[time-nuts] BC637PCI

GandalfG8 at aol.com GandalfG8 at aol.com
Sat Oct 11 14:42:49 UTC 2008

In a message dated 11/10/2008 05:12:11 GMT Daylight Time, smace at intt.net  

Has  anyone "upgraded" the oscillator on an older Datum BC637PCI card to
the MTI  crystal?  Is it as simple as changing the crystal and adjusting  the
osc gain?

Hi Scott
I've not changed the oscillator, but I did look into the possibility and  
also ran a BC637PCI for a while with an external Vectron ovened 10MHz  oscillator.
I assume you have the BC637PCI manual?
Appendix A refers to field upgrades, both converting a 635 to 637 by adding  
the GPS module and upgrading to the ovened oscillator.
In both instances a firmware "upgrade" was activated by entering a device  
specific password provided by Datum/Symmetricom. I doubt that option  is still 
longer available even though the necessary firmware changes are  obviously 
already there waiting to be activated.
I suspect, in the case of the oscillator upgrade anyway, some default  
settings including oscillator gain will be all that's changed.
When running with an external oscillator I found that loss of power would  
default any settings I had changed so no doubt that's what would happen if you  
upgraded the oscillator without the firmware change.
Page 6-15, in the rev K manual anyway, gives details of the oscillator gain  
adjustment, I found approx 70 seemed reasonable with my Vectron oscillator  
but my BC637PCI, and perhaps others, has some conditioning issues  anyway.
The basic onboard oscillator was not very good, apologies for such an  
imprecise statement but I can't find the fairly extensive notes made several  months 
ago, both accuracy and stability were relatively poor measured on an  
HP53132A using a known good Thunderbolt as a reference.
In particular I noticed quite frequent, but seemingly random, jumps in  
output frequency that would then take a little while to recover, a bit like a  fast 
acting automatic gain control with slow recovery time. These jumps  were 
significantly greater than the "normal" variations I observed and  fell outside of 
the unit specification.
Using the Vectron oscillator greatly improved things, it was a much better  
oscillator anyway and I was able to observe the significant  improvement in the 
Vectron performance when being conditioned by the  BC637PCI.
However, the frequency jumps still occured albeit with overall performance  
much improved and frequency stability much better and the jumps less than  
before, ie frequency jumps were seemingly reduced in proportion to the better  
Without the jumps I would have been happy with the Vectron performance  but 
didn't feel able to trust the combination without permanent  computer 
monitoring, and still don't know if this was a function of my  BC637PCI or something 
common to all of them.
It was interesting to play with, and I'll probably go back to it at some  
time, but I was more concerned then to set up something I could rely on and  
eventually put it to one side.
Whilst working with the BC637PCI I did find that I needed both the  BC635PCI 
and BC637PCI software, both offer options the other lacks, and I have  late 
copies of both if you need them.
I also have two manuals, version H from Datum that includes schematics, and  
the fairly similar version K from Symmetricom but without schematics.


More information about the time-nuts mailing list