ukclique > tech.* > tech.broadcast

Johnny B Good (07.03.2019, 03:21)
Way back near the beginning of December last year, I started a very on
topic thread posing the question in the subject, giving the following
quote of the first paragraph of that post below:-

"I've been attempting to ascertain whether or not BBC local AM
transmitters are locked to an NPL frequency standard which I can trust as
a secondary calibration reference."

as the reason for my interest, explaining in great detail[1] as to why
I'd been hoping the local BBC Radio Merseyside MW transmitter would turn
out to be locked to a frequency reference of exactly 1485KHz as all the
evidence at my disposal at that time was strongly suggesting.

In the end, after investing 21 quid in a u-Blox NEO M8N GPS module for
use in a homebrewed GPSDO (along with another 3 quid for an active patch
antenna with a 5 metre cable), I finally ascertained that my local BBC
station on 1485000Hz was actually transmitting on a frequency of
1484999.89810259Hz, some 18.974183mHz down (or 12.777227ppb +/-0.18ppb)
from its specified frequency.

This offset is so close to my best estimate of 13ppb about a month ago,
it does look very much like they're using a GPSDO with a 12.8ppb offset
dialled in as this TX's frequency standard. The +/- 0.18ppb measurement
error is a limitation of my measurement technique and possibly some
variation in the escapement rate of my stopwatch.

When it came to comparing the GPS module's 198000Hz output on the PPS
pin against the R4LW broadcast by way of a sanity check of my GPS
reference, there was no beat frequency to measure. Indeed, in the space
of 90 minutes, there wasn't even any perceivable phase shift between the
two thus neatly removing all doubts I may have harboured over the
accuracy of the R4LW transmitter as a frequency reference. I sanity
checked this zero beat/no phase drift finding by counting 60 beats in
exactly one minute +/-0.1s for offsets at +/- 1Hz of the frequency (the
PPS line can only be programmed in 1Hz steps).

Incidentally, this test also proved that the raw output of the PPS pin
is a trustworthy reference in its own right regardless of whatever it's
programmed to output over the 1Hz to 12MHz range. The issue of jitter for
frequencies that aren't integer ratios of the built in 48MHz VCTCXO
doesn't impact on its long term accuracy but nevertheless it's still
desirable to filter out such effects from its final output.

Put it this way, if you're looking for a superior alternative to the
10MHz WWV broadcasts as a calibration standard, the raw 10MHz output of
these u-Blox NEO M8N GPS modules is a vast improvement over the 10MHz WWV
broadcasts which might eventually become receivable in the UK when the
Sunspot activity starts to peak again in the next 5 years or so (assuming
they're still operational by then).

As for my original question, since purchasing the GPS module, the answer
has become somewhat academic as far as I'm concerned. However, for anyone
else with a more than passing interest in recalibrating their one and
only signal generator, not already possessed of a reliable frequency
standard at a conveniently higher frequency than the R4LW transmitter,
the local BBC Radio Merseyside Tx on 1485KHz (and probably a few other
BBC Local MW Radio transmitters as well) is close enough to exact
frequency to allow an initial calibration to be performed even for kit
using a 0.1ppm TCXO reference (a sub 13ppb offset is well within the
typical tolerance range of accuracy for a 0.1ppm TCXO).

[NOTES]

[1] For anyone interested (Warning! This is VERY ON TOPIC), here are the
rest of those details:-

Some twenty years ago, when I did have some form of 80 metre dipole
strung up, I was able to use the 10MHz WWV frequency and time standard
broadcast to calibrate and maintain the frequency accuracy of my, then
newly acquired, Kenwood TS140S HF transceiver. Even so, reception was
rather variable to say the least but it was good enough to allow me to
calibrate to within +/-1Hz (about an order of magnitude better than the
10Hz tuning steps of the radio's PLL synthesised tuner).

Now I'm hoping to calibrate a 50MHz 0.1ppm TCXO which I've used to
upgrade the crappy 100ppm XO chip used by Feeltech in their extremely
cheap FY6600-60M dds arbitrary wave generator[2].

Try as I might with a makeshift wire aerial dangling out of my 1st floor
'office' window, I can detect no sign of the WWV broadcasts on any of
their assigned frequencies (2.5, 5, 10, 15 and 20MHz) nor can I find any
of the Russian MF/HF frequency/time standards broadcasts to use as an
alternative calibration source.

The only strong, and seemingly frequency stable, signals in the more
useful MHz range that I've been able to tune into have been Radio
Merseyside on 1485KHz and the slightly weaker Radio City on 1548KHz[4].

Working on the basis that a BBC local am radio station is more likely to
be locked to an NPL reference than a commercial station [see note 4],
I've been monitoring Merseyside's carrier frequency with the AWG set
precisely to 1485.0000000000KHz[3] to time the beat frequency period with
a stopwatch.

Initial frequency differences (expressed in periods of 6 to 18 seconds)
strongly hinted that Radio Merseyside's 1485KHz am transmitter may
actually be synchronised to an NPL standard but I haven't been able to
track down any information to confirm this hypothesis.

