[time-nuts] GPSDO project
wb6bnq at cox.net
Sun Jun 28 22:55:46 UTC 2009
I am a little confused about your intentions. There is no mention on the RFspace
web site about a provision of being able to feed in an external reference signal
to the SDR-IQ or the SDR-14 for that matter.
If you study the Analog devices AD6620 component, you will see that it is a 3.3
volt * * * ONLY * * * process, which only adds to the problem. Additionally
there are constraints on the rise and fall times of the timing signal. This is a
horrifically complex component with intricate dependencies tied to the specific
A/D used. This is not your daddy's VFO from yester-year.
The fact that the internal DDS is a digital function means you will never, ever,
get to a perfect cardinal point, except under very special conditions. In this
case it will have an offset of 16 pico-seconds because that is the smallest
resolution the DDS can do. While small it would require very special hardware
for you to keep it locked to the GPS. As the internal clocking system runs more
then just the DDS, other considerations come into play. The "clocks" built to
run these chips are not designed for external control and studying the physical
layout of the SDR-IQ, it does not appear that they provided any means to have
external clocking ability.
Besides the computer's sound card DSP is heavily involved with regards to the
"frequency" calibration. So there are two independent clocks to keep
synchronized that are not related to the "listening time," particularly if you
record the digital bits for later listening or analysis.
Fortunately, the software will allow for you to calibrate, to some degree, but it
is not and you will not be able to approach what you expect to do with the
Thunderbolt. Keep in mind that the best you could do with a Thunderbolt on a 1
second per second comparison basis is only 1 part in 10 to the minus ninth
(1x10^-9). It gets worse if you consider shorter times like a continuous analog
aspect that you would have for active listening. The noise of the GPS system
would mask the 16 pico-second resolution of the SDR-IQ DDS.
The real problem is temperature. You could improve that by providing a chamber
for the SDR-IQ and raising, carefully, the temperature to just above your highest
After all that, I am not saying you couldn't/wouldn't/shouldn't consider it.
Keep in mind that the SDR-IQ, or for that matter, the SDR-14 DDS only sets the
center of the 190 KHz bandwidth swath that you are viewing at any one time. The
real fine aspect is done in the computer and the clocks in those are all over the
map, so to speak.
In order to have a really tight calibration, with "Timenuts" quality, would
involve some serious effort to build a clocking system for both the SDR-IQ and
the computer. The final result is you would still be offset due to the digital
processes in both the SDR-IQ and the computer, plus, perhaps, some latency in the
As I get to the point of sending this email I see that Dave Powis, G4HUP, has
gone to the trouble to lock the SDR-IQ internal oscillator to an external 10
MHz. If you look at all the information on his web site, you will see that it is
not a trivial project. As I pointed out, this only stabilizes the "window" of
the SDR-IQ still leaving the computer to tend with.
My experience with the Thunderbolt shows that it has some temperature
dependencies and the expected nominal noise of the GPS system in the short term
regard. What most people do is have a high quality "house" standard and do long
term comparisons to the GPS. A high quality "house" standard would, in effect,
filter the short term noise of the GPS system.
At any rate, this is my fantasy and I am sticking to it !
Brad Dye wrote:
> I have just purchased a Thunderbolt GPS receiver. I hope to use this
> to discipline an oscillator on 66.66666 MHz so I can use it as a LO on
> my Software Defined Receiver (SDR-IQ).
> Building such an oscillator is a little over my head so I thought I
> would ask the group if anyone knows where I could buy the missing link
> of this project. I have put a drawing of my project on the web showing
> the specs of what I need. The oscillator is shown in the blue box.
> Any help or suggestions would be greatly appreciated.
> Brad Dye K9IQY
> ex KN9IQY, KN4BK, KM5NK, WB4JCF, ZP5TQ, WA4VXU, WA9RVL, /TI2, /9Y4, /
> 6Y5, /KP4
> 52 years as a FCC licensed amateur radio operator
> 36 years as a FCC licensed first class commercial radio operator
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts