Sorry to hear that. I can totally sympathize with the tedium of having to go back and re-create what you've already done. Very happy to provide a diversion and I actually appreciate the intensity. I knew there would be some degree of heat and it would not be a proper upgrade thread without casting critical eye to aspects of Hagerman's circuit and my suggestions. There are more things to think about... Though I think the posted simulations for incorporating the Neumann time constant are not complete and indicate a more concerning trend than what's seen in practice: Jim's measured RIAA adherence values for the Bugle 3 are +/-0.5dB (30Hz to 30kHz). Maybe not record breaking, but still quite good. I've certainly heard preamps that report tighter tolerances and sound worse. Put in perspective, consider that nominal capacitances of even good quality capacitors vary by +/-5%, which results in an at worse deviation that is basically the same as the Bugle's RIAA tolerance. As his reported RIAA adherence values extend into the supersonic, I don't think there is much cause for alarm. Though the discussion and the articles about it are very interesting.
Thank you for the kinds words! Re-reading that first post, the one thing I would clarify is that for the C4 and C6 (220nF) positions, I consider the Vishay 1822s to be more the (usually) plentiful and economic choice. They are mylar caps and have some lossiness and intrinsic limitations (as you obviously well know.) Not that they don't do a good job and appear well constructed. The design notes for the Caledonian Audio DIY phono preamp build recommend the 1822, noting they "have been our favourite capacitors for audio signal." Personally, I noticed the sonic performance exceeds other mylar caps I've tried to the point of making them preferable to some caps with superior dielectrics. Of course, capacitor performance is not entirely defined by the dielectric. As John Curl mentioned in a post elsewhere in cyberspace, "Some Mylar caps sound pretty good. Other factors are: lead material, termination to the cap body, and mechanical self-resonance. Some Teflon caps are lousy sounding."
Cheers!!
The Bugle 2 simulated RIAA compliance from 20Hz, not 10Hz, to 20kHz is c. +/-0.4dB, excluding component variations.
The WP design nominally is +/- 0.009dB. The cap values are +/-1% and the R values are 0.1%. Some critical caps and resistors are effectively paralleled, reducing the standard deviation, but not the peak error. The RIAA time constant stages were designed with specific 1% tolerance PP and C0G caps in mind, and the gains/impedances were adjusted accordingly.
I have measured several tens of units and the RIAA accuracy has never been outside +/- 0.040dB 20Hz to 20kHz. I don't test RIAA into the supersonic range but there is a controlled multi pole roll off present that starts just outside the audio band. This variation is well within the inaudible range as far as any studies support or testing on my friends, and myself, can hear, and the "official" spec is within +/- 0.05dB. SOTA has also measured a number of units, and they concur, even though their measurement precision is not quite as good as mine.
Mylar (polyester film) caps have an unfortunate level of batch to batch and manufacturer to manufacturer variation in their linearity. That was unacceptable. Polypropylene is better in that regard, as is C0G, particularly from manufacturers such as Murata.
Almost all leaded caps have steel leads now rather than copper or a copper alloy. I have not noticed any difference in performance, either audible or measurable, down to the sub-ppm level, which is what the design will do. It's hard to come up with a theoretical reason why it should matter except when the capacitors are physically large- something that is avoided, especially in the SOTA designs which are entirely surface mount.
I have bench tested the used caps for distortion and dielectric absorption.
