This is a great comparision between products designed by two of the best audio engineers out there i.e. John Curl with the Parasound and Nelson Pass with the Adcom. As has been stated by both John and Nelson, they simply provided the design for these products. How closely each manufacturer followed those designs, how well they laid the circuitry out on the board, what grade of parts were used, etc... is all beyond the designer's control
Getting to the meat of the situation, in my opinion, the 5802 is sonically superior to the 5800 by a noticeable margin. Having said that, I don't really consider either amp to offer "good" sonics, regardless of what the reviews say. Personally, I find this very strange as i am typically a big fan of Nelson Pass' power amps and the basic design of this amp looks to offer great bang for the buck. Then again, see the above comments pertaining to how much actual "control" Nelson had over these products.
In my experience, both of these Adcom's tend to fall apart at a progressive rate. That is, presentation sounds less "unified" or "cohesive" as drive levels are raised. The treble and upper mid gets harder with more grain and midrange loses the little bit of liquidity and warmth that it might have at lower volumes. Bass is bloated and lacks definition at any given volume, but gets worse as spl's are increased.
The lack of bass control and definition really becomes apparent as one uses lower impedance / multiple woofer speakers. That's because such a design will pull at least twice as much current as a lighter, more manageable single woofer / higher impedance speaker would. Given that it takes surface area i.e. multiple large pistons to do bass with authority, and people buying this much power typically aren't using mini-monitor type speakers, the end result isn't quite what i was hoping for.
As a side note, i've cleaned and repaired a couple of these amps. Both models have forced fan cooling and tend to collect quite a bit of dust internally. The power switches on the 5800 tend to fail from what i've seen. While these aren't hard to replace and Adcom is great to deal with in terms of customer service / parts availability, it still shouldn't happen.
Based on past experiences with multiple different units, i think that Adcom did better with their much older Bipolar transistor designs than they do with Mosfet's. This isn't to say that all Bipolar's are better than Mosfet's, but in these specific designs and they way that they are implimented, that's my opinion.
Moving over to the Parasound, I haven't heard an HCA-3500 although i was interested in purchasing one several years ago. Having said that, i would hope that it sounded better than either of the Adcom's and "think" that it probably does. I do know that the 3500 is a TRUE dual mono design, utilizing dedicated power cords, transformers, power supply filtration, etc... for each given channel. This amp is basically an earlier "baby" version of the JC-1 monoblocks, using a similar design but built with "budget" parts ( compared to the JC's ) and crammed into one chassis. I do remember Bob Crump making a few simple yet highly beneficial modification suggestions about these amps over at AA, so if you have one or are thinking about buying one, you might want to do a search over there. Something about changing a couple of resistors, etc...
As a point of reference, John Curl, who designed the dual mono 3500, is the "Father" of all dual-mono designs. He was responsible for designing the first TRUE "dual-mono" stereo amps ever marketed to the public. Since we're on this subject, let me climb behind the podium for today's lesson : )
While "dual-mono" designs do have their benefits in terms of reduced cross-talk / improved channel separation, potential for improved imaging, greater current delivery, better control over the speakers, increased stability under heavy loads, etc... a lot of how much of a benefit ( if any ) that will be will have to deal with the total capacity of the power supplies involved. A single massive transformer with a huge array of filter caps shared between two channels could actually be superior to a dual mono design using scrawny transformers and minimal power supply reserve for each dedicated channel. Such is NOT the case with the 3500, but don't be fooled into thinking that "dual mono" or "individual power supplies for each channel" is necessarily "better" than a common supply. As i mentioned, all things being equal, "dual mono" or the equivalent for a multi-channel design, is "typically" desirable with all things being equal. Then again, how often are all things really equal??? : ) Sean