Re: ubitx circuit, wiring, source code on github

Jerry Gaffke
 

Gordon, I think you're right.

Part of what triggered me to think the github archive was out of date was mention of using
an AM broadcast as per post 35235 in the calibration instructions at   http://www.hfsignals.com/index.php/ubitx-tuneup/
And I still wasn't quite fully understanding post 35241.

As I dig deeper, am thinking the github code will do as per the calibration instructions.
All three Si5351 clocks do get updated with each step of the calibration process.
They are are all scaled by the same vcoa calibration factor.

I think the code could be made simpler and easier to understand,
Just repeatedly twiddle the value of si5351bx_vcoa and call si5351bx_setfreq()  for each of the three clocks.
Could set it to an accuracy of well under 1hz across the dial.  Till the temperature changes by a fraction of a degree.

One possible issue with the current calibration scheme is that it apparently moves vcoa by leaps of 8750hz, or 10ppm.
So the 57mhz clk1 moves in steps of 570hz with each step when calibrating?
Though that is mostly canceled out by the VFO moving in the same direction by roughly the same amount
if you are calibrating from a received signal in the general neighborhood of 12mhz when in USB mode.
So WWV at 10mhz or a ham friend at 7.050mhz should allow for moderately accurate calibration,
down around 10 or 20 hz.

Guess I'd better start playing with some hardware, see if I'm right about how this works.
Will proceed, assuming the github archive is current.

Jerry, KE7ER
 


On Sun, Dec 31, 2017 at 11:14 am, Gordon Gibby wrote:

2.  User calibration, accessed through normal usage of the encoder push button, where he seems to work at keeping you tuned into a friend as you fix your frequency display (adjusting internally both oscillators in compensatory ways?) and then separately set the BFO to desired portion of the skirt..

(see ubitx_menu file) 

 

Join BITX20@groups.io to automatically receive all group messages.