Since I've not been able to find anything more technically interesting
than the fact that the transmitter is a 2KW unit operating on the medium
wave frequency of 1485KHz, I thought I might have better luck prevailing
upon this group's expertise in the matter of such broadcasting
technicalities.

Basically, do any of the BBC's local AM transmitters utilise an NPL
reference to stabilise their frequency?

[2] In case anyone's interested, it was just a tad over 75 quid from a
Chinese supplier shipping from a London based warehouse - it took less
than a week from being ordered to being delivered.

[3] Yes, a whole 14 digits worth of resolution! Kinda makes a mockery of
the 100ppm oscillator reference it had originally been cursed with, hence
the oscillator upgrade.

Even with a 0.1ppm reference clock, that's still a good 4 digits in
excess of the best practical accuracy - it's not a fixed decimal point
display either, you can choose frequency units of μHz, mHz, Hz, KHz and
MHz and set any frequency from zero to 60MHz to an 'accuracy' of 14
digits.

[4] Actually, the exact frequency, within plus or minus a 0.18ppb of
error, turns out to be 1547995.7143Hz. This is some 4.2857Hz down on its
assigned frequency, just over twice the +/-2Hz tolerance allowed by the
DTI of old (but, unsurprisingly, allowed by the current toothless
watchdog known as Ofcom who are now charged with maintaining this rather
lax frequency standard). Yet another good reason for my referring to this
station as "Radio Shoddy" when in polite society (or, more bluntly and
onomatopoeically, "Radio Sh*tty").
bilou (07.03.2019, 12:42)
Johnny B Good avait prétendu :
> Way back near the beginning of December last year, I started a very on
> topic thread posing the question in the subject, giving the following
> quote of the first paragraph of that post below:-


Many thanks for this very informative feedback.
It seems that GPS frequency reference is the future.
This is still a mystery to me as satellites are
highly affected by the doppler effect :-)
Brian Gaff (07.03.2019, 16:54)
I was thinking the same, but since they presumably know their clock speed
being based on an atomic clock this can be compensated for by having several
sats in view and working out their speed relative to the receiver.

Brian
Johnny B Good (08.03.2019, 03:00)
On Thu, 07 Mar 2019 14:54:47 +0000, Brian Gaff wrote:

> I was thinking the same, but since they presumably know their clock
> speed being based on an atomic clock this can be compensated for by
> having several sats in view and working out their speed relative to the
> receiver.


I'm not privy to the exact details of the workings of GPS (and GLONASS
and Galileo et al) but it's not rocket science, just basically clever
navigation/position fixing computations based on the satellite borne
Caesium atomic clocks, accurately synchronised to the ground based atomic
reference clocks with nanosecond (or better) precision which all keeps
precise track of each satellite’s orbital velocity with respect to the
various ground stations involved in their day to day operation.

The Doppler effect is part and parcel of the ephemeris data set and as
well as being accounted for also provides additional dynamic positional
fix data.

The satellites transmit extremely precise time data as well as an ident
signal and ephemeris data which the GPS receivers process to calculate a
positional fix by triangulation based on distances to each satellite
calculated from the relative transit times of the GPS clock data signals,
as well as maintaining their own internal clock, synchronised to within a
nanosecond (possibly better?). One nanosecond delay represents
approximately just under a foot of distance between the satellite and the
receiver so you'd certainly hope to get nanosecond timing accuracy or
better.

The US military certainly hope to so they can deliver their ordnance to
a target building via a conveniently placed window (especially important
when delivering just the fuel part of the explosive charge to mix with
the already conveniently present oxidant using the "air burst explosive
mix it on site" method based on the "Custard powder factory explosion
principle").

AFAIR, you need a minimum of four satellite locks to get a 3D fix. Four
or more, these days, should be enough to get a fix accurate to within ten
metres or better. If a military grade GPS with its own hyper accurate
clock reference were used, just three satellites would suffice for an
accurate fix. Using four or more satellites removes the need for such
high clock reference precision in the commodity class SatNav device.

Just how accurate a fix you can get with 15 to 20 satellite fixes, as I
often observe at this location, appears to be not much better with a
variation in latitude that equates to 3 or 4 metres up and down the
longitude line. The 3D fix typically shows an elevation around the 89 to
96 metre mark.

TBH, with some 15 to 20 satellites, I'd have expected more precision
than what the u-Blox control centre evaluation software is reporting.
Still, that's not particularly important to my needs for a rock steady
frequency standard so I'll put that down to my not fully understanding
all of the u-Blox control centre's configuration options.

The satellites are all in 12 hour trans-polar orbits at an altitude of
20,180 km (according to wikipedia), well clear of atmospheric drag and
visible over a much greater range than LEO satellites and less effected
by the well documented mascons that complicate the orbital mechanics of
their low flying cousins in LEO.

