Somewhat related to this, someone posted a link for 1U chassis a while back. I thought I saved the link, but can't find it, and since the text of posts can't be searched from the archives, can't find mention of it that way either. If anyone has a decent source for 1U chassis that I could mount a pair of tbolts in, I would appreciate it. These were generic 1U rackmount chassis - not 'server' chassis with a bunch of preplanned holes.
Brent On Fri, Apr 22, 2011 at 5:37 PM, WB6BNQ <wb6...@cox.net> wrote: > Hi Pete, > > As Bob, K6RTM, pointed out the Thunderbolt and the Rubidium are two different > animals all together. > > True, you can treat a Rubidium like it was a normal crystal oscillator, but > it is > not the same. The Rubidium has a definite life span, the more you run it the > less > the life. A high quality crystal oscillator, on the other hand, just gets > better > the longer you leave it on. Aside from nominal electrical component > failures, the > crystal blank in a properly designed circuit has no short term failure > mechanism and > will last for decades with constant applied power. The Rubidium’s life span > is, at > best, 10 years. The question is how long was it running before you got it ? > > A high quality crystal oscillator has excellent short term specs but does > have drift > and aging functions that severely limit its use for long term purposes. Long > term > meaning more than a few hours for the best. That is where the Rubidium > oscillator > takes over as its drift function is measured in days to a month or more. > > For high quality measurements, the crystal excels for measurement times of > less than > 10 seconds, as the Rubidium is noisier in that time frame. That is, for > taking > readings on a one second to second basis, such as with a high resolution time > interval counter, the crystal excels. However, if the period of the > measurement is > longer, then the Rubidium would be a better choice. For portable purposes the > Rubidium also excels as its retrace is much better than a crystal oscillator. > You > also do not need to wait the thirty to sixty days for the crystal to > stabilize. The > Rubidium will be very close to its original set point in about 20 minutes. > > Adding GPS to mix has its own issues. First, you need to know the coordinates > precisely or spend a couple of days getting a damn good fix. The GPS is > quite noisy > in the short term and the oscillator that is steered by the GPS has that > noise show > up in its output. That is mitigated by having a high quality crystal > oscillator > where the GPS control loop seldom makes corrections; perhaps once an hour or > more. > That is how the Thunderbolt works and depending upon its internal crystal > oscillator, it may possibly be tweaked to perform better then the standard > factory > settings. > > As for use, it all depends upon what and how you’re making measurements. > With a > nominal 8 or 9 digit counter, for example, you may not notice all of the above > issues because they are typically beyond the resolution of the equipment in > most > cases. In other measurement processes it may be of major concern. > > As for your project boxes, I would use the rack mounted box to house the > Thunderbolt, distribution amps and perhaps a couple of other oscillators > (like the > hp 10811) along with quality power sources. Because crystal oscillators like > a > constant operating condition, do consider battery power for the lab to handle > those > occasional mains power drops. > > I would use the portable box for the Rubidium oscillator and include a battery > option depending upon your intent. The emphasis should be to have very quiet > and > stable power supplies for both projects. Even batteries have a fair amount > of noise > so make the mains power (and battery) voltage high enough to allow for > running a > quality regulation circuit. > > My two cents ! > > 73....Bill....WB6BNQ > > > g4...@btopenworld.com wrote: > >> I have a dilemma and wish to access the collective wisdom of the group to >> advise >> a solution. >> >> I am building a clock generator based on a Thunderbolt. I have an LPRO and >> would >> also locate this in the same enclosure. I will also add a distribution amp >> and a >> divide chain in due course. >> >> The ultimate purpose of the set up is to provide a self contained clock >> generator set for my other test equipment, and also an experimental >> workstation >> for Rubidium and GPS disciplined experiments. >> >> Most of my other equipment is for 19-inch rack mounting. >> >> I have two potential solutions for housing the timing kit: >> >> 1. An old dismantled HP 4U scope chassis which will fit in with my other >> equipment physically, and can be racked if necessary. The PSU would have to >> be >> built into the same enclosure. >> >> 2. A pair of Anritsu instrument cases which once house a bit error test set. >> The >> two units clip together beautifully, and are free standing. As there are two >> units, this solution would allow me to build the PSUs in one case and the >> more >> sensitive timing electronics in the other. These units cannot be racked on >> account of their form factors. >> >> Both solutions will require me to do some bespoke metalwork, but that is no >> problem for me and amounts to about the same amount of work for either >> solution. >> >> So what does the group advise? Is it vitally important to keep PSU components >> isolated from the timing electronics? I want to create the least noisy clock >> source given the components I have. >> >> Looking forward to hearing some opinions... >> >> Pete >> G4GJL >> _______________________________________________ >> time-nuts mailing list -- time-nuts@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >> and follow the instructions there. > > > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > _______________________________________________ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.