Ginger,
Thanks for your input, let me offer a few comments.
In the early design phase back a year ago I considered full feedback equalization, knowing it would have superior overload capacity.
However, I selected an IC with outstanding overload capacity, one which will present 28Vpp at it's output into 600R, translating to around 70mV millivolts of input at 1KHz. Since most MM cartridges will only ever peak around 35mV absolute max, I felt this was more than enough given the overload of the input stage to the GK-1 occurs around 15Vpp.
With the complete RIAA in the feedback loop proposed by Ginger there are now three breakpoints in this network, 75uS, 318uS and 3180uS - a move up from just two with the insertion of the 75uS breakpoint into the mix.
Present tube lore on RIAA calls for passive RIAA, and often you see it separated into discrete circuit blocks for each breakpoint, usually two. Frequently it is implemented with inductors, eschewing caps because somehow it is felt inductors are better. (I think this is extreme because inductors are wonderful aerials for hum intrusion and are VERY expensive.) The argument goes that while overload margin is not as good with passive networks, the complex phase inter-relationships found within a large, complex network incorporated into the feedback loop can interfere with imaging.
I have no proof of this, but went ahead with the present design, segmenting the equalization regimes, with the 75uS breakpoint handled passively at the output. I chose an IC with rock solid performance into a capacitive load, and vanishingly low output impedance, and when I finished the design sent it off to Mal Fear for testing.
Those of you watching the design process shrewdly will realize that many choices are made intuitively, often without full mathematical knowledge of all options. This is a fancy way of saying that design is often undertaken by gifted amateurs who have a 'feel' for the problem, but don't necessarily have all the information at their fingertips.
Of course, this is true; but the crunch comes during performance testing, and in any event it has to be said that some designs are inspired, and some are just hack reruns of well understood technology. The difference, I'd suggest, is the work which goes into the performance testing and subsequent refinement, and of course, luck plays a part as well!
Mal soon came back (he works FAST!) with reports of outstanding imaging, which I sheepishly admit had been my primary goal. There are many, many circuits which deliver everything except good imaging, and some are priced at the high end. I like robust bass, mindblowing mids, and heavenly highs, but I've found that if you can get the imaging right, these other qualities seem to fall into place. This was also the experience with the GK-1 Darl and I developed. Once imaging was right, the rest was fine, although we had to do a bit more work to optimize the bass.
Since R&D is exciting but costly, and time to market is important for any product, when Mal's response was favorable I punted one off to Mark Whitaker (Oz_Audio, wherefore art thou??) who verified Mal's experiences almost immediately. I heaved a sigh of relief and placed the phono on the market. Renzo Carlucci in Rome came back with exactly the same report, further reinforcing the good vibes, but when alerted recently by Bart Shepherd of an issue with the top end, I began to grow uneasy.
Ginger, with his incisive math mind and commendable engineering vigor has unearthed the problem; the 75uS breakpoint implemented passively at the output of the phono stage. By changing R6 and R6' (both 39K) on the present phono to 5K6 (half watt metal film in both cases) this problem is solved. I cannot condemn his suggestion to go full active RIAA, but since the product works like a charm, and is now even better, I can't see a good production reason to change it. But tweaking is another matter....
If people implement this mod, please let me know the results. I'm sure interested! But I'll keep it unofficial until Ginger builds it, auditions it, compares it, and a few get around the traps.
Thanks Ginger. Appreciated.
Cheers,
Hugh