There's a lot more info on Wikipedia here:

<https://en.wikipedia.org/wiki/Global_Positioning_System>

for anyone interested in the gory details. :-)
J. P. Gilliver (John) (08.03.2019, 03:13)
In message <o8jgE.22766$HH5.16006>, Johnny B Good
<johnny-b-good> writes:
[]
> I'm not privy to the exact details of the workings of GPS (and GLONASS
>and Galileo et al) but it's not rocket science, just basically clever

[]
Er - I think if anything is the _definition_ of rocket science ... (-:
Johnny B Good (08.03.2019, 06:28)
On Fri, 08 Mar 2019 01:13:48 +0000, J. P. Gilliver (John) wrote:

> In message <o8jgE.22766$HH5.16006>, Johnny B Good
> <johnny-b-good> writes:
> []
> []
> Er - I think if anything is the _definition_ of rocket science ... (-:


The rocketry part of the system is merely the means of delivery to place
the sats in orbit. After that, it's essentially a matter of orbital
mechanics, extreme clock accuracy, wireless telegraphy and some automated
navigational computation. Rocket science is only the start, the rest is
way more complex than that. :-)
Andy Burns (08.03.2019, 10:41)
Johnny B Good wrote:

> One nanosecond delay represents
> approximately just under a foot of distance between the satellite and the
> receiver


<https://youtu.be/JEpsKnWZrJ8>
J. P. Gilliver (John) (08.03.2019, 19:44)
In message <ObmgE.23086$HH5.13260>, Johnny B Good
<johnny-b-good> writes:
>On Fri, 08 Mar 2019 01:13:48 +0000, J. P. Gilliver (John) wrote:
> The rocketry part of the system is merely the means of delivery to place
>the sats in orbit. After that, it's essentially a matter of orbital
>mechanics, extreme clock accuracy, wireless telegraphy and some automated
>navigational computation. Rocket science is only the start, the rest is
>way more complex than that. :-)

I know, I know (-:! I was just amused to see the expression used in this
context.

I've often wondered about the origin of the expression; I'm guessing it
involved in large part the size of the calculation required to do the
orbital mechanics in the early days of the space race. (Still present in
the 70s, where AIUI various amateur satellites got a "free ride" because
it was easier to launch some things with ballast, for which some of
which the satellites could be substituted, than to redo the
calculations.)
MB (08.03.2019, 20:01)
On 08/03/2019 17:44, J. P. Gilliver (John) wrote:
> I've often wondered about the origin of the expression; I'm guessing it
> involved in large part the size of the calculation required to do the
> orbital mechanics in the early days of the space race. (Still present in
> the 70s, where AIUI various amateur satellites got a "free ride" because
> it was easier to launch some things with ballast, for which some of
> which the satellites could be substituted, than to redo the calculations.)


According to the OED

1986 Chicago Tribune (Nexis) 24 Oct. 79 Nesmith says, half jokingly,
that ‘the tape manufacturing will be somewhere close to rocket science’
to cut costs.
J. P. Gilliver (John) (08.03.2019, 22:32)
In message <q5uaku$fc3$1>, MB <MB> writes:
>On 08/03/2019 17:44, J. P. Gilliver (John) wrote:
>According to the OED
>1986 Chicago Tribune (Nexis) 24 Oct. 79 Nesmith says, half
>jokingly, that ‘the tape manufacturing will be somewhere close to
>rocket science’ to cut costs.


(I'm not sure what the "1986" and "79" in that citation mean. If one of
them's supposed to indicate first use, I _thought_ it was older than
that.) That just shows an example of _usage_ - not the _origin_ of the
phrase (i. e. _why_ we say it).
MB (09.03.2019, 01:35)
On 08/03/2019 20:32, J. P. Gilliver (John) wrote:
> (I'm not sure what the "1986" and "79" in that citation mean. If one of
> them's supposed to indicate first use, I _thought_ it was older than
> that.) That just shows an example of _usage_ - not the _origin_ of the
> phrase (i. e. _why_ we say it).


Standard OED reference.

It was the earliest quoted with that meaning.

Quite possible there are earlier ones, I have certainly found earlier
ones that their examples in the past (and passed on to them). They are
always glad to get references if people send them in.
J. P. Gilliver (John) (09.03.2019, 03:26)
In message <q5uu8m$cd6$1>, MB <MB> writes:
>On 08/03/2019 20:32, J. P. Gilliver (John) wrote:
>Standard OED reference.
>It was the earliest quoted with that meaning.


(But why two dates?)
>Quite possible there are earlier ones, I have certainly found earlier
>ones that their examples in the past (and passed on to them). They are
>always glad to get references if people send them in.

I know - I not infrequently email "cards" to my brother, who's associate
editor and historian there. I don't have a login though!
Does the entry explain _why_ we say it though?
JPG
Andy Burns (09.03.2019, 11:15)
J. P. Gilliver (John) wrote:

> MB wrote:
>> Standard OED reference.

> I don't have a login though!


Most libraries give you access ...
Similar Threads