[hpsdr] Mercury V2.9 available

[Phil Harman]

Hi All,

Mercury V2.9 is available for testing.  You will find it here;

< svn://206.216.146.154/svn/repos_sdr_hpsdr/trunk/Mercury V2.9 >

and you can use the usual update process from here;

< svn://206.216.146.154/svn/repos_sdr_hpsdr/trunk/USBBlaster-Binaries >

I recently moved to Quartus V9.1 so have added that programmer selection to 
the menu.

Mercury V2.9 adds a new feature for those using Excalibur as their 10MHz 
reference clock.

Mike, N3UC found that by using PLLs, rather than digital dividers, to divide 
the 122.88MHz  master clock  and 10MHz reference clock to 80kHz then the 
frequency accuracy of Mercury improved considerably.  In fact Mike is 
acheiving an accuracy of 2mHz  (that's two milli Hz !) using an external GPS 
locked reference.  Using this technique he came third out of 97 participants 
in the recent ARRL Frequency Measurement Tests - congratulations Mike.

V2.9 uses this PLL technique.  In order to use it you need to connect a 
short coax lead between  Excalibur J2 and Mercury AUX_CLK ( J8 - located 
directly below the FPGA). Unfortunatley we can't pick up the 10MHz signal 
from Atlas C16 since the PLLs will only work from a dedicated FPGA clock 
input - hence the need for a short coax lead.

LEDs  3 and 4 respectivley on Mercury indicate that the 122.88MHz and 10MHz 
PLLs are locked.

Be careful when connecting the coax lead to J8 on Mercury since this pin is 
connected directly to the FPGA - usual practice, make sure the power is off 
and follow anti-static precautions.

If you would like to use this feature permantly then disable the 10MHz 
TCXO's on Mercury (and Penelope) by removing power to these devices using 
the links and jumpers provided.

To use this new feature just select Excalibur as the 10MHz clock source in 
PowerSDR or KISS Konsole.  If the connection between J2 and J8 is made then 
Mercury will automatically select the PLL method - if not present it will 
use the 10MHz clock from Excalibur on Atlas C16.

I also found that the wrong Cordic code had been used in V2.8.  The latest 
version rectifies that and will produce a few less spurs.

As usual, please report your findings via the reflector.

73's Phil...VK6APH







 1258800423.0