[time-nuts] Loran - any good for timekeeping?
brooke at pacific.net
Fri Apr 21 15:51:18 EDT 2006
Years ago I built a GRI generator controlled by 3 thumb wheel switches
and used that to trigger my scope. Based on a series of articles by
Ralph Burhans in an electronics magazine. The y-axis input was from an
active whip antenna (DA-100). I could not only see the pulses and
identify them, but could also determine my location. But using the
pulse envelops is much poorer than locking at the rising edge of the
third cycle of RF.
The neat thing about the Locus LORAN-C receivers is that they are "all
in view" instead of being tied to a single GRI.
Theoritically you could use stations from different chains to get a
position fix. It's my understanding from talking to the people that run
the Middletown, Calif. station that in the not to distant future all
the U.S. LORAN-C stations will be transmitting their pulses based on UTC
rather than on some delay from the master station as it is now. Each
station has multiple Cesium standards and each has a phase adjuster.
This should have the effect of reinforcing the concept of "all in view"
and now I can see also supports adding a time code.
Here are the two paragraphs in the Austron 2100T manual that deal with
what's between the antenna input connector and the input of their A/D
converter. There's a lot of gain control and filtering prior to the A/D.
I also find that using the Austron 2084 Filter Multicoupler between the
antenna and 2100T helps since I can tune a number of the filters on the
same interfering signals. It also has a sferics blanker.
If what you're doing is less than this then maybe you don't have as
robust a front end.
First RF Amp:
The 1st RF Amplifier amplifies and filters the antenna input. There are
also four attenuator stages which permit microprocessor control of the
receiver gain. The 50 ohm (RG-58) antenna cable is coupled to the first
amplifier stage by the impedance matching transformer T1. T1 also
provides approximately 20 dB of gain. The signal is then amplified by 16
dB in (U1) and (U2) and by 8 dB in (U3). From (U3) the signal passes
through a 5-pole Bessel filter, which has a center frequency of 100 kHz
and a bandwidth of 40 kHz. Additional amplification is provided in (U4),
(U5) and (U6) before the signal leaves the 1st RF Amplifier. IC (U7)
consists of four single-pole, single-throw analog switches that control
the gain of the 1st RF amplifier. Since the operation of each switch is
the same, only (U7B) is discussed. The signal at the output of (Ul) goes
through R6 to the noninverting input of (U2). R6 and R10 form a voltage
divider when the control voltage to (U7Bt8) goes to ground, closing
switch (U7B). This causes a 32 dB attenuation of the signal.
Second RF amp:
ICs (Ul) through (U4) form a four-stage amplifier with a total gain of
38 dB. IC (U7) consists of four single-pole, single-throw analog
switches used in four microprocessor- controlled attenuators. The output
of amplifier (U2) goes to buffer (U5) which drives the 3-pole
Butterworth filter. The output of the filter is converted to a TTL level
signal by (U6). The output of (U6) is the hard-limited rf used for
acquisition. The output of (U4) is the amplified and filtered loran C
w/o Java http://www.pacificsites.com/~brooke/PRC68COM.shtml
Poul-Henning Kamp wrote:
>In message <44491E9F.1060600 at pacific.net>, Brooke Clarke writes:
>>What is a "Frame rate cyclic averaging buffer"? Is this where you use
>>a GRI generator to trigger sampling the incoming signal?
>It is a buffer which can hold one FRI worth of samples into which I
>average the received signal.
>In other words, for a 1Msps and 9660 GRI, it will be:
> 9660 * 10 * 2 = 193200 samples long
>>I've found that the quality of LORAN-C for timing depends on how close
>>you are to the transmitter. When in the 100 mile range the quality is
>>equal to better than GPS, but when it's many hundreds of miles there's a
>>lot of variation.
>Yes, the skywave at night is the killer problem.
>>I'm attaching a gif of the spectrum here from 0 to 200 kHz that goes
>>with the web page:
>Looks very typical.
>Try this: set a pulse generator to the gri-rate of a nearby LORAN-C
>chain. Connect it to the external sync trigger of your spectrum analysator
>Then set it for start=100khz, stop=100khz, bandwidth=10khz and video
>averaging (or whatever it's called)
>You should be able to see the loran-C pulses quite clearly.
More information about the time-nuts