V-Twin wrote:
Patrick Turner wrote in message ...
Have you measured the CD player output?
I bet it ain't 2vrms, even when measuring with a stream of
heavy metal grunge as a test tone, on an rms meter with 20 kHz of bandwidth.
Specs says max output 2Vrms, i don't have yet a test CD to do mesures.
No doubt that the "normal" output level is lower, since plugging the
CDP directly to the amp don't saturate it (1V sensitivity).
I think the 2V are the max voltage the output stage of the CDP can
reach.
The RIAA losses are negligible at LF, -20 dB at 1 kHz, and about -38 dB at 20
kHz.
If you had rated 4mV at 1 kHz from the cartridge, and a gain of 5,000
in the amp, and -20dB loss in the RIAA filter,
you get Vo = 0.004 x 5000/10 = 2v.
Uhm...but i want that the 2V are reached only during the peaks, like
the CDP, therefore i must consider a 20dB (Morgan Jones measured +16dB
playing the Beethoven's Ninth Simphony) larger signal coming from the
cartridge. That makes 40mV at 1kHz, or 4mV at 20Hz due the riaa curve.
The first stage of the phono preamp gains 35dB, the riaa network
equalizes the signal cutting 20dB@1kHz and further 20dB@20kHz, thus
after the riaa network i have 240mV at all frequencies. CCS loaded,
the 76 gains about 15, thus 240mV*15=3.6V, wich is more than enough.
Considering instead the nominal output of 4mV, I ends up with 24mV
after the riaa network and 360 mV at 76's output, wich should be not
too far form output voltage of CDP.
Where I'm wrong?
I have never seen 40 mV coming from a cartridge at 1 kHz.
I have never studied the CRO trace during the 9th symphonie's cannon shots.
4mV is the high side of what comes from a cartridge.
The signal going to a record cutting machine amp
has its signal equalised so relative to 1kHz, the bass is reduced a maximum of 20 dB,
and the treble is boosted at 6 dB/octave after 2,110 Hz, until about
50 kHz, when the boost is curtailed.
Therefore, if a source of constant level sine waves are fed in
to the cutter amp, the bass amplitudes are much reduced, and treble increased,
and there are limits as to how much boost to treble can be applied,
lest the cutter run out of headroom.
The noise of a cannon shot contains a potpourrie of F above 1 kHz.
The sound you hear from the vinyl is nothing like a real cannon 50 feet away,
since the system is utterly incapable of containing the dynamic range,
and a lot of limiting is going on, or they moved the cannon to 500 feet away.
A nice big bash of cymbals can easily over load the recording system.
The engineers allow for all this, and at the end of the day,
when we make a tube phono stage, the tubes' dynamic headroom ability
even with high output MM is quite able to take whatever is thrown at it
from the cartridge.
In fact, if you have 40 mV input at 1 kHz, which is a 20 dB overload,
with the amp set up with a gain at 1 kHz including RIAA of say 400, then you'd get 16 v output,
which is still below the tube limit of perhaps 64 v
but way above what most SS might achieve, using +/- 12v rails and opamps.
I am not sure what the maximum signal at 10 kHz would be compared to
what might be maximum possible at 1 kHz.
But if you feed in a range of F at the same amplitude to the record cutter to get
4mV at 1 kHz, then expect around 40 mV at 21 kHz from a record cut this way.
Fortunately, HF content in music reduces a lot above 3 kHz,
so the HF is accomodated, despite being boosted, to improve the snr.
In the preamp, the 21 khz cartridge signal of 40 mv is treated differently
depending on whether the amp has its RIAA eq done either passively,
or by way of NFB.
The passive stage needs to be able to take 40 mV without saturating, or
running into cut off or grid current, so if the 12AX7 gain is 90,
like it is in my amps, then 3.6 v output shouldn't cause any problems.
In fact, 400mV input at 21 kHz won't either, because the first stage of a decent
passively eq'd tube amp is capable of well over 36v output.
Its almost impossible to ever overload a tube phono amp.
The possibly high 21 kHz output from stage 1 is reduced 40 dB by the passive
eq so the next stage won't be overloaded either.
The feedback type of eq has the first stage recieving two signals,
one is the possible 40 mV at 21 kHz, and the other is the fed back signal,
and the difference between the two is all that is amplified, and its a heck of a lot
less than what the passive stage has to endure.
Thus RIAA FB stages have good overload capability, and far less thd,
because the network feeding back the signals from the output also
feedback any thd, which then gets reduced by nfb action, like most other amps,
except that LF errors are corrected less than HF errors, since more FB occurs as F rises.
However the feedback network itself becomes a low impedance load
for the tubes at HF, and this itself can cause overloading
of the final stage in the FB type phono amp, if output voltage is high enough,
and the design of the FB network has R values which are too low.
The V2 of the amp, if it is a ****ant tube like the 12AX7
should perhaps be coupled to a medium U triode cathode follower,
itself a case of lotsa local FB, and it then can take the brunt of the load
imposed by the FB network itself at HF.
The use of FB RIAA is almost universal in the world of SS amps,
which need all the FB thay can muster to compete with triodes.
Hence opamps are so often used.
With tubes, FB is optional because tubes have emormous
dynamic range, and triodes can be extremely linear at low signal levels,
even without the slightest use of loops of FB.
All my SE phono and line stages produce less than 0.1% thd at 10vrms output.
All have passive RIAA eq.
With a balanced set up, like what you may find at vacuumstate.com,
the thd is even a lot lower.
Anyway, to cut a long story short,
Perhaps you should just try building a two stage
triode amp, so that the gains of the two tubes equal say 4,000, +/- 50%,
with a passive RIAA filter between V1 and V2,
and have a measure, and have a listen.
The output you may get from such an arrangement may be above what you might get from
many CD players.
Once you have done this, you will know more what to do next.
Patrick Turner.
|