Tim Watts used his keyboard to write :
And if you install a proper NTP client, it will try to maintain a drift
correction factor between NTP time and the local (kernel software) clock
which improves accuracy between NTP remote queries.

I used to use one back in the 1980's, but a sync every thirty seconds
is more than enough.

--
Regards,
Harry (M1BYT) (L)
http://www.ukradioamateur.co.uk

Harry Bloomfield wrote:
Tim Watts used his keyboard to write :
And if you install a proper NTP client, it will try to maintain a drift
correction factor between NTP time and the local (kernel software) clock
which improves accuracy between NTP remote queries.

I used to use one back in the 1980's, but a sync every thirty seconds is
more than enough.

I am fairly sure that somwehere in all of this an ntp daemon 'works out'
how much the real time clock is drifting between time server updates and
adjusts the kernel clock periodically In Linux, anyway.


"ntpd uses the adjtime(2) system call to correct the local system time
without causing time jumps. Adjustments of 32ms and greater are logged
using syslog(3). The threshold value is chosen to avoid having local
clock drift thrash the log files. Should ntpd be started with the -d or
-v option, all calls to adjtime(2) will be logged.

After the local clock is synchronized, ntpd adjusts the clock frequency
using the adjfreq(2) system call to compensate for systematic drift."

(man ntpd)

Isnt it wonderful to have an operating system where all this is just
DESIGNED IN from the word go..


--
Ineptocracy

(in-ep-toc-ra-cy) €“ a system of government where the least capable to
lead are elected by the least capable of producing, and where the
members of society least likely to sustain themselves or succeed, are
rewarded with goods and services paid for by the confiscated wealth of a
diminishing number of producers.

On Tue, 21 Aug 2012 14:17:52 +0100, John Williamson
wrote:



For a cheap, quick and accurate time signal, use a GPS receiver's time.
This will be UTC accurate to within the speed of light delay from LEO,
which is of the order of 10 milliseconds. Clever software can compensate
for these delays, if you want to be *really* fussy.


I was reading about doing exactly that last night. And pondering why the
time on my GPS depends on me entering it manually?

--
Rod

On 21/08/2012 20:56, Harry Bloomfield wrote:

Amazingly, I have never known it fail to manage to gain sync in the 7
years I have worn it - thinking back to when I built a large specialised
MSF receiver and how I struggled to get stable reception. lol

I remember building a clock from a kit which used the Rugby signal.
Reception was never a problem though, since we lived about 5 miles from
it :-)

John Williamson wrote:
The time reference point is the end of the last pip. Technically, the

No, it's the start of the last pip.

Each pip starts at the x.00 second point, at 55.00, 56.00, 57.00,
58.00, 59.00 and 00.00. The first five five are 0.1s long, the last
is 0.5s long. The reference point is the start of each pip.

The pips are calibrated so that they are UTC accurate 160km from
the London transmitter so averaging out the light-speed lag
across thr country.

JGH

On Tue, 21 Aug 2012 16:48:46 +0100, Percy wrote:

As I write my Windows 7 PC is within 500 milliseconds of UTC.

Mine probably is too. Once I installed W7 and let its own time checker
get on with it, I no longer needed to be arsed with the various time
checkers I'd used over the years - some of them quite good though. The
best was Dimension 7, which also acted as a timekeeper for other PCs
on my network.

En el artículo ,
Clive George escribió:

I remember building a clock from a kit which used the Rugby signal.
Reception was never a problem though, since we lived about 5 miles from
it :-)

I got one off Ebay a couple of years ago. it works, but is sensitive to
RFI interference - if I put it next to the TV for instance, it loses the
time signal.

The 'Rugby' signal is now transmitted from somewhere in Cumbria IIRC.

--
(\_/)
(='.'=)
(")_(")

wrote:
John Williamson wrote:
The time reference point is the end of the last pip. Technically, the

No, it's the start of the last pip.

Each pip starts at the x.00 second point, at 55.00, 56.00, 57.00,
58.00, 59.00 and 00.00. The first five five are 0.1s long, the last
is 0.5s long. The reference point is the start of each pip.

So the story we were told about the last pip being longer to account for
the difference between UTC and local time when they changed it from the
standard length was a load of rubbish?

