[time-nuts] About HP10544A

Esa Heikkinen scifiscifi at sci.fi
Mon Mar 9 21:51:19 UTC 2009


> Since retrofitting an improved oscillator circuit isn't really an option
> you will need to filter the output to reduce the harmonic content.
> Try a bandpass filter driven by the buffer and terminated in 50 ohms.

Well I have to decide what to do, get another oscillator or try the 
filtering. The difference to another oscillators (like tbolt ocxo or 
LPRO) is so huge that I do not know how hard it would filter the 10544A 
to the same level, which parts to use and how much it will cost.

> Should you retrofit an improved oscillator circuit you may as well
> replace the oven controller to eliminate the oven switching frequency
> related sidebands.

Infact the switcher sidebands are now gone:
http://www.amigazone.fi/files/gpsdo/544-13.png
http://www.amigazone.fi/files/gpsdo/544-14.png

It was like thios earlier:
http://www.amigazone.fi/files/gpsdo/544-1.PNG
http://www.amigazone.fi/files/gpsdo/544-2.PNG

It was easier than I expected... When I took the output directly from 
OSC pins without using the coax connector like you suggested the 
switcher peaks was gone!

So I had a closer look to the PCB today and noticed that it has only 2 
layers and the ground net is too thin and goes around PCB totally wrong 
way so that switcher current seems to flow via the signal flow. It's PCB 
layout design fault, I think that with correct layout design this could 
be done even with 2-layer PCB correctly.

But this is not a problem of course because I will design my own PCB for 
the final system anyway, having SMA connectors for 10 MHz and EFC. I'll 
also create separate power supply for ovens and signalling stages.

-- 
73s!
Esa
OH4KJU



More information about the time-nuts mailing list