Let me see if I have this straight. You want to know if certain cap positions in a XO are inherently more sonically important than others. Mostly, because you are on a tight budget. Then, the initial post is followed by more questions up to the point were it sounds like you are interested in playing around with the XO... More of an educational exercise? While there is absolutely nothing wrong with this, there will be a conflict. Education, and cheap do not coexist in audio. And do not confuse less expensive with "cheap".
Thanks Jeff. I appreciate your welcome and your responses. Also, no experimentation or playing with the crossovers is planned. Just a straight substitution of like-for-like. And no problems with less expensive. I like bang for the buck.
Now, to answer a few of your questions (direct or implied):
1) Series caps typically influence the sound with greater magnitude over a larger bandwidth than paralleled caps in a paralleled network.
Great. That's what I was hoping to hear.
2) Caps in paralleled RLC networks tend to have the least amount of impact on the network with the exception of parallel cap operating no higher than 300Hz.
Again, nice. Good about the 300Hz cut-off too, as I didn't want to have to use anything exotic for the 350uf cap.
3) Optimizing a crossover often calls for different brand and construction caps.
Good to know. That allows for some tuning to the strengths of the different drivers.
4) While bypassing (XOs) was a very viable thing 15-20 years ago, it is seldom the solution today. This is due to the quality, and shear number of different flavors, available today.
I wasn't expecting this. I had assumed bypass caps were more accepted, and the norm? Aren't the SoniCap Gen II caps bypass caps?
A few things to note. That speaker sounded good with "all" electrolytics. The age of the caps has caused their ESR to rise, so not they sound darker, slower, and have less detail. Replacing all the electrolytics with new electrolytics would be the best bang for the buck. If you are trying to stretch your money, trial and error (an education) don't live here. I have had a couple of poor experiences with erse electrolytics, so I no longer recommend them.
I'm not opposed to using NPE's. It's just that I've read comments from retired AR engineers, stating they would have used poly caps in their crossovers, if they'd been available then. In fact, I've recapped speakers, using NPE's, and been happy with the results. Just looking to 'step-up' to the next level, if the 90's are capable of that? Which I think they are.
I'd be happy to make an official recommendation, but I need a couple of things first. Budget? Top six sonic improvements you are looking for? If you have less than six, that is OK.
Budget? Ha! I sometimes feel as though I work for the U.S. Government, as my funds seem to always get reallocated to another project, without my prior knowledge or approval. So, I usually buy a little at a time, when the mood arises! In secret!
Sonic improvements? The 90's are a really good vintage speaker, but a little more overall detail, especially in the lower mids, would be good. Plus, a bit more air, and openness in the upper mids, lower treble, would be welcome.