OR you could use a small PC type IC cooler fan blowing at the heat sink from a short distance away, driven from a Lady Heater type variable speed fan active PID Temp controller. That has NO trouble holding 0.01 Deg C at where ever you place the Temp sensor, over normal day / night room temperature variations, even with it all out on an open table without a box, insulation, OR swimming pool. ws
***** Try an insulated water cooled metal block for the baseplate heatsink The water can be recirculated and heated/cooled as required. With a recirculating system and say 25W dissipation with no temperature control you will only need about 5 cubic meters (5 tons) of water to maintain a temperature rise of less than 0.01C for 3 hours. Assuming that the 25W rubidium dissipation is the only significant source of heating for the water. That means you need a well insulated swimming pool in your basement. Controlling the temperature of a smaller amount of recirculating water is probably a simpler proposition. Bob Camp wrote: > Hi > > Water might work. It would take quite a bit of it. > > Here's my "wild guess" level math: > > 1) The basement moves 0.1 to 1 C short term / over a day. > > 2) I want to get to< 0.01 > > That takes the time constant out to>= 10X the time I'm interested in. > > 3) The time period of interest is 3 to 30 hours. > > That gets to a time constant of at least 10 days. > > At the same time you have>10 watts coming out of the gizmo. You can't put the > thermal mass inside a vacuum bottle. > > I suspect that some combination of thermal mass and active stabilization will > be needed. > > So much fun .... > > Bob > > > On Dec 23, 2009, at 2:18 AM, Don Latham wrote: > > >> sheesh! How about a right-sized water jug? >> Don >> >> ----- Original Message ----- From: "Bob Camp"<lists at cq.nu> >> To: "Discussion of precise time and frequency measurement"<time-nuts at >> febo.com> >> Sent: Tuesday, December 22, 2009 8:23 PM >> Subject: Re: [time-nuts] Cheap Rubidium >> >> >> >>> Hi >>> >>> I agree that if you simply bolt the rubidium to an old engine block and >>> toss a blanket over it, you might get some pretty good thermal stability in >>> the "hour to couple hours" time period. That's certainly a better approach >>> than putting some kind of DC heater (and it's varying magnetic field) near >>> the rubidium. >>> >>> I'm still wondering if they do indeed hit 1x10-13 (as in almost 1x10-14) or >>> not. I suspect not. I'm sure that they do indeed get into the 1x10-13's, >>> just not sure they get to the bottom of that region. >>> >>> Bob >>> >>> >>> On Dec 22, 2009, at 8:26 PM, Bruce Griffiths wrote: >>> >>> >>>> Bob Camp wrote: >>>> >>>>> Hi >>>>> >>>>> If I randomly pick up a FE 5680A data sheet, I find that it's short term >>>>> stability is 1.4/sqrt(Tau) x 10-11. Since I never doubt anything I see >>>>> on a data sheet, this immediately tells me I should get 1.4x10-12 at 100 >>>>> seconds, and I only have to wait for 10,000 seconds to get to 1.4x10-13. >>>>> >>>>> Since the temperature performance is at the 1x10-12 / C level, I would >>>>> need a room that's stable to *much* better than 0.1 C over a 3 hour >>>>> period to get there. I suspect that 0.01C might not be good enough ... >>>>> >>>>> So here's the question: >>>>> >>>>> Has anybody run any of the cheap rubidiums (FE or Efratom) in a *very* >>>>> stable temperature environment to see how close they get / what the floor >>>>> is? I've run through a lot of data on the web, but I haven't really >>>>> found what I'm looking for. >>>>> >>>>> Thanks! >>>>> >>>>> Bob >>>>> >>>>> >>>>> >>>> Figure 7 on the FE5680 page (also on the data sheet) indicates that you >>>> may need somewhat less than 3hours to achieve ADEV ~1E-13. >>>> 0.01C stability should be adequate.however its not necessary to control >>>> the room temperature to this stability if the FE5680 is in an enclosure >>>> with a sufficiently high time constant whilst having a sufficiently low >>>> thermal resistance so as to avoid overheating the FE5680. >>>> >>>> Bruce >>>> >>>> >>>> _______________________________________________ >>>> _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