The BBC get it wrong again...

The pips are calibrated so that they are UTC accurate 160km from
the London transmitter so averaging out the light-speed lag
across thr country.

That wouldn't surprise me.

--
Tciao for Now!

John.

On Tue, 21 Aug 2012, "John Williamson" writ:

Percy wrote:
On Tue, 21 Aug 2012, "John Williamson" writ:

Radio 4 pips on FM are accurate to a few milliseconds, depending on
how far you are from the radio.
How can this be if the final "pip" is 500 milliseconds long?

The time reference point is the end of the last pip. Technically, the
last zero crossing of the signal. I can set a digital watch to within
less than a frame of video with practice using the pips.

OK, understood. I'm amazed at your reaction time!

For a cheap, quick and accurate time signal, use a GPS receiver's
time. This will be UTC accurate to within the speed of light delay
LEO, which is of the order of 10 milliseconds. Clever software can
compensate for these delays, if you want to be *really* fussy.

How many GPS devices show the time to anything less than a second?

None that I know of, but the time changes on the display and the serial
output at the same time as the reference signal does. You can use a PLL
to synchronise any time point you wish within that second.

OK, thanks for the explanation.

--
P

On Tue, 21 Aug 2012 16:16:48 -0700 (PDT),
wrote:

John Williamson wrote:
The time reference point is the end of the last pip. Technically, the

No, it's the start of the last pip.

That's what I'd have thought, since the gap between pips seems to be
the same. Were it the end of the 'long' pip, it would have to start
sooner.

--
Frank Erskine

On Tue, 21 Aug 2012 18:27:21 +0100, The Natural Philosopher wrote:

wrote:
On Tuesday, August 21, 2012 2:57:05 PM UTC+1, The Natural Philosopher wrote:

if you have a windows NTP client installed or windows does this by default,
you should be with 10-20ms of 'correct'

Sadly not on Windows - see http://support.microsoft.com/kb/939322

'The W32Time service cannot reliably maintain sync time to the range of 1 to 2 seconds. Such tolerances are outside the design specification of the W32Time service.'

A 3rd part app ought to behave somewhat better.

Mathew
so insatll something better like this

http://www.timetools.co.uk/time-servers/ntp-client.htm

or Tclock2 (XP - don't know about W7). It can be used to mod. the Start
button and replace apps' buttons on the Task Bar with icons, show a clock,
day, date and an undocumented feature is that adding MM to the clock's
display shows RAM left as well.
http://www.softpedia.com/get/Desktop.../Tclock2.shtml
--
Peter.
The gods will stay away
whilst religions hold sway

I'd have though gps, though delayed was pretty accurate as otherwise its
not going to work. The other digital platforms are very inaccurate. There
are still time standard transmissions on short wave of course but any
reception frequency depends on propogation at the time of day needed.
Ther are supposed to be some frequency shifted mod signals on the Long
Wave stations run by the bbc and some others. There is the time standards in
the UK and Germany used by radio clocks as well, but these are on a very low
frequency and I've not seen any direct display receivers about at affordable
costs.

Brian

--
--
From the sofa of Brian Gaff -

Blind user, so no pictures please!
"Moonraker" wrote in message
...
How accurate is the time given on Windows? Also how about digital radio
and freesat, I guess there is a time delay at least on the last 2. I
notice that the pips are rarely used these days, and apart from the
Olympics I cannot remember the last time I heard Big Ben broadcast.
--
Residing on low ground in North Staffordshire

polygonum wrote :
On Tue, 21 Aug 2012 14:17:52 +0100, John Williamson
wrote:



For a cheap, quick and accurate time signal, use a GPS receiver's time.
This will be UTC accurate to within the speed of light delay from LEO,
which is of the order of 10 milliseconds. Clever software can compensate
for these delays, if you want to be *really* fussy.


I was reading about doing exactly that last night. And pondering why the time
on my GPS depends on me entering it manually?

I cannot think why it would need that. The GPS data received each and
every second, contains a full set of data for time and date. What GPS
satnav's sometimes need is some help to work out where they are if they
have been off for a while and their table of where the satellites
should be has become out of date, or it has been moved a large distance
to a new location whilst off - leave it to its own devices for a few
minutes and it will manage to sort itself out anyway.

