ฝัง
- เผยแพร่เมื่อ 14 ธ.ค. 2021
- Episode 984
A gift to the channel by a viewer (thank you Jay). This is a very accurate frequency standard that uses GPS data. Also referred to as a GPS disciplined oscillator. I compare the 10 MHz output to my Rubidium standard. More info at:
www.leobodnar.com/shop/index.p...
Be a Patron: / imsaiguy - วิทยาศาสตร์และเทคโนโลยี
Maybe you try again how long it takes to get accurate. I assume it is only minutes. At least mine settles pretty fast. We need this accuracy to transmit over the QO-100 satellite on 10GHz.
The antenna is active, so beware of the bias dc in the line if you connect something else to it.
The output is a mos square wave and not all equipment likes that, as you noticed, i think thats causing the change in accuracy at lower levels, it should not matter.
I added a 10 mhz lowpass filter to get a sinewave out as my marconi tester also did not like the square wave
You can look at the output of the Bodnar and your Rb box on your scope in the X-Y mode to see how they differ. I was surprised when I tried that the first time between my Rb and my GPSDO (Z3801A) standards; my intuition thought they'd be the same (a nice unmoving circle), they're not...but like you say, close enough for government work!
Thanks for the video. I enjoyed this and your NanoVNA videos. I've learned a lot about them.
Here's what I was told about rubidium standards, and I have no reason to doubt it since I was told by a physicist who works on freq standards. Rubidium is being burned off with time as the device ages. The process is slowly self-destructive. There is a voltage that is monitored to let the commercial entities (the cell companies,) know how far the rubidium has burned off, since it has to be driven ever harder as it starts running low on rubidium. That voltage that is monitored lets the field techs know when it's time to swap out the old standards. So when you buy one of those pulls you are running on borrowed time, however long it is. I had a friend who ran his 24/7 for 2 and a half years. Then it died. Mine, I only turn on the day before I need it, then I compare it against WWV, then shut it right back off as soon as I'm done with it. I've had it for 15 years. Rubidium standards are very stable. They just don't drift a lot. But they suffer from phase noise, so aren't great to use as calibrators. Nor are they essentially accurate unless you have a better standard with which to check them occasionally. I was told that a better option is a good ovenized xtal oscillator because they have low phase noise, but one that is carefully disciplined by a GPS receiver since they have the long-term accuracy. So that's what I'm in the process of building using an old, very well burned in HP xtal oven and a GPS receiver. Anyway, there's all I was told. Hopefully it's good info and helpful.
These Leo Bodnar clocks are very popular with QO-100 trannsmitters and getting the IC-9700 onto frequency on 23cm.
I found that you need a a fairly good view of the sky with the antenna to get good results from the GPS disciplined oscillator. Since the loop time constants are so long a little GPS drop out will not be noticed but it will temporally cause a slight phase shift (frequency shift) in it's 10Mhz output. I discovered this while testing a couple of GPSDOs against a Rb standard. Comparing the two on the scope the signals would phase track for 10 minutes or so then drift apart for a while then come back in phase again. Couldn't tell if it was the GPSDO or the Rb standard that was drifting but suspected the GPSDO. Finally got a second Rb standard and discover by watching both Rb's and the GPSDO that it was the GPS that was drifting. I remounted the antenna on the roof and boy did that make a difference. It was amazing to see all three independent 10Mhz sources running wild after a power on then all of a sudden bang! the Rb's and the GPS all lock up and thee 10MHz traces all track rock stable on the scope. To make the error show up better I use a 10X frequency multiplier and feed 100MHz into the scope. The idea is to use a very long loop constant on the GPS then to filter the short term GPS anomalies then use that to calibrate and adjust the Rb modules. Yes , there is a frequency adjustment on the Rb oscillators and it takes a while to get it dialed in. You have to be careful not to move the Rb oscillators because the earths magnetic field also disturbs the frequency in these micro Hertz regions.
Thanks for opening the box, was wondering what the magic sauce was inside.
I've always been fascinated that a PPS (pulse per second) from GPS can be accurately multiplied into a 10E6 per second reference signal. It just seems like the smallest bit of jitter or other threshold-level sensing on a PPS signal could cause significant and large disturbances in the 10 MHz (or other MHz level) signal coming out. Driving the Vctrl pin on an OXCO would of course smooth out and average some of the errors. Also, looks like those SMA connector grounds were soldered with true minimalism. (Full disclosure, I own one of these as well.) It might be interesting to see the error/jitter at the highest supported frequency. Maybe that's in your PT2 video, I'll go check 🙂Great video as always!
--E
de W6EL
I am fortunate that my hobbyist lab has enough nice equipment to need various reference standards. I looked at DIY GPSDO solutions and ended up building a GPSD/Clock with outputs. Basically, it is a GPS receiver, a PIC MCU, a display, a reference signal output connection and a Laptop/GUI output frequency control. My thinking was this... The GPS provides the timing standard. The additional oscillator (of various types - heated, Rb, etc.) is only used to carry over the actual reference frequency, whenever it is temporarily lost. As a hobbyist, if the GPS signal is not available for awhile, I can just simply wait to do the calibration at a later time. Meaning, I do not need all of the additional smoothing features. I use a powered antenna, good line of sight and the unit stays on 24/7. How far off of base is my thinking. Please be gentle; this is a teachable moment. I have not used it to calibrate anything, yet.
I have worked with the si5328 ICs, not top shelf but a pretty good tracking PLL with 2 input reference clocks and 2 outputs. It will free-run on the TCXO until the input error converges to a minimum and then locks to the input. Requires an ICAL to force the PLL to try and re-lock.
That little GPSDO seems to be accurate enough for my kit, once it warms up.. probably best to just keep it running 24/7.
Cheers,
Awesome video, as usual. Thanks!
I don't recall the stability of my gps ref. rubidium freq. std.. but this might be quite useful for the 'non-std' (10 MHz) ext. clock referenced devices... IF it's better than the internal reference, my swag is that it probably is.
You did not mention the issue of short-term vs long-term accuracy. Since the GPS is referenced to an atomic clock and maintained by scientists, one would expect its long-term accuracy to be exceptionally good. Its short-term variation would be expected to exceed that of the rubidium standard due to varying effects of the atmosphere on the radio signals from the GPS satellites.
Agreed. Much like using the power line as a frequency standard, the important specification is long-term stability as seen over MANY, many measurements. Shorter term frequency fluctuations are controlled differently, all the way down to nitty-gritty stuff like jitter/phase noise.
The term you're looking for is OCXO
This was interesting to see. But it is waaaaay above my level of needed accuracy. I'm such a hairy handed, knuckle-dragging, mouth breather, that simply beating the 10 mhz oscillator from my cheap little handheld frequency counter against WWV on the shortwave will get you within 10 hz or so. Way more than I need for any of my ham radio projects.
Why don’t you show the output at 10,000000 compared to you stander??
part 2
Their plot shows the power vs kelvin-hertz (K) at the higher values. Wrong unit.
Yeah, please don't bother reading the manual or learning anything about the device before you make a video and ramble on about it's performance.
somebody pi$$ in your cornflakes this morning? He compared it with his Rubidium..a few tenths of a Hz either way...what are you looking to do ?? Build a nuke for iran or something?