Re: [time-nuts] Looking for advice to get a submillisecond setup
jimlux@earthlink.net said:
So really, it's a matter of finding a place to put your Garmin receiver and
string a cable that's not too long to your *nix box running ntp.
The "place to put your Garmin receiver" may not be as simple as it sounds.
It needs a good view of the sky. Roof is best, but a window may be good
enough.
A machine room is unlikely to work, even if you have a good window. Too much
EMI. An elevator control room may have similar problems.
I've thought about building a pair of boxes that would use standard CAT-5
cables as an extender for this sort of thing. 4 pairs works nicely: power,
TX, RX, PPS.
People have done it with a Rpi, if you want to go that route.
The RPi has its Ethernet chip connected by USB so that adds a bit of fuzz.
I should try to measure that.
--
These are my opinions. I hate spam.
I think the easiest cable to make really long, if one must be long is the
antenna cable. Use 100 meters of the kind of cable they use for cable
TV. It comes double shield and has those compression type F connectors.
The cable can cary both the GPS signal and power for the amplifier that is
built into the antenna. If the cable is very long, You would need another
in-line amplifier, again powered by the cable itself.
And yes, a gps antenna needs a good view of the sky, but the receiver
itself can be 100+ meters away from the antenna
On Thu, Feb 19, 2015 at 11:36 PM, Hal Murray hmurray@megapathdsl.net
wrote:
jimlux@earthlink.net said:
So really, it's a matter of finding a place to put your Garmin receiver
and
string a cable that's not too long to your *nix box running ntp.
The "place to put your Garmin receiver" may not be as simple as it sounds.
It needs a good view of the sky. Roof is best, but a window may be good
enough.
A machine room is unlikely to work, even if you have a good window. Too
much
EMI. An elevator control room may have similar problems.
I've thought about building a pair of boxes that would use standard CAT-5
cables as an extender for this sort of thing. 4 pairs works nicely:
power,
TX, RX, PPS.
People have done it with a Rpi, if you want to go that route.
The RPi has its Ethernet chip connected by USB so that adds a bit of fuzz.
I should try to measure that.
--
These are my opinions. I hate spam.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
--
Chris Albertson
Redondo Beach, California
On 2/20/15 6:30 AM, Chris Albertson wrote:
I think the easiest cable to make really long, if one must be long is the
antenna cable. Use 100 meters of the kind of cable they use for cable
TV. It comes double shield and has those compression type F connectors.
The cable can cary both the GPS signal and power for the amplifier that is
built into the antenna. If the cable is very long, You would need another
in-line amplifier, again powered by the cable itself.
that's fine if you have a separate antenna(w/preamp) and receiver..
But something like the Garmin GPS-18x and other similar inexpensive
receivers have integrated antennas.
And yes, a gps antenna needs a good view of the sky, but the receiver
itself can be 100+ meters away from the antenna
I think you're getting into receivers that are well into the hundreds of
dollars range, if bought new.
For an inexpensive "NTP for few hundred dollars to get better than a
millisecond" end of things, I think the integrated GPS antenna/receiver
with a suitable computer right next to it is the way to go. Then you're
just running a network cable and power.
4 pair Cat5 sometimes works ok with RS232, sometimes not. At the 4800
bps of the garmins, it would probably work ok.
At least you're sending power from the same place as you're
generating/consuming the signals, so you don't have the common mode
voltage difference problem.
I like that idea in general.. a pair for power, a pair for TxD, a pair
for RxD and a pair for 1pps. I'm not sure you'd want to connect the
"ground" at both ends of the signal pairs, though. What does the supply
current to the GPS-18x look like? Maybe it really doesn't make any
difference. Hopefully your computer's RS232 "input" isn't drawing 10s
of mA.
I think you're getting into receivers that are well into the hundreds of
dollars range, if bought new.
