Aleph 30: adding an emitter follower?

[fulvio]

Hi,
everything begun few days ago, when I measured the slew rate of my new borne Aleph 30, finding it was just 4 or 5 V/us; being worried about transient intermodulation distortion (that is said to be related to low SR), I tried to get it higher.
I suspected that the guilty was the mosfets input capacitance in the amplifier section (I used IRFP350, cheaper but with higher input capacitance than IRFP240), so, just for test, I put an emitter follower between the differential stage and the gates of power mosfets; I used a BJT (BD139), with a trimmer to set its bias current.
At low bias current (less than 10 mA), the slew rate was the same, but when I increased the bias current to 15-20 mA, the SR increased to about 11 V/us; for higher current, no further improvement.
After that, I measured again the amplifier distortion, finding it was 6 to 10 dB lower than before, also at low frequency.
Some values I got:

F= 999 Hz, Vout = 14 Vrms (Pout = 24.5W/8 ohm) THD= -61 dB
(second harmonic: -63 dB, third harmonic: < -70 dB)

F= 6301 Hz, Vout = 14 Vrms (Pout = 24.5W/8 ohm) THD= -54 dB
(second harmonic: -64 dB , third harmonic: -56 dB)

Two tones IM:
F1= 11 KHz, Vout = 7 Vrms; F2= 12 KHz, Vout= 7 Vrms (total peak amplitude 20 Vpk)
1KHz IM product: - 68 dB (below each tone amplitude)
10KHz IM product: - 70 dB (below each tone amplitude)
13KHz IM product: - 65 dB (below each tone amplitude)

I’m happy of this, but now, before starting new improvements, (as using a mosfet source follower, instead that BJT emitter follower), I’m wondering if lowering distortion can cause the loss of Aleph typical sound, making it cleaner but less pleasant.
I can’t do a listening test (at the moment I don’t have any good speaker), so the question is:

- is there anybody (more expert in Aleph’s sound than me) that can tell me how much is correct to reduce distortion without loosing the quality of sound?

Thanks
Fulvio

[Zen Mod]

dunno - I drew that long time ago , but didn't tried

[Zen Mod]

NB

it's not just matter of follower , it's also matter of greater NFB , tnx to increased OLG

to introduce any follower , one needs to increase drain resistor in LTP , to compensate for Vbe ( in case of bip follower ) or Vgs ( in case of mosfet follower)

result is increased gain of input LTP , that's why OLG is increased too

[fulvio]

No, I did'nt increase the drain resistor; the differential pair works now slightly unbalanced, that is 11.5 mA in the first mosfet and 8.5 mA in the second; the 1.5 mA current increase in the first mosfet compensate the 0.6 V of BJT vbe.
An increase of 3 mA of the current generator should be right to re-balance the differential.
Fulvio

[knutn]

I would not use a follower - I think it would alter the tonal balance of the amplifier. If I were you, I would simply increase the tail current of the input differential. Then you have to change (lower) two resistor values - one that is defining the tail current - the other the drain resistance of the input differential. (Maybe it would be necessary to mount a heat sink on the current generator for the input differential). If you double the current, the Slew rate should double.

[fulvio]

Thanks Knutn, I’ll try for sure what you suggest, even if I don’t like to increase signal mosfets dissipation, because they already works in a hot ambient; I built my Aleph as two monoblocks with and a separate power supply; each monoblock is made up with two heatsinks; in the space (30 mm) between their internal sides is located the PCB; so everything inside goes near the heatsink temperature (60 – 70°C); for this reason, the signal mosfets have already a little heatsink that can’t be enlarged.
Fulvio