I think the best option for experiments where you hope for salvageable outcome would be single lenses and symmetrical-to-nearly-symmetrical doublets. Next, maybe asymmetrical doublets like a Wide Angle Rapid Rectilinear or similar by other names I can't remember at the moment...aplanat?
We've been warned by many that even the seemingly simple Cooke triplet and it's grandchildren have very critical spacing...because it accomplishes alot through careful design and assembly. Through careless assembly (found optics, guessed spacing), what was accomplished won't be. If you have the parts, experimenting is cheap.
I'm pretty sure the idea behind additional elements was to accomplish additional aberration correction. At some point someone realized spacing, curvature and index of refraction all can be combined to correct to a large extent most of the 7 most common aberrations troubling lens design. There is a patent that discusses some classic designs and assigns a numeric value to the amount of correction accomplished in each design by the three techiques mentioned above. Brian Caldwell, optics and software designer has mentioned in some cases exotic glass can be done away with by altering spacing.
Cool, but beyond most home experiments.
4-element lenses like Tessars are a variation of a triplet, with one element 'split' into two to allow specific corrections.
The beauty of symmetrical lenses is alot of bad stuff is cancelled by an identical lens (oh, you have to have them centered and collimated decently, not impossible). The beauty of a small f-number is is makes alot of stuff go away also, until it's so small diffraction problems begin to be visible. So a small aperture symmetrical lens gives you alot for you money or scrounging.
The position of a stop or aperture, if centered is theoretically ideal for 1:1 macro. Thinking about whether moving it forward or backward of center for optimizing at infinity gives me mechanical dyslexia. Maybe someone will answer that here. Some people have told me that cell spacing isn't as critical with simple lenses and long f.l. lenses, but a gut feeling tells me that centering/collimation is more critical on longer f.l. No substantiation there, just gut feeling.
I still have no clue about coverage for a given design, other than the following:
From Green, Primitive Photography, probably taken from an old classic British lens I forgot the designer of, is try to use lens elements of diameter 1/7-1/5 times overall f.l. with spacing at similar proportion.
Coverage using the rule of thumb f.l. = diagonal of negative will probably result in illumination but aberration as you move outward from center. Some of the Greene lenses I thought had 'usable' coverage on the order of 20-30 degrees.
If you want sharper images over the full negative, aim for a much longer f.l. (2x 'normal' +/-) and be prepared to baffle some of the excessive coverage/illumination capability inside the camera to alleviate flare.
If you're shooting b/w and don't have achromat cells, try a contrast filter ranging from heavy yellow to orange to red to narrow the spectrum and reduce achromaticity. It lends itself well to most subject matter with some exposure/development effort, better than shooting everything thru a blue filte any way.
Back to symmetry...since choices are poor in surplus catalogs with regard to mimicking a careful commercial design, two of 'whatever' will reward you. Oh, just don't use two negative diopter lenses. To actually see what you are focussing the net diopter value has to be positive.
Now, to dispense with the obsession for perfection, which some say is over-rated, and bowing to evaluative esthetics, see John Siskin's article on homemade lenses. He goes right ahead with a number of triplet experiments - you just don't re-create a Cooke Triplet haphazardly.
I think it was a 2002 View Camera magazine article and it is on the web somewhere, maybe John's webs site. Oops, I just corrected spelling of his name
http://www.siskinphoto.com/magazine1a.html click on image for copy of article.
Better make room for every little piece of glass you start hoarding , disaassembling every malfunctioning or lousy lens you come across, buying old camcorders, etc. You don't find pairs of identical elements this way, but you find some cheap cool stuff.
Another fun thing to do, with no idea as to quality of results, is to take a negative diopter lens and put it behind a positive lens to lengthen the f.l. and coveage. A short lens (ie. 40-50 mm) has a large diopter value so you have alot of options with negative lenses before the net diopter sum goes negative.
Long lenses have small diopter values (1 m = 1 diopter) so you have hardly any off-the-shelf negative options.
I pulled a Tessar out of an Olympus AF 35mm camera, 38 mm I think, and put (-) elements behind it but haven't exposed any film yet, just looked at ground glass (it's cheap and I'm easily amused).
I use the Gullstrand's Equation calculator on the Hyperphysics (Georgia) site to e.s.t.i.m.a.t.e. what happens as you vary spacing.
I just finished my Abominatar lens for a camera obscura, theoretical f.l. 8 m (8000 mm), but I have no idea where to measure it from. I have to mount it because it's really shaky trying go hold it and focus a light bulb across the room. I suspect coverage will be narrow because it's very long for it's diameter (roughly 2m long, 40 + 70 mm glass diameter)