--
Regards,
Harry (M1BYT) (L)
http://www.ukradioamateur.co.uk

Mike Tomlinson wrote on Aug 22, 2012:

En el artículo ,
Clive George escribió:

I remember building a clock from a kit which used the Rugby signal.
Reception was never a problem though, since we lived about 5 miles from
it :-)

I got one off Ebay a couple of years ago. it works, but is sensitive to
RFI interference - if I put it next to the TV for instance, it loses the
time signal.

The 'Rugby' signal is now transmitted from somewhere in Cumbria IIRC.



Many wris****ches do that. Mine synchronises with the transmitter in Anthorn,
Cumbria every 24 hours. If that fails it tries the one in Mainflingen,
Germany.

--
Mike Lane
UK North Yorkshire
mike_lane at mac dot com

On Tue, 21 Aug 2012 17:52:34 +0100, Nick Odell
wrote:

On Tue, 21 Aug 2012 17:23:22 +0100, Graham.
wrote:


Big Ben on FM or AM radio is sort of accurate to a second or so. It can
be heard most days on Radio 4 at 5 P.M., as well as other times.
Occasionally, to their great embarrassment, you can hear the announcer
talk over the start of the chimes.

That's standard procedure now, and has been for a year or so.

Certainly at 1800 and midnight, (Pips at 1700*, not Big Ben)
the intention was, barring accident to broadcast all the bongs
uninterrupted.

Now, the standard practice is to start reading the news headlines
after the first bong and fade them out after the third.

In both the above schemes the silence between the end of the forth
quarter chime and the first chime is filled with a station ID, time
check and "Good evening".

At midnight on New Year's Eve all 12 bongs are played in full, thank
goodness, but they still fill that poignant pause with an
announcement.
Total sacrilege IMHO.

* The 1700 pips on weekdays is followed by the start of "PM" which
used to have a theme tune, which was dropped around the time of 9-11
and never re-instated. Some of us die hard R4 listeners imagine we can
still hear it.


I think it's considered okay to talk over the chimes of BB after the
first couple because they go on so looooong. "Crashing the Pips" is a
no-no.

Announcers will talk over absolutely anything nowadays.
--
(\__/) M.
(='.'=) If a man stands in a forest and no woman is around
(")_(") is he still wrong?

On Tue, 21 Aug 2012 21:29:30 +0100, Tim Watts
wrote:

Harry Bloomfield wrote:

Mike Barnes explained :
Not especially. My everyday timepiece is an MSF wall clock, which agrees
to the second with the pips on Radio 4 FM. I just watched the minute
click over on my Windows taskbar, and according to the wall clock it was
a couple of seconds late.

That is because Windows syncs itself by default once per day at a set
time of day. If your PC is not turned on at that time, it will not sync
itself, so it then has to rely upon its internal (not very accurate)
RTC.

You can download a free utility which allows you to set how often (how
many minutes) Windows should attempt to sync itself, from every few
minutes to once per day. Mine is set to sync itself every 30 minutes
and always agrees with my radio controlled watch.


And if you install a proper NTP client, it will try to maintain a drift
correction factor between NTP time and the local (kernel software) clock
which improves accuracy between NTP remote queries.

The Windows Time service is a NTP client and can be configured to set
the clock accurately.
--
(\__/) M.
(='.'=) If a man stands in a forest and no woman is around
(")_(") is he still wrong?

On Tue, 21 Aug 2012 16:30:52 +0100, PeterC
wrote:

On Tue, 21 Aug 2012 15:27:20 +0100, tony sayer wrote:

In article , Mike Tomlinson
scribeth thus
In article , Moonraker
writes

How accurate is the time given on Windows?

Depends on how accurate your motherboard's clock is. If the CMOS
battery is running low, it will lose time.

Windows can set the time from an Internet time server. On XP, the
procedure is to right-click on the clock, Adjust Date/Time, Internet
Time, enter 0.uk.pool.ntp.org into the box, then click Update Now. OK
your way out and Windows will set the time from a server regularly. Of
course, you need to be online for it to work.


Look up Atomic.exe small prog that pings a ntp server every day.

