We have a number of HP89410A spectrum analyzers in our lab, which have been just great general-purpose workhorses over the years. But they are growing old, and Agilent no longer wants to repair them. I have been looking for a more modern replacement, but to my surprise I haven't found any other equally well performing one. I suppose there may be a workshop somewhere (preferably in Europe, if you happen to know one) which repairs obsolete instrument models, although I haven't searched for one yet.
Anyway, I'd be interested to hear if SED'ers have in mind something to recommend.
I have been primarily using the 89410's for amplifier noise measurements - it is dc coupled and performs quite well down to sub-Hz frequencies while its top frequency is the rather sufficient 10 MHz. There are audio analyzers with decent low frequency performance (SRS and R&S among others), but their top frequency seems to extend only to100 kHz or so. Actually we used such audio analyzers before getting our first HP89410A in 1994, but I'd prefer not to downgrade to those old days. In fact, I *could* use even slightly larger top frequency, like 20MHz or 50 MHz, if that's available. The true rf spectrum analyzers tend to have their frequency range starting from 3 Hz or so. I'd like it to extend below 0.1 Hz, preferably down to 0.01 Hz.
One nice feature of the 89410's is that they have a built-in source, with a pseudorandom noise option. It is very convenient to first measure the noise, then check the gain with a sinusoidal excitation (you can check the harmonic generation simultaneously) and finally the frequency response with pseudorandom noise. Our units are equipped with two input channels, which gives the capability to perform (i) the true amplitude-phase transfer function measurement and (ii) cross- correlation between the two channels - however these two-channel measurements are not used quite so often.
When both the source and the analyzer are combined the instrument does not take as much workbench space as two separate units would. The best thing about our 89410's is that they are reliable and easy to use. Just hoist the analyzer to your table, power it up, and you can pretty much rely on the nV/rtHz reading it gives, without having to do RBW calculations in ones head or any such a thing.
I have tried to get comfortable with using a Rigol DG5252 arb generator together with either Rohde&Schwarz spectrum analyzer or the Cleverscope CS328A. The combination just isn't convenient. You have two units on the table, plus a PC in case of the Cleverscope. The R&S menu logic comes from the rf world, even if it was ultimately possible to set it up for log-frequency plotting (so that you recognize the poles and zeros conveniently) and some other units than dBm - besides the R&S has 50ohm only inputs and its 20GHz top frequency is really an overkill in my appliction.
Nowadays I could roll my own instruments out of various commercially available modules, but the software with a decent user interface would take quite a lot of effort to develop. I'd prefer to avoid that, although it is an option if all else fails. I'm astonished if no-one really makes anything resembling the 89410 any more. Actually we have a VXI-based HP89640 which pretty much lies in the lab unused because it is so clumsy to move around and operate. I've been playing with the idea to equip it with a permanently attached tablet-like computer, but (i) that instrument, too, lacks the source, and (ii) in a quick test its low-frequency performance wasn't that wonderful although the inputs can be made dc coupled.
Have you guys encountered anything which would more-or-less fit the bill?