If you look at the Modern-Aire site, they have several shapes where the front tapers at the top to the depth of an upper cabinet, and the sides are straight so as to butt up to, and cover, the yellow cabinet sides you mentioned. Their typical sizes are 18" and 30" high, but you could get a 24" custom one if your budget allows. Or you could get an 18" and have a box made for above it to cover that area and to which you could attach replacement moulding. The box could be made of either wood matching your moulding or sheetrock. Vent-a-hood has a stainless direct replacement if you decide to stick with the same size you now have. I'm attaching the link to the Modern-Aire page that shows their shapes. PS-10, PS-11, and PS-26 all have the characteristics I was talking about above. Here is a link that might be useful: Modern-Aire Hoods

I don't think you will notice after a while. Mine is off by about 1" on the left, because the *&^&%$ range hood is a full inch less wide than the rangetop (I swear I bought a 48" range hood, not 47"!). But then again, we are DIY so I wasn't about to make my poor DH uninstall then reinstall all of the lower cabinets after it took him a full day in the first place; if we were paying someone for it, it might be a different story. If you choose to keep it, just make sure you don't do a stacked tile BS - you would probably definitely notice it then. Example of what not to do: http://www.peeltile.com/img/media/backsplash/layout.jpg

In principle, any hood with an internal blower could be gutted of the blower and operated with an in-line or roof mounted blower. (GreenDesigns may have meant that using it while leaving the blower inside would be a poor plan.) I can imagine construction details that might cause non-optimal internal hood flow patterns, but I have no detailed knowledge of the internal construction of myriad hoods to know whether such are prevalent. There may be nanny jurisdictions that would claim that code approval of the hood was violated when it had its blower removed. Any hood available for order without a blower likely is fully compatible with an external blower. This, of course, can be confirmed with the manufacturer. The most important adaptation requirement is that the blower be continuously controllable, and that the hood control be a continuous motor control. A hood with just a few motor speed positions on a switch might require a compatibly-wired motor. While a multi-position hood control could in principle be replaced by a continuous control, par. 2 may apply. When I was doing my kitchen reno, I chose a compatible hood/roof blower combination (supplied by Wolf) because the entire scope of the project (gutting and reconstructing part of a house) made playing with the details of the blower circuit a side project of tertiary importance that I couldn't afford the time to deal with. As it happens, inside my Wolf (Independent) hood is a diac/triac motor control circuit, about the simplest design available that goes back to the '70s. The compatible Wolf (Broan) roof blower uses a typical induction motor that is easily controlled by this type of circuit. The cost versus time value trade-off needs to be considered when dealing with issues like this. kas

Assuming a cooktop with a cooking area (contour around the union of hot pan base areas) of some dimensions L x D, a hood entrance aperture of dimension L + 6 x D + 6 (inches) would (in still air) provide sufficient capture up to about 36 inches of height over the cooktop. If there are strong cross drafts, then a larger aperture is recommended. Blower airflow (actual) capability should be of the order of 90 CFM/sq. ft. of hood aperture. If roof penetration and wall penetration are both out of the question, then adapting the methods of apartment type recirculating hoods may work, if not prohibited by the safety statement associated with the range. Such hoods need adequate layers of grease capture filters, and charcoal filters. Vent-a-hood and Broan/Best/NuTone may have units suitable for modest air flow requirements; for higher power cooking requirements, their approach would have to be adapted into a larger, more powerful unit. Perhaps custom hood fabricators such as ModernAire could rise to the occasion. A simple web search for a commercial counterpart led to this as one example of perhaps others: captiveaire.com This particular unit is too low in CFM for a cooktop. And because it is made of 430 stainless steel, in salt air it may need a dilute phosphoric acid wash now and then to brighten it up. See also http://ul.com/wp-content/uploads/2014/04/ul_RecirculatingHoodSystems.pdf and http://restaurantexhaustfans.com/how-effective-are-ductless-commercial-hoods/ And here is a snarky, albeit imperfect tract, part of which addresses this subject. The comments about cooking effluent constituents is pertinent: http://www.treehugger.com/kitchen-design/most-screwed-badly-designed-inappropriately-used-appliance-your-home-kitchen-exhaust.html (Please note that a hood cannot "suck" more than it "blows.") Anyway, the important point that I didn't elaborate upon in my first message is that when the hood exhaust air is directed to go "sailing out" along the ceiling, its velocity rapidly drops as the moving exhaust effluent's momentum interacts with the friction induced by the room air inertia. To succeed in the original plan without ducting would require a very powerful and loud "screaming-mimi" fan to accelerate the effluent such that much of it made it to whatever door was open. There would still be significant undesirable mixing with the room air. Some improvement might be achieved by also accelerating a cylinder of the same velocity clean air surrounding the effluent. I believe that achieving even partial success this way is much less practical than any duct scheme, including one that has to penetrate concrete. kas