Calibration on my Flex 6700 GPS seems off at 10 Meters?

Roy Laufer

Hello,

I have noticed a problem with my Flex 6700 with GPS lock using SSDR 1.4.16.142.

I believe it was spot on the frequency on the 10 Meter band previously, but everyone has started to sound like Mickey Mouse and they seem about 70 Hz high. I went to WWV 20.000 MHz, and my 6700 reads the peak as 20.000.056! (By comparison WWV 10 reads as 10.000.028 Hz, so this appears to be off by the same percentage.).

I tried shutting down my 6700 and removed the power, then I gave it some time to lock back into the GPS, but WWV 20 still seems to read as 56 Hz high?

Am I missing something? Do I have to reset it all the way to factory settings?

Any suggestions?

Thanks!

73,
Roy, AC2GS

Roy Laufer

By the way, I tried a full factory reset (now I have to start re-entering all the old settings...), and still "no joy" - still seems to be roughly 62 Hz high on 10 Meters.

Roy AC2GS

Walt - KZ1F

Have you done the math to determine ppm? GPSDO is more accurate than the other two methods, but it still isn't perfectly spot on. Remember too gps's done provide civilian accuracy equal to military accuracy., there is deliberate drift. I was attributing that to selective availability, which was turned off but the GPS does show my house moving so there is error somewhere.

George KF2T

Have you done a calibration from the Radio Settings box? That can apply the proper offset to WWV and square you up.

Roy Laufer

With the GPS module installed (and properly locked on the signal) there is no "calibration procedure". It's all supposed to be done automatically.

Roy, AC2GS

Roy Laufer

Somehow I doubt that the point of the additional GPS unit was meant to allow a frequency offset of 60+ Hz on the 10 Meter band - it didn't have that offset before, and the hardware has no method to correct for it, so I assume that there is something else "going wrong", than that my expectations are set too high?

Perhaps FRS will be able to offer a few words of wisdom in a few days when their Labor Day rivalries wind down.

Roy, AC2GS

Walt - KZ1F

How many satellites does your radio's GPS see?

Roy Laufer

13 visible, 10 tracked, status: Locked.

Jay Nation

Roy

I just verified that the following works.

Go to the GPS tab in radio setup, click uninstall. switch to the Receive tab. click the start button.
It will calibrate. when I did it, the Offset value was set to 0, both before and after the calibration. Then click the GPS tab again and click the Install button. GPS should restart, without losing lock. 

It may not solve your problem, but you can manually run the calibration.

73, Jay - NO5J

Walt - KZ1F

Yep, that would be the definitive source. Does it show your shack moving a few feet over time, even when locked?

Roy Laufer

Well, I did as you suggested.

I set it on 25 MHz and calibrated it, it snapped the panadapter to 25.000.062 MHz with an offset of 0 ppb. I went back to GPS and re-"installed" it.

I still have WWV coming in at 25.000.062 according to my 6700.

So, no change here...

Roy, AC2GS

Roy Laufer

Uh... what would "the definitive source" be? I would assume that 10-11 Sats tracked and the GPS status as locked would mean that I should be in a good position, not 62 Hz out of my expected frequency. The Long. and Lat. seem rock steady, the altitude drifts a little now and again, but it always has.

All this added tech bought me the precision of a used FT897???

Roy, AC2GS

KY6LA_Howard

Since BOTH WWV and GPS give exactly the same frequency ..it very much sounds like a readout error rather than a device calibration error. Perhaps you should open w help ticket with FRS

k3Tim

Agree with Howard.  Before a ticket you might want to try removing the GPS module and reseating it, centering it carefully on the connector.

The GPS / WWV should be within a a thousandths of a cycle from each other, sans propagation of course.

GL,

K3Tim

Roy Laufer

The GPS unit came pre-installed, and it appears to be reporting its data correctly, so I am reticent to poke around inside, until I hear something from FRS on this. I'm hoping that there is some kind of GPS based recalibration option that is unknown to most of us, but I am concerned that it might require a short trip back to where it was "born".

62 Hz off on 25 MHz is 2.5 ppm which is awful for the 6700 with just a TCXO, with a GPSDO its downright absurd.

It's never easy...

