Hello
Wolf,
Thank
you for your reply; the Thunderbolt E 1 PPS and 10MHz specs
follow:
1
PPS Interface Specification:
--
BNC Connector 0 V to 2.4 V ±10% into 50 Ω
--
10 microsecond wide pulse with the leading edge synchronized
to UTC within 15 nanoseconds (one sigma)
--
Rise time <20 nanoseconds
10
MHz Interface Specification:
--
BNC connector; waveform is sinusoidal 7 (+/-2) dBm into 50 Ω
--
10 MHz accuracy: 1.16E-12 (one day average)
--
10 MHz stability: Allan Deviation ~ 1E-11, 1-512 seconds;
5E-13, 8192 seconds
Holdover
(GPS disabled): 8us in 24 hours
I
once asked Trimble tech support about the rationale for the
fixed 10us pulse width (not adjustable via software or
hardware); they did not have any additional information
available, other than to confirm that the pulse was not
adjustable in amplitude or duration.
I
read Andy’s message, in which Andy said:
“Won’t
the sound card's input filter, the anti-alias filter, extend
this pulse, however short it is, into something the sampling
rate will always see? So there is no need to deliberately
lengthen it. The only real advantage to doing so, perhaps,
would be to increase the pulse energy within a one sample
period leading to a height amplitude in the capture. The 1
PPS from my HP-5061 is only 1us in width, but can still be
detected on a soundcard stream at any sampling rate chosen.”
Andy’s
comment (above) reminds me that my reason for sending a
question on this topic in March 2014 was that I had
initially connected the Thunderbolt E directly to the sound
card, verified that the amplitude was not excessive, but
failed to pass the sync tests mentioned in SpecLab
documentation and posts. Given that I had no experience with
external sync in SpecLab at that time, I may have made any
number of mistakes in that exercise.
73,
Jim
AA5BW
Hello Jim,
I still wonder why the sync pulses from the Thunderbolt E are
so short - you mentioned 10 us in your email, but what is the
exact pulse length ? With a slightly modified software (SL) it
may be possible to use them as they are.. if they are longer
than a sample from the soundcard.
73,
Wolf DL4YHF
(not QRV on LF+MF unfortunately..)
Hello
Stefan,
I
have a question that is a digression from Tony’s question
but perhaps related:
Bill
(WB6BNQ), Graham and Wolf (on March 3, 2014) suggested
(thanks!) the following to make the narrow 1pps pulse from
my Trimble (Thunderbolt E) GPSDO] compatible with an
external USB sound card:
A
pulse stretcher based on either a modern version of 74123,
such as a 74HC123, or a stretcher made from half of a
74HC74 Dual D flip-flop [resistor from not-Q to not-R;
parallel speedup diode (anode to not-Q), and capacitor
from not-R to ground], to stretch the 1pps pulse to well
over 10us (at least a few milliseconds)
I
am preparing to interface two more Trimble Thunderbolt E
GPSDOs to an external USB sound card for SpecLab , for a
VLF loop-TX test this year (60 meter horizontal square
loop with 6 kW PA power; low ERP but GPS disciplined so
maybe some interesting signals); and I wonder if the above
solutions (one-shot or flip-flop) for interfacing the
narrow Trimble 1pps signal to an external USB sound card
for SpecLab remain the recommended approach. Do you have
any thoughts on this?
Thank
you, and apologies if this obfuscates Tony’s thread.
73,
Jim
AA5BW
Hi Tony,
OK so if your RX and PC time is stable, we can give it a
try.
-You need a stereo input soundcard and feed the PPS signal
to the right channel. Just run it at 48 kS/s. The PPS signal
must be DC decoupled and the level must be limited to 80% of
the ADC range, simply use 100nF and a resistive divider. The
values will depend on your GPS module and its output level.
-Download the latest version of SpecLab and install it to
the recommended folder.
-Turn the RX to 136.00000 kHz USB and use a 2.5 kHz wide SSB
filter. Connect the RX to the left channel of the soundcard
input.
-Download the usr file in this mail and open it into SpecLab
(File -> Load settings from...)
-Click on Components -> Show Components. A box opens. In
the top left corner, click on "SR cal". Another box opens.
The Status must be green, showing pps peaks OK. If the level
is to high, click on "scope" and see the peak form. Then you
need to adjust the divider of your GPS module to reduce or
increase the PPS level. The box must be green for most of
the time in the end. Click to scope again, there is a pulse
counter PC and an error counter EC. The error should be
lower than 2 % or so, for example PC=1000 EC=16. Just let it
run a few hours and see if the tracking is stable.
If you have problems in this step, tell us what the problem
is.
-Wait a few hours and check if text files appear at
C:\Spectrum\data After the first 40 minutes, a file should
appear each 10 minutes.
-Ask someone to transmit a short EbNaut message at 1s
symbols and make sure to record on this time.
So far. Then we can continue with part 3 :-)
73, Stefan
Am 23.10.2016 19:09, schrieb Tony Baldwin:
Hi Stefan
My receiver is an Elad FDM-DUO SDR with a 10 MHz GPS input.
Unfortunately SpecLab doesn't have that radio for an SDR
input, so I have to use the sound input via USB.
I could send the 1 PPS to my computer via RS 232 as well if
that helps.
73, Tony, EI8JK.
On 23/10/2016 17:51, DK7FC wrote:
Hi all,
I started to write the second part of this tutorial but
when finished i was not happy with the result.
The first thing one can say is that you will need a very
frequency stable receiver. There is no chance without an
external reference signal. For generating wav files using
SpecLab you will also have to use a ref signal to
compensate the drift of your soundcard. Some people have
already prepared things in the past. You can use a 1 PPS
signal for soundcard drift compensation or a down divided
10 MHz GPS reference or a stereo soundcard receiving VLF
MSK signals on the other channel to compensate the drift.
Also the SpecLab configuration depends on your receiver
type.
Now i could simply make some assumptions what your
equipment may be but this may be not a good way.
So now i think it's better that we start with one example,
one (or more) of the stations who are interested in EbNaut
decodes on LF should describe the available system
components. Then we
can work it out what must be done.
We will need to have a stable RX frequency and a drift
compensated soundcard frequency in the end.
So who of those who are interested wants to start?
73, Stefan
