British vs. American RIAA Stage Design
John Byrns
Hi All,
The typical vintage American RIAA magnetic phono preamp uses a cascade
pair of ECC83 triodes, with a feedback equalizer from the output of the
second stage to the cathode of the first. This basic design was used with
minor variations by most of the top American Hi-Fi companies, like Dynaco,
Fisher, H.H. Scott, Marantz, and McIntosh. Marantz and McIntosh followed
the two gain stages with a cathode follower, although only Marantz drove
the RIAA feedback equalizer network from the output of the cathode
follower. Dynaco, and Scott used positive feedback between the cathodes
of the two stages to increase the gain. McIntosh also used this positive
feedback trick, but from the cathode follower to the cathode of the second
gain stage.
The typical vintage British RIAA magnetic phono preamp uses a EF86 pentode
with an op amp style feedback equalizer network, a series input resistor
somewhere between 68k and 120k is used between the input and the grid of
the pentode. Mullard, Leak, and Quad all used variations on this theme.
The Mullard and Leak circuits seem to have done all the equalization with
the feedback loop, while the Quad design appears to use a shunt equalizer
in the plate circuit, for the low frequency boost, and feedback for the
high frequency rolloff. The time constants in the Quad circuit look a
little strange to the eye, but I assume they work out on paper. As always
Quad takes a slightly different tack.
I am curious if anyone has any historical insight on how these two
distinctly different styles of phono preamp evolved, and why each design
seems to predominate in the respective country. I realize of course that
some American designs used pentodes, and I would guess that a few British
designs used the cascade triode design, but each country seems to have
largely preferred the one design durring the golden years of tube stereo.
I am also wondering if there is much of a noise penalty to be paid for the
series resistor in the grid circuit of the British design?
Another question that comes to mind relates to the possibility of using a
pentode as a current source to drive a shunt equalizer network in the
plate circuit, for both the high and low frequencies. Would that work, or
is there likely to be a overload problem, perhaps at high frequencies?
Any thoughts on how these two distinct RIAA stage designs evolved would be
appreciated.
Regards,
John Byrns
Woodside Electronics
John Byrns <jby...@enteract.com> writes
I have seen both types of feedback used on both side of the pond. I
suspect that the trends tend to come from tube manufacturers data sheets
with suggested designs and also published articles in the individual
countries by well known writers such as Williamson and Hafler.
>
>I am also wondering if there is much of a noise penalty to be paid for the
>series resistor in the grid circuit of the British design?
There are disadvantages with both the series (UK style) and the shunt
(US style) types.
The shunt type should have better noise performance but the loading of
the feedback resistor in shunt with the anode load of the output valve
can reduce the gain and increase the distortion of the tube prior to
feedback. The main dissadvantage is that it is difficult to maintain the
RIAA at HF because the gain of the stage can only fall to unity where
the HF rolloff will shelve.
With the series feedback design the loading effects of the feedback
resistor are not so bad but the higher these values are made the more
noise. The main advantage of this design is that the HF roll can be
maintained as the stage becomes an intergrator and never shelves, so the
RIAA HR frequency and phase response are maintained.
Both of these designs will benefit from a lower output impedance stage
e.g. a DC coupled cathode follower or an SRPP design.
>
>Another question that comes to mind relates to the possibility of using a
>pentode as a current source to drive a shunt equalizer network in the
>plate circuit, for both the high and low frequencies. Would that work, or
>is there likely to be a overload problem, perhaps at high frequencies?
>
>Any thoughts on how these two distinct RIAA stage designs evolved would be
>appreciated.
>
>
>Regards,
>
>John Byrns
John Widgery
--
Woodside Electronics