Some GPS receivers that I've used in the past (for stratum-1 NTP clocks, back when you had to build them yourself, anyway..) can be configured to optimize TDOP (Time Dilution of Precision) rather than the usual HDOP (horizontal dilution of precision.) Usually these receivers will get into a survey mode, where over a longish period of time (hours or days), they'll compute your mean position on the earth. Then they enter "position hold mode" or maybe called something similar in any particular receiver. At this point, the receiver can optimize computations to recover the correct time -- it knows the path length to each GPS spacecraft because of your position and the satellite orbital elements and what the time should be. As it has fewer unknowns in computing the pseudo-range to each spacecraft, it can do a better job. The result of all that is a 1-PPS output synched to the start of each second. That's used to discipline a PLL by steering the TXCO or OCXO and poof! Clever implementations might try to relate measured temperature of the oscillator vs. corrections applied and use that to help ride-out loss of GPS synchronization steering. I think the clock in your product is a TXCO; didn't look large enough for an oven, which usually has some insulating structure around the crystal. Hard to tell from the view from the back in the previous video.
Two points: The GPS antenna needs to be mounted outdoors to get a lock on enough satellites. Yes, rubidium's are very stable, but there's no reason to expect them to be 'absolutely' on frequency unless they have been dialed in sometime in the last few years.
"The GPS antenna needs to be mounted outdoors to get a lock on enough satellites." Depends on the building construction. My house is sufficiently RF transparent that, from inside, I can receive more-or-less all of the GNSS satellites that are in view. Same for L-band satellite radio; no problem. I intentionally avoid metal foils and metal roofing in my house construction for this very reason.
It takes time to connect to each satellite, and it(Antenna) needs as many as Sats that are possible to "see" and compare to another Sat. Take the antenna outside and compare the adaptation time to indoors with fewer Sat Clock references to begin and calibrate from.
Hi, Thanks for the demo. I didn't see your antenna stuck to a metal sheet. It's recommended to use 9"X9" minimum. I have my antenna stuck to the top of a metal cabinet and no metal in the way. I installed a CW-TIM GPS module in the frequency counter. It has 2 LEDs one red one that flashes for every satellite it sees on power up. When it locks onto a satellite signal a green LED flashes for each satellite it's locked onto. The specs say that 10MHz output is stable when the green LED starts flashing, even just once. I use it as my standard. I want to put it in it's own case and add a distribution system.
Fairly similar to my results. I keep both my GPSDO and Efratom FRS powered on full time. I check the two against each other periodically and note that the drift rate of the GPSDO will occasionally speed up then or slow back down against the Rubdium. I have other oscillators in the garage that I can check that still agree with the Rubidium, so I believe that it is the GPSDO that has the short term variance.
In part 1, you also made mention that you saw the frequency drift when you adjusted the output level on the GPSDO, you could have demonstrated that, and also done a real time check to see how many satellites you were locked onto. Since your GPS antenna was just laying on your bench ( unless you moved it without saying) you may have not even had a good lock to compare to your other source.
You need to put the antenna at least stuck to a window(with some double sided tape) but preferably outside so it can receive as much satellites as possible, it does averaging. If it receives more satellites it also becomes faster more accurate.
There is an old version of the software that shows the number of satellites it receives, with more satellites it becomes more accurate. I have the dual channel model and always wondered if it would be more or less accurate against a rubidium, mind you: the gps is referenced to two cesium clocks so basically the gps signal would be more accurate but then you have the radio transmission which the rubidium does not have. So to get the answer you would need direct acces to a cesium clock ;-)
A rubidium clock is a secondary frequency standard. It is only as accurate as what it was adjusted to when compared to a higher order standard. How did you set up your rubidium standard? If there is a noticable difference in the frequency of the GPSDO & the rubidium after a few days, I would suspect that it was the rubidium that was out, not the GPSDO. Also, it would have been intersting to have seen the configuration screen for the GPSDO, to see the number & quality of the satellites received.
You'd think it could store the OXCO voltage when it is turned off to minimize the settling time. A good GPSDO should achieve 1e-12 or better. A Ublox LEA-5T-0-003 50-channel 15ns 1pps Timing GPS receiver will do much better than the Leo. You can use pretty much any quality OCXO to lock to the 1 pps. There are plenty of schematics on the web. If you are going to spend what the Leo costs, you might as well get your money's worth and do it right.
![[LIVE] : ONE ลุมพินี 82 วันนี้!! คู่เอก | ยอดไอคิว vs อับดัลลาห์](http://i.ytimg.com/vi/hSsyTl_9bqs/mqdefault.jpg)
Some GPS receivers that I've used in the past (for stratum-1 NTP clocks, back when you had to build them yourself, anyway..) can be configured to optimize TDOP (Time Dilution of Precision) rather than the usual HDOP (horizontal dilution of precision.) Usually these receivers will get into a survey mode, where over a longish period of time (hours or days), they'll compute your mean position on the earth. Then they enter "position hold mode" or maybe called something similar in any particular receiver. At this point, the receiver can optimize computations to recover the correct time -- it knows the path length to each GPS spacecraft because of your position and the satellite orbital elements and what the time should be. As it has fewer unknowns in computing the pseudo-range to each spacecraft, it can do a better job.