Its also there in Windows 7 ..

Then take into account the effects of some software. The version of Opera
that I'm running atm causes the clock to gain by about 5s between the
2-minute-interval ping of npl.

Please explain how it does this.
--
(\__/) M.
(='.'=) If a man stands in a forest and no woman is around
(")_(") is he still wrong?

The Natural Philosopher formulated the question :
Harry Bloomfield wrote:
Tim Watts used his keyboard to write :
And if you install a proper NTP client, it will try to maintain a drift
correction factor between NTP time and the local (kernel software) clock
which improves accuracy between NTP remote queries.

I used to use one back in the 1980's, but a sync every thirty seconds is
more than enough.

I am fairly sure that somwehere in all of this an ntp daemon 'works out' how
much the real time clock is drifting between time server updates and adjusts
the kernel clock periodically In Linux, anyway.


"ntpd uses the adjtime(2) system call to correct the local system time
without causing time jumps. Adjustments of 32ms and greater are logged using
syslog(3). The threshold value is chosen to avoid having local clock drift
thrash the log files. Should ntpd be started with the -d or -v option, all
calls to adjtime(2) will be logged.

After the local clock is synchronized, ntpd adjusts the clock frequency using
the adjfreq(2) system call to compensate for systematic drift."

(man ntpd)

Isnt it wonderful to have an operating system where all this is just DESIGNED
IN from the word go..

PC's actually use two clocks. A battery backed CMOS RTC clock which
maintains the time whilst the system is off. When the PC is turned on,
the data is transferred into memory, where the PC itself then does the
counting to maintain that separate from the hardware clock's time
interrupt driven.

So there are two possibilities for drifting accuracy - the CMOS RTC's
crystal and the CPU's crystal whilst the PC is on and running. A time
sync adjustment has to correct both of these clocks.

I use this VB script to alter the time interval between my PC syncing
itself via the time servers.

You just need to run it once (ever), to make permanent changes to the
sync interval...

'xp_time_sync.vbs - Change the Internet Time Update Interval
'© Doug Knox - revised - 5/10/2002
'This code may be freely distributed/modified
'Downloaded from www.dougknox.com
'Thanks to Gregory Phillips for catching an error when clicking Cancel.

Option Explicit
On Error Resume Next

'Declare variables
Dim WSHShell, p1, p2, cn, newtime, mycheck, ev, X, Y

'Set the Windows Script Host Shell and assign values to variables
Set WSHShell = WScript.CreateObject("WScript.Shell")

p1 =
"HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servi ces\W32Time\Config\UpdateInterval"
p2 =
"HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Servi ces\W32Time\TimeProviders\NtpClient\SpecialPollInt erval"

'This section writes the correct values to the Registry

ev = WSHShell.RegRead(p2)
ev = ev / 3600

cn = ""

Do While IsNumeric(CN) = False

cn = InputBox("Enter the number of hours"& vbCR & "between Internet
Time updates" & vbCR & vbCR & "The current Setting in hours is: " & ev
& vbCR & vbCR & "The default value is 168 hours (7 days).","Value
Entry","24")

If IsNumeric(cn) = False Then
MsgBox "Please enter a number!",4096,"Error!"
End If

If cn = "" Then
Exit Do
End If

Loop

If cn "" AND cn 0 Then

X = InStr(cn,".")
cn = Left(cn,X+2)

newtime = cn * 3600
WSHShell.RegWrite p1, newtime, "REG_DWORD"
WSHShell.RegWrite p2, newtime, "REG_DWORD"
MsgBox "The Internet Time update interval has been changed." & vbCR &
"Reboot your computer for the change to take effect.",4096,"Finished"

Else

If cn = "" Then
MsgBox "The Internet Time update interval has NOT been
changed.",4096,"Cancelled"
Elseif cn = 0 Then
MsgBox "The Internet Time Update interval has NOT been changed." &
vbCR & "The value must be greater than 0 (zero).",4096,"Cancelled"
End If

End If

--
Regards,
Harry (M1BYT) (L)
http://www.ukradioamateur.co.uk

Clive George used his keyboard to write :
On 21/08/2012 20:56, Harry Bloomfield wrote:

Amazingly, I have never known it fail to manage to gain sync in the 7
years I have worn it - thinking back to when I built a large specialised
MSF receiver and how I struggled to get stable reception. lol

I remember building a clock from a kit which used the Rugby signal. Reception
was never a problem though, since we lived about 5 miles from it :-)

From a small company by the name of something like Cambridge Kits?

