Quote History Quoted:
I looked for an amplifier / active splitter and never found a good one that was reasonably priced. I did the math on my Trimble Thunderbolt's output, the loss through a passive splitter, and the signal requirements of my gear, and everything was in spec so I went with a passive splitter and it works great. The Thunderbolt is feeding My Flex 5000, a HP spectrum analyzer, HP frequency counter, and two Fluke signal generators. I can't feed anything else but what I have works great.
I'd be interested if you find something good so I can have a signal available for other stuff if needed.
View Quote
if the passive approach works, use it. quality 2 way and 3 way splitters suitable for use at 10MHz are available from mini-circuits and others.
if a passive approach doesn't work, there are generally two options:
1) daisy chain the instruments.
high end scopes and the like (such as you list above) have a REF IN and a REF OUT bnc. if the signal at REF IN is valid, REF OUT will be a buffered copy. this way you can take a single 10MHz reference signal, connect it to the REF IN on the HP SA, then use the REF OUT to HP counter REF IN, and so on. in general the REF OUT bnc on HP/Agilent gear provides a good enough copy for 99.9999% of typical benchtop use. (where we don't use this approach in our performance lab is with network jitter/wander testing used for formal equipment certification tests; here the 10MHz feed from a cesium reference ("atomic") clock is applied directly at the test instrument without intervening buffers/amplifiers of any type.)
2) use a distribution amplifier.
in theory, a good reference frequency distribution amp will have a gain of 0db at 10MHz (or perhaps +1dB to help with cable losses), will not add jitter or spurious noise (sidebands) to the outputs, and will also provide infinite input-to-output and output-to-output isolation so noise introduced anywhere does not propagate through the system. in practice, though, all amplifiers (even unity gain) contribute phase noise and spurious noise, and in addition have finite isolation characteristics. herein lies the rub: it is essential that you understand that active components do "things" to a signal which you are expecting to be a reference (that is, very high accuracy, and very low jitter (=phase noise), wander, and drift). whether or not the degradation introduced by the distribution amp is detrimental to your application is completely under your purview and related to what you are trying to accomplish.
for most amateur radio applications, inexpensive distribution amps will provide excellent results. for the most part, the parameters i would most be interested in knowing from an amp vendor are phase noise (this will show up in SSB and FM if it is bad enough) and output-to-output isolation (one of the "distribution clients" can pull the other clients around, or pollute their references). unfortunately, minimizing phase noise and maximizing output-to-output isolation is hard to design and expensive to build.
just as an academic side note, the specifications of laboratory reference distribution amps from Symmetricon (recently purchased by Microsemi) are linked below. none of the Symmetricon hardware is in the realm of affordability for typical amateur radio use -- however it is useful to compare specifications with low cost or DIY kit distribution amps.
http://www.microsemi.com/products/timing-synchronization-systems/time-frequency-distribution/1u-distribution-amplifiers
PDF spec sheets:
http://www.microsemi.com/document-portal/doc_download/133406-6502b
http://www.microsemi.com/document-portal/doc_download/133413-4036b
one final note:
you can do yourself a big favor by terminating (with 50ohms) the unused outputs on any distribution amp. this prevents internal reflections on the amplifier and generally results in quieter reference outputs. BNC's with integral 50 ohm terminations are cheap and effective.
ar-jedi