For an inexpensive "NTP for few hundred dollars to get better than a
millisecond" end of things, I think the integrated GPS antenna/receiver
with a suitable computer right next to it is the way to go. Then you're
just running a network cable and power.
I paid $36 for two Motorola Oncore receivers. These are 55 nanosecond (1
sigma) timing receivers. I think today they cost about $25 each. I paid
$27 for the timing antenna that is a helix the inside a pointed redone.
The dome has screw holes that fit a common galvanized pipe flange. This is
not an expensive setup and the parts are all there on eBay. You need to
add a power supply. I use a plug-in power cube.
But as I said. Use what you have all GPS receivers have more precision
than NTP can make use of. You don't need high-end gear if the requirement
is only "sub millisecond".
Mostly I'd say don't spend money on performance you don't need but in some
cases you get great performance for less money.
Chris Albertson
Redondo Beach, California
Standalone receivers don't have to be expensive. Take a look at the GPS
receiver modules at sparkfun.com. They are under $100 (some way under),
and some either require or can take an external antenna, and they
provide 1 PPS output. Garmin themselves sells receiver boards without
integrated antennas.
Now, they are navigation not timing receivers, so the 1 PPS accuracy is
likely only a microsecond or so, not in the nanosecond range. But that's
plenty for NTP. And because they are recent receiver designs, they have
higher sensitivity and faster acquisition than older receivers. Some
support WAAS corrections.
-Dave
On 20/02/2015 10:25, Jim Lux wrote:
And yes, a gps antenna needs a good view of the sky, but the receiver
itself can be 100+ meters away from the antenna
I think you're getting into receivers that are well into the hundreds
of dollars range, if bought new.
I am unsure which country you are in but the UK supplier
http://www.galleon.eu.com/computer-time-clock.html has a range of
reasonably priced units that may fit your requirements.
Regards
Peter Torry
On 20/02/2015 16:40, Chris Albertson wrote:
I think you're getting into receivers that are well into the hundreds of
dollars range, if bought new.
For an inexpensive "NTP for few hundred dollars to get better than a
millisecond" end of things, I think the integrated GPS antenna/receiver
with a suitable computer right next to it is the way to go. Then you're
just running a network cable and power.
I paid $36 for two Motorola Oncore receivers. These are 55 nanosecond (1
sigma) timing receivers. I think today they cost about $25 each. I paid
$27 for the timing antenna that is a helix the inside a pointed redone.
The dome has screw holes that fit a common galvanized pipe flange. This is
not an expensive setup and the parts are all there on eBay. You need to
add a power supply. I use a plug-in power cube.
But as I said. Use what you have all GPS receivers have more precision
than NTP can make use of. You don't need high-end gear if the requirement
is only "sub millisecond".
Mostly I'd say don't spend money on performance you don't need but in some
cases you get great performance for less money.
----- Original Message -----
From: "Jim Lux" jimlux@earthlink.net
To: time-nuts@febo.com
Sent: Friday, February 20, 2015 10:25 AM
Subject: Re: [time-nuts] Looking for advice to get a submillisecond setup
On 2/20/15 6:30 AM, Chris Albertson wrote:
I think the easiest cable to make really long, if one must be long is the
antenna cable. Use 100 meters of the kind of cable they use for cable
TV. It comes double shield and has those compression type F connectors.
The cable can cary both the GPS signal and power for the amplifier that
is
built into the antenna. If the cable is very long, You would need
another
in-line amplifier, again powered by the cable itself.
that's fine if you have a separate antenna(w/preamp) and receiver..
But something like the Garmin GPS-18x and other similar inexpensive
receivers have integrated antennas.
And yes, a gps antenna needs a good view of the sky, but the receiver
itself can be 100+ meters away from the antenna
I think you're getting into receivers that are well into the hundreds of
dollars range, if bought new.
For an inexpensive "NTP for few hundred dollars to get better than a
millisecond" end of things, I think the integrated GPS antenna/receiver
with a suitable computer right next to it is the way to go. Then you're
just running a network cable and power.
