e.g. this item here: http://cgi.ebay.co.uk/6-1-2-FOAM-SPEAKE ... m153.l1262
Considerably cheaper ($2.95 per pair) than the high markup $25 'kits' you see everywhere else.
To remove the old foam, I used a product called Evo Stik Adhesive Cleaner:
£4 down the local market. It worked really well on the bass drivers, but strangely on the mids it was at best 'sort of helpful' i.e. it worked but didn't seem to be quite the right solvent for the job - lots of scraping with a screwdriver
Where the old foam was stuck to the cone I found you could basically just gently tear it away, leaving the cone slightly thinner for about 2mm around its rim. Some thread on here mentions using water to soften the cardboard before doing this - not sure it would have helped really, the amount of cardboard that came away with the glue was pretty minimal.To glue them I used UHU 'all purpose adhesive'
as it's a flexible rubbery sort of glue that on the tube it says is suitable for cardboard, metal etc - seemed to work pretty well and didn't react with the foam in any way. In fact, one of the cones that I'd done wrong I was able to peel it away and re-glue it to the metal rim without having to needlessly further damage the paper cone - so quite handy that. Another advantage is you can get it just about anywhere e.g. the local pound shop or whatever.
Top tip - seriously watch out for the fibreglass they use for paddding/sound absorbtion - even the tiny wisp of it that will most likely come out stuck to the metal of the magnet housing, will spread itself around and get everywhere. By the time I'd finished, I had all these tiny invisible glass splinters in my arms - you probably want to avoid making the same mistake
(just try to get rid of it when you take the cone out - maybe with some sellotape or brush it off with a damp bit of tissue or something)The battery method involved as Jörgen says, gluing the surround to the cone first, then connecting a battery (I used a 9v - bit stronger than a 1.5, as well as easier to hook up some wires with a dodgy rubber band sort of solution
) to lift the cone, which makes it easier to apply glue to the metal rim; then disconnect and reconnect the battery several times to sort of 'wiggle' the cone in to place in contact with the glue. I generally left it connected or kept connecting/disconnecting for a while until it'd started to dry, which is a matter of maybe 20-30s or so with this glue. I think the mistake I made was in checking if it was centred a bit too much, by pushing on the dust cap first in the centre, then slightly closer to the outside to feel for scraping. I generally didn't find any, but I think maybe this pushing round the outside may have moved it a bit off centre just as the glue took hold. I suspect Jörgen's way of using a variable power supply to allow you to gently lower the cone into contact with the glue would have worked a lot better than my fully-on -> suddenly no stabilising magnetic field way 
Incidentally, if you do feel any scraping more on one side than another, you're supposed to push the cone away (towards the opposite side of the dust cap) from where it scrapes.As I say I think more than one of mine aren't done as well as they could be, but for now I can't really be bothered messing with it any more as I might just make it worse so I'll come back to it later - for now they're good enough that I can just enjoy listening to them for a bit. If I were doing it all over again/if I'd known about that method, I'd still do it myself, but would definitely go for the variable power supply method - there's a thread on here that alludes to a special CP guide they were preparing a few years ago to avoid people making the slight mess I have, but I haven't been able to find out anything more about it from the CP people so I decided to just do the best I could/try to figure it out etc - ho hum