Roy AC2GS

Doug Hall

There's no "deliberate drift" in the time info. The time info is good to nanoseconds. An oscillator disciplined to the GPS constellation will be nearly spot on.

k3Tim

When you disabled the GPS and calibrated, was the frequency error under control?  From your earlier post, I could not make that determination.  This data point may helpful to the FRS support folks.

  BTW, here is a dissertation by Prof. Dr. Steve about the frequency error due to setting the DDS:

https://community.flexradio.com/flexradio/topics/synthesizer-frequency-error

Splitting hairs for sure.  


and, yes, it is always something! 

Good Luck...

k3Tim

Walt - KZ1F

What the GPS does is use the information from multiple days to effectively triangulate its location in 3 dimensional space, this takes a minimum of four steady signals. Yes Doug, I tend to agree with you...but... If e erything was perfect with the 11 locked satellites it should compute the receivers location down to the foot or so many inches. Where the GPS status gives its computed four dimensional coordinates they should be invariant, right? But they aren't, your coordinates drift. Independent of the flexradio if you look at the GPS location from your smartphone or tablet it will show you, if you're in your driveway, that maybe you're in the back of your yard of something like that. So working that variability backwards then I'd say perhaps the accuracy is also drifting. Definitive source was Steve's explanation. The .000005 ppm is drift over 24 hr, how you freq appears off I don't know. But if the GPS dialog shows your shack moving, that is likely part of the explanation.

Walt - KZ1F

With selective availability, there is/was. I don't buy civilians and foreign governments have access to the same GPS accuracy, like putting a cruise middle in the left garage bay, that the military enjoy. Selective availability allegedly ended in 2000. Maybe it did and something very similar took its place. Rhetorical question, watch your location in Google maps and you'll see it thinks you're moving. All that is based on what time that satellite says it is and where it says it is, the rest is hs trig.

DH2ID

Splitting some more hairs: GPSDO only uses the satellite's atomic
clock, not the coordinates. You probably know where your Flex is, hi
;-)
BTW I built a 10 MHz frequency normal that is slaved to our central
atomic clock and is on spot with my rubidium frequency normal.I need
that to calibrate my measurement instruments.

Roy Laufer

No, disabling the GPS and manually starting the calibration for 25 MHz caused the panadapter to select a frequency of 25.000.062 Hz, where it was "hearing" WWV. It did not re-adjust that to read 25.000.000 - it kept it with the apparent 2.5 ppm odd error in place. Enabling the GPS option did not cause the oscillator to indicate any other frequency for WWV other than 25.000.062. I tried shutting everything off and disconnecting the power and at first the reported frequency for WWV was 24.999.990 MHz, but that was because the TXCO heater hadn't yet stabilized the frequency. A little while later, it was back at 25.000.062.

Perhaps, when the TXCO goes too far away from its frequency the GPSDO isn't able to correct for it? I dunno.

I hope FRS has some advice other than "ship it back and let us take a look".

I'm sure someone from FRS will be back here by the time that the Labor Day weekend is over.

Roy AC2GS

DH2ID

Some links:

http://tf.nist.gov/general/pdf/2297.pdf

https://en.wikipedia.org/wiki/GPS_disciplined_oscillator

73, Alex DH2ID

Walt - KZ1F

Are you really sure you want to go with the answer? How do you derive location from time at an unknown source? All the sats have the same time. Because they're in different locations the time it takes the signal to get to you differs. Yes, it doesn't use, directly, the location but to compute time it has to subtract how long it took that signal to arrive, right?

Doug Hall

You're dealing with jitter. Whether you are looking for positional accuracy or time accuracy you have to average the GPS info over a long enough period of time to get the accuracy you want. Surveyors do this, and so does my Spectracom NetClock. In the long term it's extremely accurate. Otherwise the guys selling GPS Disciplined Oscillators would be out of business. 

DH2ID

Walt, you are absolutely right, of course,  in that GPS receivers triangulate their position in relation to the satellites received, distance to the satellites and a virtual "center of the earth". GPS time is transmitted through the GPS code.

But this does not interest a GPSDO at all, it discipines it's clock by the offset to the satellite's cesium clock. It is all in the paper listed above.
Theoretically a GPSDO could even be wrong, if the cesium clocks of
the satellites are wrong, but these clocks are monitored from the ground
and regularly corrected.

And - no aging: ... the output is steered to compensate for aging and frequency
drift.... My rubidium normal slowly ages.

