What defines a good tonearm

For a serious answer I have found that the execution of the particular arm is more important than the design principal chosen. Conventional arms, unipiviot arms, linear tracking arms, etc can all sound great [ or not!]. Every design principal has strengths and weaknesses and great designers can maximize the first and minimize the second. If one design principal were clearly better then we would not still see the variety of arms we do. Unless of course audiophiles were just mindlessly searching for novelty and the latest thing regardless of sound quality; what a ridiculous thought!

Stanwal,
And I guess thats my point. I mean, there are DIY unipivot arms on the web using knitting needles with material costs under $5.00 that are being touted as being more than acceptable, so how much more does it cost to make a great unipivot that is priced at $5,000.00. I know they are beautiful, and the workmanship is superb, and they have adjustability and all that, but if unipivot is the answer, why bother with the difficulty of building a fixed bearing design. Or is there some deficiency in unipivots that the fixed bearing is trying to fix.

I remember using a Pickering table that had a unipiviot arm consisting of a large steel needle which fit into the bottom of a plastic cup on the tonearm. Did it work? Sure. Did it sound like modern arms? Hell no! The question is , what do YOU find acceptable! Just because it is a unipiviot does not mean that it the equal of a GOOD unipiviot. My Graham arm is a unipiviot and cost considerably more then $5, as do the VPIs that I am a dealer for. If you cannot hear the difference between arms then get the cheapest one you can. If you can get the best you can afford. No one can prove to you using logic that a good tonearm is a necessity, only your ears can tell you the answer. I am constantly amazed by what SOME of the DIYers think is acceptable; others are among the most sophisticated of audiophiles. Many of the expensive arms on the market were designed by individuals who were dissatisfied with the commercially available ones, starting with SME and on through Graham, Basis, etc. I am a great believer in saving money but an even greater believer in good sound; just like in every other sphere in life, while money is no guarantee of quality, good costs more than bad when everything is taken into consideration.

I'll vote with Sgunther for the Trans-Fi Terminator and submit the following list as ideal criteria for a tonearm:

1. Linear tracker for perfect tangency. It's disingenuous to suggest that small errors in tangency of pivot arms are of neglible significance. If this were so there would be no market for 12" arms. IIRC, extrapolating from a Bob Graham paper, the average percentage reduction in tracking angle error in going from a 12" arm to a linear arm, is greater than the reduction in error in going from a 9" arm to a 12" arm. You can hear this difference.

2. Low pressure soft air bearing(1-2 psi) eliminates bearing chatter created by air turbulence in high-pressure bearings of Kuzma, Walker, etc.

3. Wand is mechanically coupled by point bearings to a moveable carriage of sufficient mass to absorb excess cartridge body energy, while still being completely isolated from external plinth vibrations on an air cushion.

3. Separately adjustable vertical and horizontal inertial mass accomodates cartridges of widely varying compliance. As close as it can be to a universal tonearm.

4. Short wand for maximum stability and neglible mechanical resonance associated with wand composition. On this arm I've tried short wands using various custom wood, alloy, and carbon fiber constructions-- they all sound very close in character. Arms like Reed suggest that comparing woods is like choosing a golf club. The short arm takes this sport out of the equation-- which some owners may miss. The short wands are cheap and swappable--avoiding the mechanical impedance barrier introduced by removeable headshells or detachable arm wands.

5. Uninterrupted shielded wiring from cartridge pins to phono stage-- minimal signal degredation and virtual elimination of RFI/EMI issues. Anyone who has tried uninterupted wiring will also discover that many hum problems are introduced at the mechanical junctions of wiring. The best solution is to get rid of the junctions.

Most of the above points are outside the common vocabulary of discussion about tonearms. The only quibble concerns the matter of the short 2" wand. Since the Trans-Fi's short flat wand can be loaded for variable inertial mass in both horizontal and vertical planes, the issue is not about the short wand's dynamic behavior in combination with cartridge compliance. The matter boils down to changes in geometry over warps. IMO putting this little point aside, the Trans-Fi is the least burdened of any tonearm by the theoretical compromises that lead most designers to increasingly elegant and expensive workarounds.

Stanwal,
Don't misunderstand me, I think the ingenuity and craftsmanship that goes into something like a triplanar or graham or schroeder or any of the other high end arms is well worth the money spent. Much like sophisticated medical or lab equipment, these low volume works of machine art are expensive to create and duplicate. This isnt stamping a shape out of a piece of tin, but real machineing of different metals pieced together to form a working product.
My question is more why is there two basic lines of thought on how to make a pivoted tonearm. Either unipivot or captured bearing and what is the inherent deficiency in each that the other is trying to fix.
My assumption is that a unipivot is more free to move in all directions, and is therefore more likely to track the groove accurately. I also assume that a captured bearing is less likely to allow a cartridge to chatter or bounce around outside where it is supposed to be. And then we have the Townshend silicone trough to damp all this motion, which does seem to work from my experience, although I havent tried it with a unipivot.
Is it your opinion that either of the two pivoting tonearm types is the right answer, and that the other type, while it can be made to sound great, has an inherent impediment to perfection, if that was achievable.