4 pair Cat5 sometimes works ok with RS232, sometimes not. At the 4800 bps
of the garmins, it would probably work ok.
At least you're sending power from the same place as you're
generating/consuming the signals, so you don't have the common mode
voltage difference problem.
I like that idea in general.. a pair for power, a pair for TxD, a pair for
RxD and a pair for 1pps. I'm not sure you'd want to connect the "ground"
at both ends of the signal pairs, though. What does the supply current to
the GPS-18x look like? Maybe it really doesn't make any difference.
Hopefully your computer's RS232 "input" isn't drawing 10s of mA.
One needs to be careful with extending the 18X "RS232" signal. It really is
not true RS232 but more like a 5 volt CMOS like signal.
If you are using it in a wood frame house, it might work fine indoors. But
in a steel/concrete multi-story commercial or industrial building, I would
be more surprised if it worked than not.
tm
On Fri, Feb 20, 2015 at 1:10 PM, Tom Miller tmiller11147@verizon.net
wrote:
----- Original Message ----- From: "Jim Lux" jimlux@earthlink.net
To: time-nuts@febo.com
Sent: Friday, February 20, 2015 10:25 AM
Subject: Re: [time-nuts] Looking for advice to get a submillisecond setup
On 2/20/15 6:30 AM, Chris Albertson wrote:
I think the easiest cable to make really long, if one must be long is the
antenna cable. Use 100 meters of the kind of cable they use for cable
TV. It comes double shield and has those compression type F connectors.
The cable can cary both the GPS signal and power for the amplifier that
is
built into the antenna. If the cable is very long, You would need
another
in-line amplifier, again powered by the cable itself.
that's fine if you have a separate antenna(w/preamp) and receiver..
But something like the Garmin GPS-18x and other similar inexpensive
receivers have integrated antennas.
And yes, a gps antenna needs a good view of the sky, but the receiver
itself can be 100+ meters away from the antenna
I think you're getting into receivers that are well into the hundreds of
dollars range, if bought new.
For an inexpensive "NTP for few hundred dollars to get better than a
millisecond" end of things, I think the integrated GPS antenna/receiver
with a suitable computer right next to it is the way to go. Then you're
just running a network cable and power.
4 pair Cat5 sometimes works ok with RS232, sometimes not. At the 4800 bps
of the garmins, it would probably work ok.
At least you're sending power from the same place as you're
generating/consuming the signals, so you don't have the common mode voltage
difference problem.
I like that idea in general.. a pair for power, a pair for TxD, a pair
for RxD and a pair for 1pps. I'm not sure you'd want to connect the
"ground" at both ends of the signal pairs, though. What does the supply
current to the GPS-18x look like? Maybe it really doesn't make any
difference. Hopefully your computer's RS232 "input" isn't drawing 10s of mA.
One needs to be careful with extending the 18X "RS232" signal. It really
is not true RS232 but more like a 5 volt CMOS like signal.
If you are using it in a wood frame house, it might work fine indoors. But
in a steel/concrete multi-story commercial or industrial building, I would
be more surprised if it worked than not.
I've had poor luck extending "fake" RS232 using cat5 wire. It works well
if you use differential signaling Convert the cos level serial to RS422
and you can go almost a mile using cheap cat-5 wire. And I've also have
worse luck extending a 1PPS plus using cat-5. The solution is RS422
signaling for the plus. But I finally gave up as running a longer
antenna cable has easier.
tm
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/
mailman/listinfo/time-nuts
and follow the instructions there.
--
Chris Albertson
Redondo Beach, California
On 2/20/2015 4:25 PM, Jim Lux wrote:
/For an inexpensive "NTP for few hundred dollars to get better than a
millisecond" end of things, I think the integrated GPS antenna/receiver
with a suitable computer right next to it is the way to go. Then you're
just running a network cable and power.