Some excerpts I find interesting:

1. From the excellent article in Wikipedia
    https://en.wikipedia.org/wiki/GPS_disciplined_oscillator
... A GPSDO works by disciplining, or steering a high quality quartz or rubidium oscillator by locking the output to a GPS signal via a tracking loop. The disciplining mechanism works in a similar way to a PLL but in most GPSDOs the loop filter is replaced with a microcontroller that uses software to compensate for not only the phase and frequency changes of the local oscillator, but also for the "learned" effects of aging, temperature and other environmental parameters....

2. From the paper by Michael A. Lombardi
    http://tf.nist.gov/general/pdf/2297.pdf
....GPSDOs can and do fail, particularly
when the GPS signal is unavailable in a
local area. There are many possible
failure modes that have been well docu-
mented elsewhere [21], but the most
likely cause of failure is probably RF
interference and jamming (either inten-
tional or unintentional). GPS signals are
very susceptible to interference due to
their low power levels. A receiver can
lose lock on a satellite due to an interfer-
ing signal that is only a few orders of
magnitude more powerful than the
minimum received GPS signal strength,
which is –160 dBW on earth for the L1
carrier, equivalent to 10-16 W. [22] One
“jamming” incident at NIST was caused
by a GPS receiving antenna with a loose
connector. The signals leaking from this
connector jammed other receivers whose
antennas were located 100 meters away....

Jay Nation

I don't think They're fudging the time codes, they do fudge on the positioning code though.

Telco's and other internet backbone providers use the timing data to keep all their equipment synced if that drifted we would all be loosing our connections.

It's not about what time it is, it's more about about everyone dancing to the same beat.

Most telco equipment is synced to onsite caesium or rubidium clocks but they also use GPS as the backup source and to verify that the onsite clocks are maintaining sync.

Any drift in those signals could/would produce bit errors at every office in the data path.

Some offices/locations/equipment can get their timing by reading the timing in signals sent from other offices with the real clocks, but that's being phased out in favor of onsite sync solutions.

IOW things would get very bad, rather quickly.   

73, Jay - NO5J

Walt - KZ1F

I think I may actually have the answer. When I was writing the portable facsimile to flexlib I discovered slice freq was a double, essentially floating point. I did use wwv at 10.000 often for testing both the spectrum display and waterfall. 10.000 isn't a clean 5digit number, it was 10. a bunch of zeros and a 6, as I recall. When the frequency is formatted for display that trailing digit is dropped so you won't see it in the slice flag but if you zoom in enough to see accuracy to the Hz, it will be reflected. I believe you are seeing the rounding error of fixed digit decimal value to its double precision floating point value. I could be wrong but I bet that's it. That just came to me as I was trying to fall asleep.

Roy Laufer

I considered a "rounding error", except that, for WWV it's not off by the last digit, but by the last two digits "62". Also early on when I did this with my Flex 6700 I marveled at its 10 MHz WWV peaking exactly at 10.000.000 MHz - no rounding errors at that time.

I am beginning to think that for some reason in the past week my TXCO drifted out of specs and the GPSDO isn't correcting the drift.

I think I have tried every obvious "trick" that I could think of - I'm hoping that FRS has an undocumented magic "pointy stick" that might re-awaken the self calibration back into service...

We shall see what happens when "the parents" (FRS) come home.

73,
Roy AC2GS

Walt - KZ1F

To be sure, I am curious what the issue is...and the remedy. I didn't want to say anything to divert your issue but I've seen something kinda sorta similar with a straight cw signal. It strikes me, more times than not, a cw signal peaks just to the left of 'on frequency. Occasionally, its spot on, never to the right on on freq. On the rounding issue, that was just an educationed guess, as it strikes me a tad unintuitive a discrete freq, 10MHz, for instance, would be a floating point number. But it is. Even 10.1MHz is 10,100,000Hz. But, that trailing 6 could just as easily be a trailing 60. I can't wait to get Steve's take.

Tim - W4TME

Roy -  I highly recommend that you accurately determine the frequency difference using the Frequency Analysis feature in Fldigi to make sure your radio is actually off frequency.
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CCUQFjA...

Roy Laufer

Wouldn't this suffice?

(this is the panadaptors of three slices set to WWV calibration tones at 25.000.000 MHz, 15.000.000 MHz, and 10.000.000 MHz)

Would you still like me to install and run the other third party software?

73,
Roy, AC2GS