--
Regards,
Harry (M1BYT) (L)
http://www.ukradioamateur.co.uk

Brian Gaff explained :
Ther are supposed to be some frequency shifted mod signals on the Long Wave
stations run by the bbc and some others. There is the time standards in the
UK and Germany used by radio clocks as well, but these are on a very low
frequency and I've not seen any direct display receivers about at affordable
costs.

The radio clocks and watches receive the VLF time standards.

--
Regards,
Harry (M1BYT) (L)
http://www.ukradioamateur.co.uk

On Tue, 21 Aug 2012 20:51:33 +0100, Andy Burns wrote:

Ramsman wrote:

I noticed during the BBC Olympic broadcasts that the stadium clock was
about seven seconds behind my watch.

If you were in the stadium I expect the clock would have agreed with
your watch, most of the the delay if watching on digital TV will come
from compressing it before transmission and decompressing it after
receiving to display the picture.

Walking around the TV section of a large department store recently, SWMBO
noticed that the TVs all seemed to be showing the same channel at
slightly different times - some were a good second out.

Jethro_uk used his keyboard to write :
On Tue, 21 Aug 2012 20:51:33 +0100, Andy Burns wrote:

Ramsman wrote:

I noticed during the BBC Olympic broadcasts that the stadium clock was
about seven seconds behind my watch.

If you were in the stadium I expect the clock would have agreed with
your watch, most of the the delay if watching on digital TV will come
from compressing it before transmission and decompressing it after
receiving to display the picture.

Walking around the TV section of a large department store recently, SWMBO
noticed that the TVs all seemed to be showing the same channel at
slightly different times - some were a good second out.

That is due to the TV's internal delays in decoding the digital signal
plus maybe differences in the time the TV's were powered up (not sure
on that last).

--
Regards,
Harry (M1BYT) (L)
http://www.ukradioamateur.co.uk

On 22/08/2012 08:25, Brian Gaff wrote:
I'd have though gps, though delayed was pretty accurate as otherwise its
not going to work. The other digital platforms are very inaccurate. There
are still time standard transmissions on short wave of course but any
reception frequency depends on propogation at the time of day needed.
Ther are supposed to be some frequency shifted mod signals on the Long
Wave stations run by the bbc and some others. There is the time standards in
the UK and Germany used by radio clocks as well, but these are on a very low
frequency and I've not seen any direct display receivers about at affordable
costs.

It isn't hard to make one and they are standard in the small weather LCD
display monitor devices sold in the likes of Aldi, Lidl and Maplin.

The hard part is buying the bits in one off quantities for a sensible
price. The fully built products are from about £10 on Amazon. You are
probably better off buying one and taking it apart for the modules.

http://www.amazon.co.uk/s/?ie=UTF8&k...376axjd3b2_ b

I don't know of any cheap kits these days although you can slug an old
LW receiver with 2.2x the standard capacitor in parallel to move the
tuning so that 60kHz sits where 200kHz was marked on the dial.

Regards,
Martin Brown

On 22/08/2012 09:01, Mark wrote:
On Tue, 21 Aug 2012 16:30:52 +0100, PeterC
wrote:

On Tue, 21 Aug 2012 15:27:20 +0100, tony sayer wrote:

In article , Mike Tomlinson
scribeth thus
In article , Moonraker
writes

How accurate is the time given on Windows?

Depends on how accurate your motherboard's clock is. If the CMOS
battery is running low, it will lose time.

Or gain it. The only thing guaranteed is that there is a small
dependence on temperature and a smaller one on battery voltage. Most are
not properly trimmed and so drift +/- 15s/day worst case.

Some time sync software will show you the log of adjustments.

Windows can set the time from an Internet time server. On XP, the
procedure is to right-click on the clock, Adjust Date/Time, Internet
Time, enter 0.uk.pool.ntp.org into the box, then click Update Now. OK
your way out and Windows will set the time from a server regularly. Of
course, you need to be online for it to work.


