Since I have one each 7L18, 492 and 494P, I can give you a comparison of the performance and features the way I see it. My 492 has options 1, 2 and 3. This option set is very common, but once in a while, someone will sell a 492 on ebay that does not have all 3. Be wary and ask if you don't see it listed. My comments refer to a 492 with options 1, 2 and 3.
All 3 instruments are more alike than different, mostly compared to the previous generation of spectrum analyzers.
All 3 have a YIG preselector for the microwave frequencies (from 1.5 or 1.8 GHz and up). It's option 1 on the 492.
All 3 have digital storage (on the 492, it's option 2)
All 3 have phase locking of the YIG oscillator at narrow spans (on the 492, it's option 3), which greatly improves the stability of the 1st LO at narrow spans and allows a relatively narrow RBW filter (100Hz on the 492, 30Hz on the 7L18 and the 494P and 10 Hz on some of the 494APs)
Since they use the YIG oscillator and YIG preselector, they use a high 1st IF which pretty much eliminates IF feedthrough, which was a common problem with the previous generation of analyzers (491 for instance).
Now for the differences:
The 492 has similar center frequency accuracy as the 7L18 (a few MHz), where the 494 has a very precise (and accurate) center frequency display down to 10Hz (counter down to 1Hz.)
The 7L18 only goes down to 1.5 GHz, where the 492's go down to 50 kHz and the 494's go down to 10 kHz. I am not sure where the difference between the 492 and 494 comes from since they use mostly the same RF parts.
I wish the 7L18 had an input for an external reference as the internal reference oscillator (only used to phase lock the YIG at narrow spans) is not very good. Even with phase lock, the instrument drifts quite a bit at narrow spans until fully warmed up, which can take a while (the plug-in runs fairly warm after a while). In that regard, the 492 (at least the one I have) has a better oscillator and is stable at the 100 Hz RBW must faster (within 15-20 minutes instead of an hour). The 494 (I have a 494P) is considerably better with regard to stability (great OCXO), with similar sensitivity.
As shown on the pictures here the 7L18's 30Hz RBW is useable. The instrument had warmed up for a while, but the trace was still moving slowly across the screen as I took the picture.
The 494P and AP have a GPIB interface which, when coupled with John Miles' excellent PN and 7470 emulator software, are put to very good use. The 492 and 7L18 don't have outputs (get a digital camera), and the 492P and AP/BP have GPIB but are much more limited in their usefulness because of the lack of precise/accurate control of the center frequency.
The 494/P/AP also have keypad entry for center frequency, span and RBW, so going to a specific frequency is just a matter of punching a few buttons. Compared to the 492 and 7L18, it's night and day. The 494/P/AP also have 10 setup memories, so you can store your most frequent setups and quickly return to those. The 494 also has self calibration and many useful software features, should you ever have to repair it. In that regard, I believe it would be easier to repair than a 492 (depending on the problem), even though it is a little more complex. The 494AP has markers and a few other software bells and wistles, which the 494P lacks. These are nice, but to me the 494P is 95% of a 494AP, while the 492 is 60% of a 494P. Depending on what you do, your evaluation of the feature sets may be different.
The biggest difference between the 7L18 and the 492, in my mind, (aside from coverage down to 50kHz for the 492) is portability and convenience in the shack. The 7000 series mainframes are *BIG*. The 492 weighs a ton, but it's not too big, it is portable (much more so than a 7603 for instance) and sits on the floor with the handle down while taking a minimum of space and while all controls are quite accessible. Otherwise, performance is roughly comparable (the 492 is a little more sensitive). The 492 is more stable, but the 7L18 has a 30Hz filter. Both have digital storage, memory and on-screen display.
If your interest lies with HF, VHF and UHF only (no microwave), you may want to consider the 7L14 plug-in or the 495 or 496. These are low band only versions of the other instruments (7L14->7L18, 495->494, 496->492). I believe aside from the YIG preselector and frequency coverage, a lot of what I said above is applicable to these instruments.
One more gotcha: some 492 (I forgot which option that is, maybe option 8?) only go to 8 GHz instead of 21 GHz. If you are into ham radio microwaves at 10 GHz, stay away from that option. Fortunately, they are fairly rare, but just be careful.
In summary, they are all fine instruments, and priced according to their usefulness. While I have quite a bit of older test gear, I am not a collector in the common sense. I am not interested in stuff just because it's old and it was made by Tek (I mean no disrespect to true collectors, I admire the hobby, but don't have the interest to pursue it myself). The equipment I am looking for has to be useful and relatively convenient (within the limits of what I am willing to spend for convenience).
The 7L18, 492 and 494P are quite respectable instruments, even compared to modern gear, in terms of sensitivity, spurs, stability and relatively fool-proof operation. Older style SAs, such as the 1L and 5L series and some of the earlier 7k series, to me, are just antiques. They may well be quite useable and useful, they are too far (in performance and convenience) from the instruments I use at work everyday so that going from one to the other would be painful. If I were not exposed to modern gear in my work, I might well be quite satisfied with one of those. When switching from an HP 8565E to a Tek 494P, I don't feel cheated (unless I need to go to 50 GHz :-)
My recommendation? It depends on what you want to do and how much money you have available for this. In my case, after actually using a 7L18 for a while, then a 492 for a while (all the while using the fancier stuff at work), I broke down for a 494P.
When associated with an inexpensive GPIB/USB controller such as sold by SparkFun Electronics, you can have a lot of fun and also do useful work of course.
PS: Here is a message from John Miles giving more info about how the 492 and 494 differ internally. I hope John won't mind...
John Miles wrote: > They have very similar RF guts, but functionally, they're very different > beasts. Daylight and dark. > > The 492 and 494 have the same RF hardware, and they both tune to 21 GHz. I > don't know where the "18 GHz 492" mixup got started, but it has no basis in > reality as far as I know. Maybe they built a few early models with > different specs, but I've never seen one in the wild, and I've seen a lot of > 'em. > > The 494 can be tuned by direct keyboard input; the 492 can't. > > The 494 can store settings and displays in non-volatile RAM; the 492 can't. > > The 494 can count the applied signal frequency at 1 Hz precision all the way > to 21 GHz; the 492 can't. > > The 494 is tunable at 1 Hz precision; the 492 is tunable at 1 MHz precision. > > The 494 is accurate to within 100 Hz or so; the 492 is accurate to within > +/- 5 MHz. The 494 has a high-quality OCXO internal standard; the 492 > couldn't use one if it had it. > > The 494s have a poorly-designed 30-Hz ovenized filter that you will probably > have to carve open with a Dremel tool and repair before it will > self-calibrate. The 492 doesn't have the ovenized filter, and doesn't > self-calibrate at any rate. > > GPIB functionality, including HP-GL/2 plotting, is much more advanced in the > 494 than the 492. The 494A adds markers and quite a few more firmware > features to those above. > > If the price differential isn't excessive, a 494 is a much better analyzer > in all respects. > > -- john, KE5FX > > >