock and you use 4 or more range spheres
(planes) to give better data for location fix.

the frequency used is 1.57542 GHz, the chip rate for the pseudo noise code
is 1.023 MHz. The codes used can be found in the interface control document
ICD-GPS-200 and that can be downloaded from the web at
http://gps.losangeles.af.mil/

give it a go

Cheers
Greg




..

You use the ephemeris or almanac data to tweak the
"bruce varley" <bxvarley@weqstnet.com.au> wrote in message
news:3fcdebc0$1@quokka.wn.com.au...

I don't want to do this myself, but the question came up at work today and
I
wondered what the answer is: What would be involved in building ones own
GPS
receiver, from discretes and MSI, plus maybe a micro? I'd guess that one
would need RF skills (frequency in the GHz?), including resonator design,
and a fair amount of signal processing stuff like balanced modulators,
PLLs
and the like. Would it be at all a realistic home project?

 

[Ken Taylor]

The OP's reference to 'discretes and MSI' makes it a tad difficult. As you
say, not impossible, but certainly more difficult than the run-of-the-mill
home project. However if you wanted to dabble with one of the one or
two-chip solutions, it may be fun. Except for the code. Good luck trying to
get that out of anyone!

Cheers.

Ken

"gcd" <gcdnoSPAM@austarmetro.com.au> wrote in message
news:3fcf1d67@news.comindico.com.au...

Hi,
the gps system began devlopment in the 1970s and the 1st satellites were
up
in the late 70s ( 1978 if im not mistaken).

Given the military back then were about 5 - 10yr ahead of commercial
elctronics, technology available to you in the mid to late 80s would be
sufficient to do the job.

A GPS receiver is no different to a direct sequence spread spectrum
receiver, because that's what they are.

So if you are comfortable dabbling with spread spectrum, then you can
dabble
with gps as well.

The complexity is more in the math rather than the hardware (imo).

All the satellites are on the same frequency and each has a different
psuedo
noise gold code so that each satellites ephemeris data can be extracted.
After you've got code lock, you then extract the data and use that
information to determine your position.

You need to deal with orbit determination equations and use the data from
the satellite to ammend those equations and then do some kinematics.You
then
compensate that location fix with models for errors such as inosphere,
troposphere etc.

The whole idea of using more than 3 satellites to create a fix is so you
don't have to use a very accurate clock on the ground. You create pseudo
ranges based on your inaccurate clock and you use 4 or more range spheres
(planes) to give better data for location fix.

the frequency used is 1.57542 GHz, the chip rate for the pseudo noise code
is 1.023 MHz. The codes used can be found in the interface control
document
ICD-GPS-200 and that can be downloaded from the web at
http://gps.losangeles.af.mil/

give it a go

Cheers
Greg




.

You use the ephemeris or almanac data to tweak the
"bruce varley" <bxvarley@weqstnet.com.au> wrote in message
news:3fcdebc0$1@quokka.wn.com.au...
I don't want to do this myself, but the question came up at work today
and
I
wondered what the answer is: What would be involved in building ones own
GPS
receiver, from discretes and MSI, plus maybe a micro? I'd guess that one
would need RF skills (frequency in the GHz?), including resonator
design,
and a fair amount of signal processing stuff like balanced modulators,
PLLs
and the like. Would it be at all a realistic home project?

 

[onestone]

You can take the simple route and use off the shelf RF front ends, and
chip sets, like those from Sirf. Even this just isn't realistically
worth the effort. Developing all your own software is obviously quite
feasible, since there are a lot of people out there who do it, but why
waste a lot of time? You won't save money, given that a modern 16
channel unit in OEM format can be had for around AUS$100, much cheaper
in volume.

Al

bruce varley wrote:

I don't want to do this myself, but the question came up at work today and I
wondered what the answer is: What would be involved in building ones own GPS
receiver, from discretes and MSI, plus maybe a micro? I'd guess that one
would need RF skills (frequency in the GHz?), including resonator design,
and a fair amount of signal processing stuff like balanced modulators, PLLs
and the like. Would it be at all a realistic home project?

--
Please remove capitalised letters to reply
My apologies for the inconvenience
Blame it on the morons that spam the net

 

[gcd]

Hi,
I reckon it makes it easier. The code autocorrelation can now be done in a
single fpga/cpld rather than dozens of discrete or custom ICs. And you can
cut and paste that block and (depending on the size of the fpga) get lots of
code lock ccts in the one fpga.

the fpga would then spit out the ephemeris data and phase or time tag it to
allow a micro to do the kinematics.

