I think if you have a SWR meter and just need to tune a few HF antennas you probably should just stick with the SWR meter. I think the analyzers have made people obsess a bit about true resonance, that is X=0, when the best adjustment for many antennas just needs to present low VSWR to the coax cable.
People will go on and on about how "low SWR doesn't mean it radiates," and that's true. But getting the SWR as low as possible by trimming/tweaking for an antenna
where you expect low SWR and for which you don't have HUGE coax cable losses is just fine, and almost always leads you to the optimum. If you put up a dipole or a multiband vertical and stick to sane adjustments like trimming the dipole or adjusting the lengths of the vertical sections within their design range and a shunt matching coil, you will do just as well with a SWR meter. There are only a few antennas where simply adjusting for low SWR is going to add lots of loss, and you won't necessarily be able to notice that even if you have an analyzer. It's kind of a pathological weird situation.
It can be much FASTER to find the lowest SWR on mobile antennas, bigger shunt matched verticals, beta matched yagis, and other antennas with shunt matching elements. You can leave the shunt coil out, tune for a particular value of X (usually something like X= -j10 to X = -j20 for most situations) and put the shunt coil back in and only fiddle with that to get lowest SWR. But if you juggle element length and coil stretching/compressing you can ALSO find the lowest SWR point. It will just take longer.
For most HF antennas, especially simple direct-feed like dipoles or more complicated commercial antennas, a SWR meter is really all you need. Even at VHF it's okay as long as you make sure you're shooting for very low SWR. At VHF cable loss can mask extremely high SWR, but if you're unsure, measuring the SWR at the antenna can help there. That isn't even necessary at HF. IF you are putting up a dipole, hook the SWR meter up at the shack end, trim for the lowest SWR across the band, and get on the air.
A MFJ analyzer is an incredibly versatile tool and a good value but I almost never use it to adjust anything for exact resonance
It's just a good RF multimeter. I would be lost without it, but that means I also kind of have a good idea what I NEED it for and what it's just nice for. I use mine to measure coils, capacitors, transformers, and traps as much as I use it on actual antennas as finished. In fact, with a combination of EZNEC, careful cutting, and the MFJ-259B for pre-tuning matching networks, I haven't had to trim an upper-HF antenna at all in a long time.
I still trim single element antennas like dipoles because it's faster than careful modeling. It is nice to not have to run in and out of the house when trimming a dipole or adjusting a vertical, and it's nice to have the '259B when you need to set five matching networks for a multiband antenna (
http://www.n3ox.net/projects/stepperswitch) but maybe you live with that while you're saving money a little at a time, or you borrow an analyzer if you're a set-and-forget kind of guy.
The analyzer can be a real time-saver if you're experimenting a lot, and allows a lot of room to grow by allowing you to measure coils and caps for homebrew matching networks and so forth. But if you're just installing a few antennas and doing light experimentation with simpler types, the cost vs. usage tradeoff might not work out so well for you. There are a few resonance zealots who will make you feel like you can get a better signal by knowing when X=0 but those people don't have a thorough understanding of the role of resonance in antenna systems and how VSWR relates to that in HF systems.