Look up Atomic.exe small prog that pings a ntp server every day.

Its also there in Windows 7 ..

Then take into account the effects of some software. The version of Opera
that I'm running atm causes the clock to gain by about 5s between the
2-minute-interval ping of npl.

Please explain how it does this.

Who knows. It is more common for old games that disable PC interrupts
for long periods of time to make the main system clock run slow.

The typical daily drift on an untrimmed RTC chip from a 32kHz xtal is
about 15s or worse if the designers fail to put the right loading
capacitors round it. Depends a bit on ambient temperature.

I do find it annoying that some very expensive computer based kit does
not have correctly trimmed xtal RTC built in. The manufacturers fix for
this was to add GPS so that the unit can get a correct time as it boots!

Regards,
Martin Brown

Percy writes:

On Tue, 21 Aug 2012, "John Williamson" writ:
€¦ I can
set a digital watch to within less than a frame of video
with practice using the pips.

OK, understood. I'm amazed at your reaction time!

Not reaction time. People are much better at hitting things to
time when they know the beat in advance than they are at
reacting. An undergraduate teaching (ie not serious) experiment
tried to measure reaction times using a regular flashing light.
I could hit the button within a few milliseconds of the light
coming on once I got the rhythm.

--
Jón Fairbairn
http://www.chaos.org.uk/~jf/Stuff-I-dont-want.html (updated 2010-09-14)

In article , Mark i@dontget
lotsofspamanymore.invalid writes

The Windows Time service is a NTP client and can be configured to set
the clock accurately.

To be more accurate, it's a SNTP (simple NTP) client which not all NTP
servers support.

--
(\_/)
(='.'=)
(")_(")

Mike Tomlinson wrote:

In article , Mark i@dontget
lotsofspamanymore.invalid writes

The Windows Time service is a NTP client and can be configured to set
the clock accurately.

To be more accurate, it's a SNTP (simple NTP) client which not all NTP
servers support.


Which sounds like its dumb and may not attempt to apply linear corrections
to the system clock, rather just relying on resetting the clock
periodically.


--
Tim Watts

In article , Tim Watts
writes

Which sounds like its dumb and may not attempt to apply linear corrections
to the system clock, rather just relying on resetting the clock
periodically.

Pretty sure it is, so anyone wanting greater accuracy could use one of
the NTP clients mentioned in this thread. Being able to specify more
than one server is also an advantage.

--
(\_/)
(='.'=)
(")_(")

In article ,
John Williamson wrote:
Radio 4 pips on FM are accurate to a few milliseconds, depending on how
far you are from the radio.

Only if you use analogue reception. On DAB or FreeView they are well out.
'Internet' reception too, I'd guess.

--
*A backward poet writes inverse.*

Dave Plowman London SW
To e-mail, change noise into sound.

In article ,
Moonraker wrote:
How accurate is the time given on Windows? Also how about digital radio
and freesat, I guess there is a time delay at least on the last 2. I
notice that the pips are rarely used these days, and apart from the
Olympics I cannot remember the last time I heard Big Ben broadcast.

Big Ben is broadcast several time a day on R4. But it is not accurate most
of the time.

--
*Don't squat with your spurs on *

Dave Plowman London SW
To e-mail, change noise into sound.

On Tuesday, August 21, 2012 2:31:49 PM UTC+1, John Williamson wrote:
Andrew May wrote:

On 21/08/2012 14:17, John Williamson wrote:



Big Ben on FM or AM radio is sort of accurate to a second or so. It can

be heard most days on Radio 4 at 5 P.M.,



I never get to hear the beginning of the PM programme. Do they have Big

Ben then? I know they have it at 6pm at the start of the news.



Possibly not, I was having trouble remembering which it was. I know I

hear it most evenings when I'm listening to Radio 4, though. ISTR they

play it at either 9 or 10 PM, too.



A couple of years ago Big Ben was being repaired so they had birdsong

instead. It wasn't as accurate.



Obviously. How d'you expect a bird to tell the time without a watch? :-)

Could tweet for it. ;-)




--

Tciao for Now!



John.

