On 17/03/17 00:03, Mike Baker wrote:
Hello, Time-nutters--
Any thoughts on what the likely accuracy of smart phone time
displays might be? I am thinking that the stacking of delays
along the path to its receive antenna plus any internal processing
delays would accumulate to some unknown degree. Any
thoughts on this?
Mike Baker
Well..
Both the Android phones I carry (my Motorola "MoTo G", and a Samsung S6
for work, on different carriers here in the UK) are visually spot on
with the Thunderbolt and Lady Heather's display, and with the various
PC's I have NTP'd to the pool, or my local NTP server, and also with the
BBC time signal on Analogue FM radio, and the various HF time-signal
broadcasts.
However...
How would you get a low latency "time event" out of a phone to test it
against something else, without introducing other indeterminate delays?
I have in the past "Attempted" to use NTP over a mobile internet link,
with terrible results. When it did get a response from a time server,
the jitter was truly awful, in the multiple 10's of milliseconds range,
and that was from a fixed location in the clear, with a good phone signal.
As others have said, the GSM/3G network uses accurate timings
internally, so probably the handsets do maintain good sync to that, and
to whatever root standard the overall system is synced to. Be it GPS
or ???
AndroiTS GPS Test (V 1.48 Free) is good, but a battery hog I find. On
my Moto G, I find that it can handle not only the US GPS system, but
three other systems too, including Glonass, and I think the new Chinese
system. I don't recognise the last symbol, maybe Galileo. Not bad for
a consumer gadget.
I do not know if the phones could use GPS time directly, but that's only
a software task at a guess.
Regards.
Dave B.
G0WBX.
AndroiTS GPS Test (V 1.48 Free) is good, but a battery hog I find. On
my Moto G, I find that it can handle not only the US GPS system, but
three other systems too, including Glonass, and I think the new Chinese
system. I don't recognise the last symbol, maybe Galileo. Not bad for
a consumer gadget.
THIS is why the phones don't really track time so well. Not that they
can't but doing so requires battery power. It would need to keep
receivers and internal oscillators powered up full time. These
phones have an incredible degree of fine grained power management.
Processors can slow their clocks, graphic units can power off. The
best clocks required stable conditions and even temperature controlled
ovens for their crystals.
Bt the question raised here was more interesting: How accurate is the
DISPLAYED time. No need for an internal probe you can look at the
screen. The problem is that human eyes are not good at this and I
doubt anyone could detect even a 100 ms error. You need to rig up a
photo detector aimed at the screen.
One question was if the software stack from GPS/Cell receiver to
screen added any delay. Of course it does but did the designers
account for any fixed delay? You have to measure to know.
Only one manufacturer, Apple seems to give a worst case spec and they
only give it for a phone pared to a watch, 50 ms. But it might be
much better. I guess it is most of the time. Again, to answer what
was asked, you'd need some kind of optical sensor.
Chris Albertson albertson.chris@gmail.com wrote on Fri, 17 Mar 2017
at 14:38:17 -0700 in CABbxVHsRa41HoB=xu4nk4T_c39uh2BCOMu17Nb7H3rA8uHXLQg@mail.gmail.com:
AndroiTS GPS Test (V 1.48 Free) is good, but a battery hog I find. On
THIS is why the phones don't really track time so well. Not that they
can't but doing so requires battery power.
This statement doesn't seem to be well-supported. I think it's
basically untrue if we're talking about timing at the tens of
millisecond level. Anything more precise seems relatively useless in
a smartphone without specialized mechanisms to get the time off of the
phone.
A phone's GPS receiver takes a lot of battery. But GPS is not the primary
mechanism that phones use get their time.
They get time from the cellular phone network (whether from the layer
two timing information or at a higher layer with something like
NTP). The effort required to keep a phone's clock in sync, even with a
really bad local oscillator, is lost in the noise of all the other
things the phone has to do. It's just not a battery issue.
