This is a complicated question. First of all, there is definitely no crystal filter involved. However, the multiplier can have its own phase noise due to either the amplifier or the SRD. There can also be AM to PM conversion and vice versa. I have found the technique of multiplying a source to measure its phase noise to be somewhat unreliable. It should however at least give you an upper bound on phase noise, since the multiplier will always increase noise by at least 20 log N, if you are close enough to the carrier to avoid filtering effects.
I should also mention that after the techs got done with their sweeper tests, they reconnected the DUT to an HP 608 signal generator as the drive source and did a phase noise test. Zeta had one at every bench, dozens of them. Even in 1975, the 608 was getting long in the tooth. However, at the time, nothing else had the low phase noise floor of the 608, even the amazing HP 8640, the latest and greatest out of Bill and Dave's garage. To be sure, the 8640 was better at 20 kHz offset, but at 1 MHz offset and especially 10 MHz offset, no tunable signal generator could touch the 608. Even today, no tunable signal generator to my knowledge has ever eclipsed the 608 in phase noise floor. We needed the low phase noise floor on the generator so that we could measure the phase noise floor of the multiplier as opposed to the drive source. Even with the 608, you had to be careful what level you drove the multiplier at. That could affect the phase noise. At certain drive levels, it would get enhanced. Basically, the Of course, if you don't need tunability, you can just use a 100 MHz crystal oscillator like they had in the "bricks". In case you are wondering, the reason why the 608 could beat everything else is that it had a "post selector" filter at the output AFTER the amplifier. If you have ever used a 608 you will remember the knob you had to peak after any frequency change. This was to tune the post selector. This gives it an advantage that is impossible to overcome. For whatever reason, no other sig gen has a post selector filter. Rick Karlquist N6RK Matt Ettus wrote: > Thanks Rick. My main reason for buying this is to be able to better > measure the phase noise of my 100 MHz oscillator. I figured this > would give a truer measure than a brick since there is no PLL > involved. But then I thought there might be too much filtering going > on. If the filters are at least several MHz wide, then it should be > fine, but if there is a 100 MHz crystal filter in there, then it won't > really give me a true measure. > > What do you think? > > Thanks, > Matt > > > On Mon, Apr 28, 2008 at 4:15 PM, Rick Karlquist <[EMAIL PROTECTED]> wrote: >> I worked for Zeta Labs many years ago. This multiplier >> undoubtedly has a step recovery diode that is being driven >> at a level of a good watt or two. In most cases there was >> a hefty transistor to make this power. You probably need >> +10 dBm or so to drive it, maybe as much as +20 dBm. Take >> the cover off and look at the input section and it should >> be apparent if there is any additional gain. You can always >> start at 0 dBm and work up, looking at the power out as you >> go along. The adjustment of these can be very tricky, so >> proceed with extreme caution. Especially the SRD assembly. >> There were rows and rows of technicians with sweepers who >> tuned these up. The sweepers would sweep typically 80 to >> 120 MHz. Some of the techs liked to listen to the radio >> as they worked, and since this frequency range encompased >> the FM band, the radio would make a woodpecker like sound >> all day long. >> >> The Zeta multipliers were fairly similar to the multiplier >> in the well known "brick" frequency sources (California >> Microwave, etc). Many people worked at both companies at >> various times. >> >> Rick Karlquist N6RK >> >> >> >> >> Matt Ettus wrote: >> > Just bought a >> > >> > Zeta Labs X76 Multiplier 7600 MHz Output Model 5856-01 >> > >> > on ebay. Anybody have any info on these? What kind of drive should I >> > give it? >> > >> > Thanks, >> > Matt >> > >> >> >>> _______________________________________________ >> > 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.