On Wed, 22 Aug 2012 10:14:35 +0100, Mike Tomlinson
wrote:

In article , Mark i@dontget
lotsofspamanymore.invalid writes

The Windows Time service is a NTP client and can be configured to set
the clock accurately.

To be more accurate, it's a SNTP (simple NTP) client which not all NTP
servers support.

AFAIK it's based on SNTP on Windows 2000 but NTP on later OSs.

The default settings are not great (check once per day and always use
time.windows.com). By configuring to check more frequently and by
using a better server, the accuracy can be pretty good.
--
(\__/) M.
(='.'=) If a man stands in a forest and no woman is around
(")_(") is he still wrong?

Harry Bloomfield wrote:
Jethro_uk used his keyboard to write :
On Tue, 21 Aug 2012 20:51:33 +0100, Andy Burns wrote:

Ramsman wrote:

I noticed during the BBC Olympic broadcasts that the stadium clock was
about seven seconds behind my watch.

If you were in the stadium I expect the clock would have agreed with
your watch, most of the the delay if watching on digital TV will come
from compressing it before transmission and decompressing it after
receiving to display the picture.

Walking around the TV section of a large department store recently,
SWMBO noticed that the TVs all seemed to be showing the same channel
at slightly different times - some were a good second out.

That is due to the TV's internal delays in decoding the digital signal
plus maybe differences in the time the TV's were powered up (not sure on
that last).

nope. Its purely how much buffering they do on the signal. Decoding a
compressed video always takes some time, and to an extent,the more time
it takes the better a job you can do.

Our Sonyt STBS are about a second and a half out of synch with 'real
time' and the DTB adapter on this computer is more like 15 seconds.

The wifes DTB on her Mac is about 5 seconds delayed


--
Ineptocracy

(in-ep-toc-ra-cy) €“ a system of government where the least capable to
lead are elected by the least capable of producing, and where the
members of society least likely to sustain themselves or succeed, are
rewarded with goods and services paid for by the confiscated wealth of a
diminishing number of producers.

On Tuesday, August 21, 2012 10:32:07 PM UTC+1, The Natural Philosopher wrote:

[ntpd]
Isnt it wonderful to have an operating system where all this is just
DESIGNED IN from the word go..

You say designed in but ntpd is not a built-in part of any 'operating system' but rather a 3rd party tool that is optionally installed. Few, if any, distributions install ntpd as part of their base install, not least because the benefits of its use are severely limited on machines that don't stay continuously on.

If anything it is more common to find ntpdate installed and is usually invoked once at startup. Now that ntpd can perform these 'one-shot' time sets ntpdate will eventually be retired however so we may start to see ntpd installed by default but perhaps not automatically run as a daemon.

Mathew

On Wednesday, August 22, 2012 11:27:54 AM UTC+1, The Natural Philosopher wrote:
Harry Bloomfield wrote:

Jethro_uk used his keyboard to write :

On Tue, 21 Aug 2012 20:51:33 +0100, Andy Burns wrote:



Ramsman wrote:



I noticed during the BBC Olympic broadcasts that the stadium clock was

about seven seconds behind my watch.



If you were in the stadium I expect the clock would have agreed with

your watch, most of the the delay if watching on digital TV will come

from compressing it before transmission and decompressing it after

receiving to display the picture.



Walking around the TV section of a large department store recently,

SWMBO noticed that the TVs all seemed to be showing the same channel

at slightly different times - some were a good second out.



That is due to the TV's internal delays in decoding the digital signal

plus maybe differences in the time the TV's were powered up (not sure on

that last).



nope. Its purely how much buffering they do on the signal. Decoding a

compressed video always takes some time, and to an extent,the more time

it takes the better a job you can do.



Our Sonyt STBS are about a second and a half out of synch with 'real

time' and the DTB adapter on this computer is more like 15 seconds.



The wifes DTB on her Mac is about 5 seconds delayed

I heard, during the Olmypic refering to Usain Bolt that it takes a human 0.1 second to respond to a sound, so that makes all clocks with sound as indication out by 100ms according to the true time, unless you use software to 'pip' 100ms beforehand the true time is reached.