The only reason modern smartphones keep bad time is because their
designers can't be bothered to do better, or possibly the network is
providing "bad" time to the phone. (Unless I'm missing something.)
--jhawk@mit.edu
John Hawkinson
On my iPhone, I run an NTP client, Emerald Time, that displays the offset between NTP time and the device's internal clock.
I'm on T-Mobile US and the offset is typically in the low tens of milliseconds or better.
It's certainly accurate enough as a clock where all I'm looking at is one minute resolution.
Glen Hoag
hoag@hiwaay.net
Sent from my iPhone
On Mar 18, 2017, at 09:52, John Hawkinson jhawk@MIT.EDU wrote:
Chris Albertson albertson.chris@gmail.com wrote on Fri, 17 Mar 2017
at 14:38:17 -0700 in CABbxVHsRa41HoB=xu4nk4T_c39uh2BCOMu17Nb7H3rA8uHXLQg@mail.gmail.com:
AndroiTS GPS Test (V 1.48 Free) is good, but a battery hog I find. On
THIS is why the phones don't really track time so well. Not that they
can't but doing so requires battery power.
This statement doesn't seem to be well-supported. I think it's
basically untrue if we're talking about timing at the tens of
millisecond level. Anything more precise seems relatively useless in
a smartphone without specialized mechanisms to get the time off of the
phone.
A phone's GPS receiver takes a lot of battery. But GPS is not the primary
mechanism that phones use get their time.
They get time from the cellular phone network (whether from the layer
two timing information or at a higher layer with something like
NTP). The effort required to keep a phone's clock in sync, even with a
really bad local oscillator, is lost in the noise of all the other
things the phone has to do. It's just not a battery issue.
The only reason modern smartphones keep bad time is because their
designers can't be bothered to do better, or possibly the network is
providing "bad" time to the phone. (Unless I'm missing something.)
--jhawk@mit.edu
John Hawkinson
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.
I have that app but don't see an option to display "the offset between
NTP time and the device's internal clock." Please guide me.
Jeremy
On 3/18/2017 9:37 AM, Glen Hoag wrote:
On my iPhone, I run an NTP client, Emerald Time, that displays the offset between NTP time and the device's internal clock.
I'm on T-Mobile US and the offset is typically in the low tens of milliseconds or better.
It's certainly accurate enough as a clock where all I'm looking at is one minute resolution.
Glen Hoag
hoag@hiwaay.net
Sent from my iPhone
On Mar 18, 2017, at 09:52, John Hawkinson jhawk@MIT.EDU wrote:
Chris Albertson albertson.chris@gmail.com wrote on Fri, 17 Mar 2017
at 14:38:17 -0700 in CABbxVHsRa41HoB=xu4nk4T_c39uh2BCOMu17Nb7H3rA8uHXLQg@mail.gmail.com:
AndroiTS GPS Test (V 1.48 Free) is good, but a battery hog I find. On
THIS is why the phones don't really track time so well. Not that they
can't but doing so requires battery power.
This statement doesn't seem to be well-supported. I think it's
basically untrue if we're talking about timing at the tens of
millisecond level. Anything more precise seems relatively useless in
a smartphone without specialized mechanisms to get the time off of the
phone.
A phone's GPS receiver takes a lot of battery. But GPS is not the primary
mechanism that phones use get their time.
They get time from the cellular phone network (whether from the layer
two timing information or at a higher layer with something like
NTP). The effort required to keep a phone's clock in sync, even with a
really bad local oscillator, is lost in the noise of all the other
things the phone has to do. It's just not a battery issue.
The only reason modern smartphones keep bad time is because their
designers can't be bothered to do better, or possibly the network is
providing "bad" time to the phone. (Unless I'm missing something.)
--jhawk@mit.edu
John Hawkinson
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.
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.
Look at the line below the live time display. The data elements are the AM/PM indicator, a colored dot indicating the quality of the phone's time vs NTP time, the local time zone, followed by the offset. It's explained on the instructions that come up when you click the info button in the lower right corner of the screen.
