[hpsdr] external 10MHz reference input

[H.A.Meijer]

Hi Brian,
Thanks for sharing your experience concerning the freq jumping of the 
internal 10MHz tcxo.
I have done some research for the 10MHz tcxo (FOX924b) that is used and 
have come to the conclusion
that is a low-cost one, with not very high 
spec'shttp://www.foxonline.com/pdfs/fox924.pdf 
<http://www.foxonline.com/pdfs/fox924.pdf>.
So first thing I will try is to replace it with an other one, I think 
maybe the CFPT-9301 ( http://www.farnell.com/datasheets/43326.pdf )
It has better temp stability then the FOX924b  and the fase noise spec's 
in it's data sheet.

Bert


<http://www.farnell.com/datasheets/43326.pdf>
>
>
>
>     Hi Helmut,
>
>     Yes I know this 'buffer' does clamp the ref. input between 0 and
>     +/- 3.5V.
>     I did try to change R142 to 120Ohm, then I get a kind of
>     symmetrical 10Mhz square wave @+/-2tt,  out of my sinus wave 10MHz.
>     But unfortunately that did not cure the jumping in freq, with the
>     external 10Mhz standard.
>     Yes that 10Mhz standard is stable and does not show me that jumping.
>     I almost start to believe there is a FGPA firmware problem some
>     how not looking correct to the 10MHz reference all of the time.......
>
>
> I spent a month or so on this problem with Apache Labs and with Phil a 
> couple of months ago. Phil spent a LOT of time trying to figure it 
> out. Like you I proposed that there might be a problem with the DPLL 
> as coded into the FPGA but after extensive review and testing, Phil 
> determined the DPLL is working properly. He finally determined that 
> the problem is *definitely* the 10MHz internal reference. It is quite 
> easy to hear the frequency jumps if you listen to a high-order 
> harmonic of the reference oscillator. The problem does not appear when 
> the 122.8MHz clock is free running. It appears that the problem has to 
> do with the digital temperature correction of the reference 
> oscillator. It stays well within its specifications even with the 
> frequency jumps, but the step function used for temperature correction 
> is *REALLY* annoying. The whole process got Apache Labs looking for a 
> better 10MHz oscillator that uses analog temperature compensation. 
> Abhi told me that my 100D was being shipped to me with a different 
> 10MHz reference and my preliminary tests have shown that the frequency 
> steps are gone on the 100D. I don't know but I presume that they are 
> using this new reference on new production radios.
>
> As for the problems with the external reference input, I experienced 
> those as well too. Changing R141 solved the problem. I am now using my 
> LPRO-101 Rb reference without difficulty or problem. Because of the 
> frequency jumping I now use the Rb reference 100% of the time with my 
> ANAN-10.
>
> -- 
> Brian Lloyd, WB6RQN/J79BPL
> 706 Flightline Drive
> Spring Branch, TX 78070
> brian at lloyd.com <mailto:brian at lloyd.com>
> +1.916.877.5067

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20130518/7457252b/attachment-0004.htm>