/
Given that the subject of a GPS unit integrated with the antenna has been touched, maybe some kind soul
could give me some information on this unit :
https://dl.dropboxusercontent.com/u/15089947/NavSymm1.jpg
https://dl.dropboxusercontent.com/u/15089947/NavSymm3.jpg
https://dl.dropboxusercontent.com/u/15089947/NavSymm2.jpg
Google is of no help...
TIA
Alberto
This email has been checked for viruses by Avast antivirus software.
http://www.avast.com
Looks like you need the "other half" of this unit. It looks like it is
designed to use fiber optic cable. You would need to other box to accept
the fiber cables, decode the data and interface it to a normal computer
port.
A good, brand new timing receiver is less then $100. So I'd not want to
spend that much making something else work.
On Sat, Feb 21, 2015 at 1:54 AM, Alberto di Bene dibene@usa.net wrote:
On 2/20/2015 4:25 PM, Jim Lux wrote:
/For an inexpensive "NTP for few hundred dollars to get better than a
millisecond" end of things, I think the integrated GPS antenna/receiver
with a suitable computer right next to it is the way to go. Then you're
just running a network cable and power.
/
Given that the subject of a GPS unit integrated with the antenna has been
touched, maybe some kind soul
could give me some information on this unit :
https://dl.dropboxusercontent.com/u/15089947/NavSymm1.jpg
https://dl.dropboxusercontent.com/u/15089947/NavSymm3.jpg
https://dl.dropboxusercontent.com/u/15089947/NavSymm2.jpg
Google is of no help...
TIA
Alberto
This email has been checked for viruses by Avast antivirus software.
http://www.avast.com
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/
mailman/listinfo/time-nuts
and follow the instructions there.
--
Chris Albertson
Redondo Beach, California
Symmetricon swallowed Navstar Systems Ltd. In 1993. So if there were any information available, they would likely have it, but I fear it may be long gone.
I did find this reference in the time-nuts archive:
<http://www.navsync.com/docs/mushroom_data_sheet.pdf >
Which looks like the same animal ...
HtH
David Partridge
Thanks for all the advice received on and off list.
My main doubt was about the quality of cheap receivers but you cleared
that doubt. To answer a few questions, the antenna would be put on top
of the building on top right corner of
http://referentiel.nouvelobs.com/file/2458497.jpg (the only square
with some kind of roof) in Paris. So I am confident we can receive a
good signal here. I don't think the tower could be a problem.
Regards
Matt
2015-02-21 17:10 GMT+01:00 David C. Partridge david.partridge@perdrix.co.uk:
Symmetricon swallowed Navstar Systems Ltd. In 1993. So if there were any information available, they would likely have it, but I fear it may be long gone.
I did find this reference in the time-nuts archive:
<http://www.navsync.com/docs/mushroom_data_sheet.pdf >
Which looks like the same animal ...
HtH
David Partridge
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
On 2015-02-24 03:48, Matt wrote:
Thanks for all the advice received on and off list.
My main doubt was about the quality of cheap receivers but you cleared
that doubt. To answer a few questions, the antenna would be put on top
of the building on top right corner of
http://referentiel.nouvelobs.com/file/2458497.jpg (the only square
with some kind of roof) in Paris. So I am confident we can receive a
good signal here. I don't think the tower could be a problem.
The tower could create multipath reflections, and electrical equipment
on a roof, such as elevator motors, could add noise.
On top of the tower or further away south west would reduce reflections.
Take care. Thanks, Brian Inglis
The tower could create multipath reflections, and electrical equipment
on a roof, such as elevator motors, could add noise.
On top of the tower or further away south west would reduce reflections.
This might be a very small problem if the goal were nanosecond level
timing. But the goal here is "sub millisecond" so the tower is a
non-issue. All you need is a view of most of the sky. don't worry about
"noise" either. NTP will NEVER notice effects smaller than a microsecond
Chris Albertson
Redondo Beach, California