I've designed & built several thermal air panels, both active (with blower) and passive (convective.) Key to making the convective versions work well is the quality of the backflow prevention flaps- they need to be light enough to open at very low pressure, but tight enough to prevent backcycling at night. Sizing the openings sufficiently large is also important for convective designs. If they're too restrictive you get a high delta-T, but low flow.
Also note, high temp output air == high loss when the outdoor temps are low. On active designs I try to keep the output temps under 110F (but over 80F.) Convective versions are prone to this loss. If after building the thing you find it's putting out 130F+ air when it's under 30F outdoors (a 100F delta-T between collector & outdoors) you may want to consider putting a small blower on it. You can control the with a simple snap-disc type thermostat switch mounted directly on your heat exchanger plate/downpout. This is a nice cheap version with the right temp range (turns on at 110F, turns off a 90F):
http://www.grainger.com/Grainger/WHITERODGERS-Snap-Disc-Fan-Control-Switch-2E245?Pid=search
It's enough of a switch to run a substantial blower, but they don't need much. The size & type of blower depends on the size of the collector and the cross sectional area of the air path through the collector.
See also:
http://www.builditsolar.com/References/Measurements/CollectorPerformance.htm
Wall mounting is absolutely the right approach for an Indiana latitude- better convective forces, and better heat rejection during the summer. Roof mounting at an optimal angle for space heating would require a VERY steeply pitched roof (50 degrees or higher). With a wall mount the performance is undercut only slightly, but allows the use of overangs to limit summertime gains, and the mid-day summertime sun angles cause much of the direct sun to be reflected by the oblique angle to the front face of the glazing, a phenomenon of optics called "total exterior reflectance". The angle at which that happens is a function of the difference in refractive index of the air vs. that of the glazing material.
In a New England climate the annual offsetting of the heating bill on the collectors I've measured in-situ with a good heating-fuel history have been everywhere from 1 therm to 3.5 therms per square foot of glazed area (which is equivalent to ~0.75-2 gallons of heating oil.) That offset performance is a function of both the efficiency of the heating system and that of the air collector. The more efficient heating system, the less difference it'll make, but with an unobstructed southern exposure with an ~80-85% AFUE system in decent shape, 4x8 (32') decently built thermosiphoning collector will offset better than 50 therms, but probably not 100. With a blower resulting in a lower operating temp will usually improve performance by 20-50%, for a small cost in power use.
While I generally advocate active systems if designed well, there are fewer code & maintenance issues with a thermosiphon. The most recent version I helped design & implement was a ~80square foot building-integrated thermosiphon built into the wall of a rec-hall type building. I don't have good heating fuel performance data on this intermittent-use building to compare but I've measured 130F output air when it was 15F outside, 68F indoors, which is not a mark of good efficiency. I'm sure even at a buck a therm gas (current rates) and a 80% efficiency gas furnace they use to heat the place it'll still pay off the ~$800 in materials expended (teenage sweat-equity went into the assembly) in less than a decade.