EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: electricspark on June 30, 2015, 02:02:16 pm
-
Hi everybody. This is my first post and maybe a daft question;
I would like to make a GPS disciplined oscillator. I like the look of the one at http://gpsdo.i2phd.com/ (http://gpsdo.i2phd.com/) as it looks to be a nice simple circuit. My only difference will be the reference frequency.
I would like to use 1Hz from a GPS module I have and feed it directly to the 74HC4046 comparator as in the circuit. I would like to divide the 10Mhz from my VCTCXO down to 1Hz and also feed it into the comparator.
My question is, would this work with the two signals being compared being such low frequencies? Would there be any drawbacks?
Any advice would be appreciated as you can probably tell, my PLL knowledge is not that great!
-
Comparing phase of two 1 Hz signals in a PLL will produce a very slow servo.
-
OK. My ignorance is really coming to the fore now. What are the implications for a slow servo – Will the 10Mhz output drift up and down? Will it take a long time to lock (I don’t mind this, I can just leave it permanently on).
Will my proposed mods to the http://gpsdo.i2phd.com/ (http://gpsdo.i2phd.com/) circuit make it unworkable?
-
Trying to lock two 1Hz signals with just an analogue PLL is going to take pretty much forever (as TimFox says, a _very_ slow servo).
With some intelligence on the board to help you, things get a lot easier. Try taking a look at
http://www.ik0otg.net/index.php?option=com_content&view=article&id=53&Itemid=59&lang=en (http://www.ik0otg.net/index.php?option=com_content&view=article&id=53&Itemid=59&lang=en)
Which should at least give you some ideas.
-
Brooks Shera came up with a digital filter based design a few decades ago that has been built by a lot of people.
see the link in James Miller's description of his building of it:
http://www.jrmiller.demon.co.uk/projects/freqstd/frqstd.htm (http://www.jrmiller.demon.co.uk/projects/freqstd/frqstd.htm)
It is much easier if you have a GPS with 10kHz output (see James Miller's site again).
If you go on TimeNuts lists they have extensive discussion on the Brooks Shera design including its drawbacks which are rather esoteric and only important if you become a time nut!
I've been playing with GPSDOs for a while but have yet to actually build a complete system - I got a board from James Miller and made it up and also have a couple of GPS modules with 10kHz output and one with 1pps and a stack of OCXOs off ebay but I'm keen to design my own (when I get time) so have just been experimenting sofar.
-
If you can compare two 1 Hz signals digital would be an easier to implement. One thing to keep in mind is that the Oscilator you are training needs to be stable beyond your update rate to actually lock. Ie, The oscillator need to hold time over a minimum several second so that you are able to close the control loop.
-
Thanks for the links and thoughts.
I have zero experience with PIC’s and would like to keep things really simple without spending any more money. As far as I can tell, my NEO-6 GPS module outputs a 1pps signal. How about me multiplying this to 10Hz or 100Hz (using another 4046) and then feeding the main PLL?
-
The Neo-6G/Q/M/P/V can be configured to output 0.25 Hz to 1 kHz, and the Neo-6T can be configured to output 0.25 Hz to 10 MHz. See Section 11 of the Neo-6 Protocol Manual.
Cheers,
/dev
-
Of course if you want to do the work and fun with the design, this is probably a fun project. One question though: Is it to get to the end result or the project and building fun you want? If it s the former, then maybe this worth buying.
http://www.ebay.com/itm/281701995263 (http://www.ebay.com/itm/281701995263)
-
Something like this ? -> http://archive.siliconchip.com.au/cms/A_108508/article.html (http://archive.siliconchip.com.au/cms/A_108508/article.html)
-
Thanks for the links and comments. I want this to be an interesting project plus a means to finally make my frequency counter & signal generator agree with each other!
It would be great to use what I have as it is. But, I have ordered a CP2102 module so I can connect the Neo-6M to my PC USB port and have also downloaded u-centre 8.17 so I can alter the frequency to 1Khz. Hopefully my original plan will then work!
Having said this, the item Lightages suggested is not too expensive at all & even comes with an antenna. I think this will be an option if I can’t get this project to work properly.
-
Thanks for the links and comments. I want this to be an interesting project plus a means to finally make my frequency counter & signal generator agree with each other!
That is easy... just use the one with the better timebase to drive the other one.
You can use the GPS 1PPS output to calibrate it once in a while.
-
Have a look here: http://www.dpieshop.com/trimble-icm-smt-360-gpsgnss-timing-receiver-with-10mhz-clock-p-1634.html (http://www.dpieshop.com/trimble-icm-smt-360-gpsgnss-timing-receiver-with-10mhz-clock-p-1634.html) and http://www.trimble.com/timing/ICM-SMT-360.aspx (http://www.trimble.com/timing/ICM-SMT-360.aspx)
Note that those el cheapo GPS receiver modules lose satellites (and as a result the 1pps) on the long run, after a week or so. Looks like the Kalman filter saturates or something like that. Anyway resetting the receiver and/or doing a cold start solves that but that is not what you want. Those are fine for navigation but not for timing.
-
Of course if you want to do the work and fun with the design, this is probably a fun project.
In my experience controllers with rise times longer than a second are never fun.
-
Note that those el cheapo GPS receiver modules lose satellites (and as a result the 1pps) on the long run, after a week or so.
That's hard to believe. Details please?
-
Details? Just what I wrote. Have had that with the low cost Chinese G-Top and Fastrax modules. Didn't try others, but there is a load of Chinese $10 other ones around. Under low signal conditions, but not that low it was unable to do a proper cold start. After it looses all satellites and doing a new cold start, they found 8 - 10 satellites with a snr above 15 dB and could do a valid 3D estimation again.
-
Sounds like buggy firmware. Didn't see anything like that with my ublox modules, ever, but then again, the ublox stuff is pretty good... and has proper docs.
-
I have fooled around with my Neo-6M and it will accept upto 100Khz time pulse output - This will certainly make things a lot easier!