The result of all that is a 1-PPS output synched to the start of each second. That's used to discipline a PLL by steering the TXCO or OCXO and poof! Clever implementations might try to relate measured temperature of the oscillator vs. corrections applied and use that to help ride-out loss of GPS synchronization steering. I think the clock in your product is a TXCO; didn't look large enough for an oven, which usually has some insulating structure around the crystal. Hard to tell from the view from the back in the previous video.
Two points:
The GPS antenna needs to be mounted outdoors to get a lock on enough satellites.
Yes, rubidium's are very stable, but there's no reason to expect them to be 'absolutely' on frequency unless they have been dialed in sometime in the last few years.
"The GPS antenna needs to be mounted outdoors to get a lock on enough satellites."
Depends on the building construction. My house is sufficiently RF transparent that, from inside, I can receive more-or-less all of the GNSS satellites that are in view. Same for L-band satellite radio; no problem. I intentionally avoid metal foils and metal roofing in my house construction for this very reason.
It takes time to connect to each satellite, and it(Antenna) needs as many as Sats that are possible to "see" and compare to another Sat.
Take the antenna outside and compare the adaptation time to indoors with fewer Sat Clock references to begin and calibrate from.
Put the scope in X-Y mode and examine the Lissajous Figure
Exactly! Would this not be the easiest way to measure the relative phase drift of the two signals?
Would it look that good with a square wave on one input? I have only ever generated lissajous from two sines.
I prefer direct waveform observation. It's very easy to see the difference.
The GPSDO should be more stable than the Rubidium. The GPS system time base is a bigger badder cesium time base.
Hi, Thanks for the demo. I didn't see your antenna stuck to a metal sheet. It's recommended to use 9"X9" minimum. I have my antenna stuck to the top of a metal cabinet and no metal in the way. I installed a CW-TIM GPS module in the frequency counter. It has 2 LEDs one red one that flashes for every satellite it sees on power up. When it locks onto a satellite signal a green LED flashes for each satellite it's locked onto. The specs say that 10MHz output is stable when the green LED starts flashing, even just once. I use it as my standard. I want to put it in it's own case and add a distribution system.
you might like this: th-cam.com/video/wem0DoPe5O0/w-d-xo.html
Fairly similar to my results. I keep both my GPSDO and Efratom FRS powered on full time. I check the two against each other periodically and note that the drift rate of the GPSDO will occasionally speed up then or slow back down against the Rubdium. I have other oscillators in the garage that I can check that still agree with the Rubidium, so I believe that it is the GPSDO that has the short term variance.
In part 1, you also made mention that you saw the frequency drift when you adjusted the output level on the GPSDO, you could have demonstrated that, and also done a real time check to see how many satellites you were locked onto. Since your GPS antenna was just laying on your bench ( unless you moved it without saying) you may have not even had a good lock to compare to your other source.
I use it and compare with my rubidum in this newer video: th-cam.com/video/kdj6kfDrVCc/w-d-xo.html
It is very stable and accurate.
You need to put the antenna at least stuck to a window(with some double sided tape) but preferably outside so it can receive as much satellites as possible, it does averaging.
If it receives more satellites it also becomes faster more accurate.
There is an old version of the software that shows the number of satellites it receives, with more satellites it becomes more accurate.
I have the dual channel model and always wondered if it would be more or less accurate against a rubidium, mind you: the gps is referenced to two cesium clocks so basically the gps signal would be more accurate but then you have the radio transmission which the rubidium does not have.
So to get the answer you would need direct acces to a cesium clock ;-)
A rubidium clock is a secondary frequency standard. It is only as accurate as what it was adjusted to when compared to a higher order standard. How did you set up your rubidium standard? If there is a noticable difference in the frequency of the GPSDO & the rubidium after a few days, I would suspect that it was the rubidium that was out, not the GPSDO.
Also, it would have been intersting to have seen the configuration screen for the GPSDO, to see the number & quality of the satellites received.
i got one of them 10MHz GPSDO SAMSUNG GPS with a OCXO and it works good I got it for $90 and it had 4 outputs
i got my gps antenna stuck to the tin roof
You'd think it could store the OXCO voltage when it is turned off to minimize the settling time. A good GPSDO should achieve 1e-12 or better. A Ublox LEA-5T-0-003 50-channel 15ns 1pps Timing GPS receiver will do much better than the Leo. You can use pretty much any quality OCXO to lock to the 1 pps. There are plenty of schematics on the web. If you are going to spend what the Leo costs, you might as well get your money's worth and do it right.
Fundamental Q: what reference are we referencing to - the 'Standard' of 'the Standard'. or what the scope thinks?
With the antenna inside the house I don’t think it’s a good idea.
GPSDO require a good view of the sky, otherwise the frequency is dependent on the OCXO.
did you have the antenna inside or outside?
this video is actually a "hey, look at me, I have a Speedmaster" hahahaa :D Bonus points if it was a "moon" model. (my fav watch btw)
what do you have for a Rubidium Standard?
It is based on the Fe-5680A Rubidium. added a large heatsink. also have a distribution amp to supply the 10MHz to various instruments.
@@IMSAIGuy Thanks!
what happens if it loses GPS? Or if satellites go out of view, new ones come in?
Yeah it locks on to new ones as they pass by.
i know that. what i was asking about was whether or not those transitions introduce variations in the output.
5 milihertz difference?
2min 3sec = 123sec, or 8 mHz. Therefore, short term stability is 1/(10E+6 x 123) ~ 1ppb. Not much!
What is the problem with usb-c? Still using mini usb plugs, why?