That is exactly right. Let's see why that's true.
One convenient way to specify a tube's noise is to compare it to the Johnson noise of an ideal resistor, multiplied by the tube's gain. In other words, a tube can be thought of as an ideal (noiseless) tube with a normally moisy resistor at its input. This spec is generally called "equivalent input noise resistance."
One of the very quietest tubes in audio is the triode-connected D3a, just like Alexander used. Run balls-out, its equivalent input noise resistance is about 60 ohms, i.e., it acts as an ideal tube with a 60 ohm (room temperature) resistor at the input. Now, a typical MC cartridge will have an internal resistance of just a few ohms, plus another ohm or so for wiring. Call it 6 ohms in total. There is, of course, an intrinsic Johnson noise to the phono cartridge, which sets a noise floor that a preamp can't help. All the preamp can do is add noise.
OK, so with a source impedance of 6 ohms and an equivalent noise resistance of 60 ohms, it can be seen that the tube (and this is a pretty darn quiet tube running at its limits) is adding 10dB of noise beyond that of an ideal preamp. For a more sane operating condition, the ENR will be more like 100 ohms or so.
An ECC88 is worse. Best case, and this is rarely achieved in practice, the ENR is about 350 ohm. So there's almost 20dB more noise than an ideal preamp. Paralleling gives a 3dB improvement, not super helpful.
What does a transformer do? If you use a 1:10 stepup (and for simplicity, I'll ignore the DCR), the cartridge now appears as a 600 ohm source. Suddenly, the ECC88 is only adding 2dB more noise than the theoretical minimum.
Transformers are a bit tricky to optimize, but they present the best solution for tube designers of RIAA stages.
SY,
Are these noise values valid for both common cathode and common grid connected tubes? I ask this because I also tried a MC head amp with an ECC88 connected in common grid and it sounded magically. The cartridge was directly connected in the cathode, which has some dangers (1 or 2 mA DC flowing through the coil). Using an 1:4 transformer there would be perfect impedance matching and the B+ voltage could be raised.
Laszlo
Are these noise values valid for both common cathode and common grid connected tubes? I ask this because I also tried a MC head amp with an ECC88 connected in common grid and it sounded magically. The cartridge was directly connected in the cathode, which has some dangers (1 or 2 mA DC flowing through the coil). Using an 1:4 transformer there would be perfect impedance matching and the B+ voltage could be raised.
Laszlo
Attachments
a friend of me had make good experiences with a Graaf WTB one - go to second PDF in post #8 under
Graaf Graafiti WTB One - Schematic needed
the 4AH5 from National provide much more better SNR than PC900.
An excellent alternative to both SUTs and separate "head" amps is Allen Wright's hybrid cascode.