Sent from my iPhone
On Mar 18, 2017, at 16:43, Jeremy Nichols jn6wfo@gmail.com wrote:
I have that app but don't see an option to display "the offset between NTP time and the device's internal clock." Please guide me.
Jeremy
On 3/18/2017 9:37 AM, Glen Hoag wrote:
On my iPhone, I run an NTP client, Emerald Time, that displays the offset between NTP time and the device's internal clock.
I'm on T-Mobile US and the offset is typically in the low tens of milliseconds or better.
It's certainly accurate enough as a clock where all I'm looking at is one minute resolution.
Glen Hoag
hoag@hiwaay.net
Sent from my iPhone
On Mar 18, 2017, at 09:52, John Hawkinson jhawk@MIT.EDU wrote:
Chris Albertson albertson.chris@gmail.com wrote on Fri, 17 Mar 2017
at 14:38:17 -0700 in CABbxVHsRa41HoB=xu4nk4T_c39uh2BCOMu17Nb7H3rA8uHXLQg@mail.gmail.com:
AndroiTS GPS Test (V 1.48 Free) is good, but a battery hog I find. On
THIS is why the phones don't really track time so well. Not that they
can't but doing so requires battery power.
This statement doesn't seem to be well-supported. I think it's
basically untrue if we're talking about timing at the tens of
millisecond level. Anything more precise seems relatively useless in
a smartphone without specialized mechanisms to get the time off of the
phone.
A phone's GPS receiver takes a lot of battery. But GPS is not the primary
mechanism that phones use get their time.
They get time from the cellular phone network (whether from the layer
two timing information or at a higher layer with something like
NTP). The effort required to keep a phone's clock in sync, even with a
really bad local oscillator, is lost in the noise of all the other
things the phone has to do. It's just not a battery issue.
The only reason modern smartphones keep bad time is because their
designers can't be bothered to do better, or possibly the network is
providing "bad" time to the phone. (Unless I'm missing something.)
--jhawk@mit.edu
John Hawkinson
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.
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.
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.
From: Jeremy Nichols
I have that app but don't see an option to display "the offset between
NTP time and the device's internal clock." Please guide me.
Jeremy
Jeremy,
Top right corner of the display. My iPad Pro 9.7 shows +0.003 at the
moment.
See:
https://itunes.apple.com/gb/app/emerald-time/id290384375?mt=8
where it's -0.001
SatSignal Software - Quality software written to your requirements
Web: http://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv
Ahh, I got the wrong app. Now, with the right app, my iPad shows -.001 and
my iPod Touch shows +0.216. Thank you!
Jeremy
On Sun, Mar 19, 2017 at 7:01 AM David J Taylor <
david-taylor@blueyonder.co.uk> wrote:
From: Jeremy Nichols
I have that app but don't see an option to display "the offset between
NTP time and the device's internal clock." Please guide me.
Jeremy
Jeremy,
Top right corner of the display. My iPad Pro 9.7 shows +0.003 at the
moment.
See:
https://itunes.apple.com/gb/app/emerald-time/id290384375?mt=8
where it's -0.001
SatSignal Software - Quality software written to your requirements
Web: http://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv
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.
--
Sent from Gmail Mobile
As noted, the device clock’s offset from the current best estimate of correct time is on the second line at the far right. The line that begins with “24”, “AM” or “PM.”
If you tap on the offset, it will trigger a re-sync of NTP time. The offset can be based on a single NTP server, so it may bounce around by several ms each time you re-sync.
-Denny
On Sun, Mar 19, 2017 at 11:03 AM, Jeremy Nichols jn6wfo@gmail.com wrote:
Now, with the right app, my iPad shows -.001 and
my iPod Touch shows +0.216.
It's marginally interesting (or time nutty) to listen to the per second
ticks that can be enabled. If you have (at least) two devices you can
listen to remote (bluetooth, airplay) audio delay.