PXI Atomic clock

January 25, 2013 1:26 PM

Does anybody make a Rubidium Beam Atomic Clock that I can plug into my PXI Chassis to discipline its 10 Mhz clock to approximately 10^-10 parts. I don't want a card that I can put a external 10 Mhz clock into, I simply don't have the space in my system.
posted by Confess, Fletch to Science & Nature (9 answers total)

Hmm... as for your specific request, I can't help you, but... Google uses GPS to maintain it's systems clocks to a very high degree of accuracy. Dunno if that is helpful or not.
posted by Freen at 2:38 PM on January 25, 2013

Kyocera

I just found this with a google search for "High Accuracy Oscillators"
posted by Long Way To Go at 3:06 PM on January 25, 2013

I doubt this comment/link will be particularly helpfull, but it might be of interest.
posted by Packed Lunch at 3:21 PM on January 25, 2013

This is sort of up my alley. Care to memail details?
posted by fatllama at 9:31 PM on January 25, 2013

Details here, too, I might also be able to offer a suggestion.
posted by drhydro at 10:17 PM on January 25, 2013

This question belongs over on the Time-Nuts mailing list. I don't have any PXI experience, but I've recently been elbows-deep in a cranky rubidium physics package, and those folks helped me out big-time. There's no better bunch of freq-freaks.

In general, I'm gonna guess that an FEI-5680A or another similarly thin module could probably be wedged onto a PXI carrier board, but the task of disciplining the system clock against it would be still unsolved.

Best of luck! I'm gonna fave this and check back, please update it if you find a solution. :)
posted by Myself at 1:28 AM on January 26, 2013

The one from Stanford Research Systems is used by quite a few people. They have a good reputation and I am happy with most of their other equipment.

I don't have first hand experience with their rubidium clock, but I've checked the specs and also discussed some details of it with their engineers at a fair. The only extra feature I would have liked to see is the option to float some of the outputs. If you like, I can talk to a co-worker who's used it extensively.

Maybe you already know all about this, but some stand-alone instruments (e.g. Tektronix AWGs) don't work well unless they are allowed to source the clock for the whole system. I don't know of any PXI modules acting like that. However, at least Aeroflex claim that the chassis clock bus itself (even when externally sourced) isn't up to sufficient standards. They recommend daisy-chaining the clock with coax between the front panel connectors of their modules instead of using the backplane.

There might also be an issue about feeding an external clock through the chassis connector, that it's better to route it in through the front panel of the module in the Star trigger slot. This I'm not sure of though, it's just a very vague memory.
posted by springload at 6:32 AM on January 26, 2013

Or wait, did I misunderstand you? Do you want an external clock to drive your PXI rack, or a clock in the shape of a PXI module because you're out of space in the surrounding lab?
posted by springload at 6:47 AM on January 26, 2013

Response by poster: I have a system in a 19 inch rackmount that is very full. My system has a NI PXI chassis that is not completely full. Also in the system is a SRS DS360 that does not generate a stable enough signal for my purposes. I thought of one of the aforementioned SRS FS725 and I really like it but I'm not sure of where in my system I'd put it.

I thought if I could find a PXI mountable Rubidium Clock I could then send its output to my DS360 and also sync my PXIs's internal 10Mhz clock to improve it's accuracy. I'd also thought of the GPS disciplined system, but it doesn't really get me out of the 'where do I put it' problem.

I think I'm just stuck with the SRS FS725 and making some horribly klugey shelf inside my rackmount cabinet or xip-tying it to the door or some other wacky idea.

Thanks everybody.
posted by Confess, Fletch at 8:19 PM on January 26, 2013