Re: [time-nuts] ``direct'' RS-232 vs. RS-232 via USB vs. PPS decoding cards
On Wed, Feb 15, 2017 at 11:48 PM, Chris Albertson
albertson.chris@gmail.com wrote:
On Wed, Feb 15, 2017 at 10:30 PM, Ruslan Nabioullin
rnabioullin@gmail.com wrote:
On 02/15/2017 01:17 PM, Chris Albertson wrote:
Why set up a dedicated NTP server if you only have two computers
that will use it? Your server will be accurate to a few
microseconds but your two computers will only by good to a few
milliseconds because ethernet is not nearly as good as PPS.
Well Ethernet can be extremely accurate if PTP is used (a whitepaper
specifies <= 100 ns accuracy if the LAN is optimized for it).
But PTP requires special hardware. You may not have this.
A cheap why to distribute time is use PPS on a back channel. You
build a little amplifier and rig it so all computers receive PPS from
the same GPS source. NTP runs on all these computers and you add the
"Atom" ref clock. The trouble is you need to run more wires,
Ethernet and PPS. But notice that only 1/2 of the wires in a
standard Ethernet cable are in use.
--
Chris Albertson
Redondo Beach, California
On Thu, Feb 16, 2017 at 7:55 AM, Chris Albertson
albertson.chris@gmail.com wrote:
But PTP requires special hardware. You may not have this.
I have to disagree.
I run PTP on a Raspberry Pi using its onboard USB connected NIC, and
onboard NICs on HP and Dell servers, I see +- 5 microsoconds jitter in
the one way delay across 4 fanless HP switches, and when PTP is
running on a Pi with a GPS hat with PPS, I see +50 -30 microseconds
jitter in the reported offset from master across the slaves.
If the requirement is for sub millisecond then PTP on commodity
hardware is a workable solution.
If sub microsecond accuracy is required, then the NIC, Ethernet switch
and time source hardware requirements change.
Cheers
Yo Chris!
On Wed, 15 Feb 2017 23:55:02 -0800
Chris Albertson albertson.chris@gmail.com wrote:
On Wed, Feb 15, 2017 at 11:48 PM, Chris Albertson
albertson.chris@gmail.com wrote:
On Wed, Feb 15, 2017 at 10:30 PM, Ruslan Nabioullin
rnabioullin@gmail.com wrote:
On 02/15/2017 01:17 PM, Chris Albertson wrote:
Why set up a dedicated NTP server if you only have two computers
that will use it? Your server will be accurate to a few
microseconds but your two computers will only by good to a few
milliseconds because ethernet is not nearly as good as PPS.
Well Ethernet can be extremely accurate if PTP is used (a
whitepaper specifies <= 100 ns accuracy if the LAN is optimized
for it).
But PTP requires special hardware. You may not have this.
No, PTP works optimally with special hardware, it can work with any
ethernet port.
I will give you that software mode PTP is no more accurate than
plain NTPv4.
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
sub Millisecond is EASY. my Apple 27" iMac is doing that right now
using just Internet pool servers. Yes I have a very good Internet
connection. 100 Mbit fiber and then the last meter is 1000BaseT
But still, milliseconds are really EASY. It is sub microseconds that
requires things like PTP and special hardware. It's at the uS level
where you have to work hard
If anyone needs to break a millisecond, just run the PPS to the
machine that needs it edit /etc/ntp.conf and you are now in the "few
uS" level. It costs nothing (if PPS is available.) But breaking that
uS barrier is not easy
On Thu, Feb 16, 2017 at 9:59 AM, shouldbe q931 shouldbeq931@gmail.com wrote:
On Thu, Feb 16, 2017 at 7:55 AM, Chris Albertson
albertson.chris@gmail.com wrote:
But PTP requires special hardware. You may not have this.
I have to disagree.
I run PTP on a Raspberry Pi using its onboard USB connected NIC, and
onboard NICs on HP and Dell servers, I see +- 5 microsoconds jitter in
the one way delay across 4 fanless HP switches, and when PTP is
running on a Pi with a GPS hat with PPS, I see +50 -30 microseconds
jitter in the reported offset from master across the slaves.
If the requirement is for sub millisecond then PTP on commodity
hardware is a workable solution.
If sub microsecond accuracy is required, then the NIC, Ethernet switch
and time source hardware requirements change.
Cheers
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