--

Ineptocracy



(in-ep-toc’-ra-cy) – a system of government where the least capable to

lead are elected by the least capable of producing, and where the

members of society least likely to sustain themselves or succeed, are

rewarded with goods and services paid for by the confiscated wealth of a

diminishing number of producers.

wrote:
On Tuesday, August 21, 2012 10:32:07 PM UTC+1, The Natural
Philosopher wrote:

[ntpd]
Isnt it wonderful to have an operating system where all this is
just DESIGNED IN from the word go..

You say designed in but ntpd is not a built-in part of any 'operating
system' but rather a 3rd party tool that is optionally installed.

well in that case you could argue that nothing except the linux kernel
is part of linux.

Not even X-windows or a window manager. Nothing. Not even bash or sh.

In fact ALL of linux is '3rd party' because there is no 'Linux company'
that 'supplies one monolithic product'


Few, if any, distributions install ntpd as part of their base
install, not least because the benefits of its use are severely
limited on machines that don't stay continuously on.


hardly.

If anything it is more common to find ntpdate installed and is
usually invoked once at startup. Now that ntpd can perform these
'one-shot' time sets ntpdate will eventually be retired however so we
may start to see ntpd installed by default but perhaps not
automatically run as a daemon.

whatever..
Mathew


--
Ineptocracy

(in-ep-toc-ra-cy) €“ a system of government where the least capable to
lead are elected by the least capable of producing, and where the
members of society least likely to sustain themselves or succeed, are
rewarded with goods and services paid for by the confiscated wealth of a
diminishing number of producers.

In article ,
The Natural Philosopher wrote:
That is due to the TV's internal delays in decoding the digital signal
plus maybe differences in the time the TV's were powered up (not sure
on that last).

nope. Its purely how much buffering they do on the signal. Decoding a
compressed video always takes some time, and to an extent,the more time
it takes the better a job you can do.

Anything digital is always behind real time. Due to the way it is sampled.
Even digital microphones, compared to analogue ones. Which makes it
impossible to use a mix of the two. Hence there are very few around - the
exception being some radio mics.

And any processing after that initial sampling may slow things down
further.

--
*If you don't pay your exorcist you get repossessed.*

Dave Plowman London SW
To e-mail, change noise into sound.

"Harry Bloomfield" wrote in message
. uk...
polygonum wrote :
On Tue, 21 Aug 2012 14:17:52 +0100, John Williamson
wrote:



For a cheap, quick and accurate time signal, use a GPS receiver's time.
This will be UTC accurate to within the speed of light delay from LEO,
which is of the order of 10 milliseconds. Clever software can compensate
for these delays, if you want to be *really* fussy.


I was reading about doing exactly that last night. And pondering why the
time on my GPS depends on me entering it manually?

I cannot think why it would need that.

Tomtoms rely on it to get the local time.
The sats broadcast in UTC.
There is an option to sync the local time to the sat time in the menus
somewhere but it doesn't change the hours AFAIK.
Satnavs only really need the local time to work out the likely traffic
conditions.

The GPS data received each and every second, contains a full set of data
for time and date. What GPS satnav's sometimes need is some help to work
out where they are if they have been off for a while and their table of
where the satellites should be has become out of date, or it has been
moved a large distance to a new location whilst off - leave it to its own
devices for a few minutes and it will manage to sort itself out anyway.

However it still will want you to set the local time.

on 22/08/2012, dennis@home supposed :
However it still will want you to set the local time.

My WinCE satnav come ICE system just needs you to tell it which time
band you are in, then sets the clock from RDS or maybe satnav.

--
Regards,
Harry (M1BYT) (L)
http://www.ukradioamateur.co.uk

"Dave Plowman (News)" wrote in message
...

Anything digital is always behind real time. Due to the way it is sampled.

You won't find much delay in the telephone network due to sampling.

You get delays when its being encoded and/or compressed like on mobile
phones or VoIP.

The telephone system uses a straight 8KHz sample rate with no additional
delays due to correction or compression.

Even digital microphones, compared to analogue ones. Which makes it
impossible to use a mix of the two. Hence there are very few around - the
exception being some radio mics.

And any processing after that initial sampling may slow things down
further.

The thing that will slow it down is bitstream D/A convertors if someone uses
them because they are cheaper than the alternative ladder convertors. But
only by a few milliseconds.