I doubt you'd use hdl for it, schematic capture would be the go. But yes,
when you start talking hdl and fpga it's certainly no atmel or pic project.

but if you wanted to, you could. nothing like a challenge to help you learn

Cheers
Greg



"Ken Taylor" <ken123@xtra.co.nz> wrote in message
news:bqo45h$24ea55$1@ID-76636.news.uni-berlin.de...

The OP's reference to 'discretes and MSI' makes it a tad difficult. As you
say, not impossible, but certainly more difficult than the run-of-the-mill
home project. However if you wanted to dabble with one of the one or
two-chip solutions, it may be fun. Except for the code. Good luck trying
to
get that out of anyone!

Cheers.

Ken

"gcd" <gcdnoSPAM@austarmetro.com.au> wrote in message
news:3fcf1d67@news.comindico.com.au...
Hi,
the gps system began devlopment in the 1970s and the 1st satellites were
up
in the late 70s ( 1978 if im not mistaken).

Given the military back then were about 5 - 10yr ahead of commercial
elctronics, technology available to you in the mid to late 80s would be
sufficient to do the job.

A GPS receiver is no different to a direct sequence spread spectrum
receiver, because that's what they are.

So if you are comfortable dabbling with spread spectrum, then you can
dabble
with gps as well.

The complexity is more in the math rather than the hardware (imo).

All the satellites are on the same frequency and each has a different
psuedo
noise gold code so that each satellites ephemeris data can be extracted.
After you've got code lock, you then extract the data and use that
information to determine your position.

You need to deal with orbit determination equations and use the data
from
the satellite to ammend those equations and then do some kinematics.You
then
compensate that location fix with models for errors such as inosphere,
troposphere etc.

The whole idea of using more than 3 satellites to create a fix is so you
don't have to use a very accurate clock on the ground. You create pseudo
ranges based on your inaccurate clock and you use 4 or more range
spheres
(planes) to give better data for location fix.

the frequency used is 1.57542 GHz, the chip rate for the pseudo noise
code
is 1.023 MHz. The codes used can be found in the interface control
document
ICD-GPS-200 and that can be downloaded from the web at
http://gps.losangeles.af.mil/

give it a go

Cheers
Greg




.

You use the ephemeris or almanac data to tweak the
"bruce varley" <bxvarley@weqstnet.com.au> wrote in message
news:3fcdebc0$1@quokka.wn.com.au...
I don't want to do this myself, but the question came up at work today
and
I
wondered what the answer is: What would be involved in building ones
own
GPS
receiver, from discretes and MSI, plus maybe a micro? I'd guess that
one
would need RF skills (frequency in the GHz?), including resonator
design,
and a fair amount of signal processing stuff like balanced modulators,
PLLs
and the like. Would it be at all a realistic home project?

 

[Wayne Peacock]

Hi All,

This could be of interest to some of you. Not a complete DIY gps
receiver but a quick way of using an lassen SQ module.

A USB powered GPS receiver for your PC / Laptop

http://www.embedtronics.com/GPS/lassenSQ.html

Regards
Wayne Peacock
_____________________________________________________________________
Embedtronics
www.embedtronics.com

Embedded System Design
Adelaide, Australia

Embedded Ethernet, Data & GPS Logging Systems, Nokia F-Bus Interfacing
AVR Microcontroller Designs, MP3 Players.

Email: wayne@embedtronics.com
_____________________________________________________________________

 

[Ken Taylor]

"Wayne Peacock" <wayne.peacock@embedtronics.com> wrote in message
news:3FD59103.3050601@embedtronics.com...

Hi All,

This could be of interest to some of you. Not a complete DIY gps
receiver but a quick way of using an lassen SQ module.

A USB powered GPS receiver for your PC / Laptop

http://www.embedtronics.com/GPS/lassenSQ.html

Regards
Wayne Peacock
_____________________________________________________________________
Embedtronics
www.embedtronics.com

Embedded System Design
Adelaide, Australia

Embedded Ethernet, Data & GPS Logging Systems, Nokia F-Bus Interfacing
AVR Microcontroller Designs, MP3 Players.

Email: wayne@embedtronics.com
_____________________________________________________________________

That's not quite the MSI/discrete solution that the OP was hoping for, but

it is a *very* cool way of doing it yourself! Nice link.

Ken