I presently work in a Melos SE75 single ended with 6Kg6 output tubes
Each output tube is driven by a BUZ80 mosfet follower. The first stage is a 12AU7 in cascade with 20K plate resistors. There's even a mosfet follower between the two cascaded 12AU7. (???) The last 12AU7 is driving a BUZ80 mosfet follower which drives two paralleled BUZ80 driving two output tubes. As I think that the mosfet follower between the cascaded 12AU7 is overkill, I bypassed it. The BUZ80 has an input capacitance of 750pf. This is probably why the 12AU7 is driving one BUZ80 who is driving two BUZ80 instead of driving directly two parallelled BUZ80. Do a 12AU7 with 20K plate resistor could perhaps drive 1500 pf ? Thanks.
Each output tube is driven by a BUZ80 mosfet follower. The first stage is a 12AU7 in cascade with 20K plate resistors. There's even a mosfet follower between the two cascaded 12AU7. (???) The last 12AU7 is driving a BUZ80 mosfet follower which drives two paralleled BUZ80 driving two output tubes. As I think that the mosfet follower between the cascaded 12AU7 is overkill, I bypassed it. The BUZ80 has an input capacitance of 750pf. This is probably why the 12AU7 is driving one BUZ80 who is driving two BUZ80 instead of driving directly two parallelled BUZ80. Do a 12AU7 with 20K plate resistor could perhaps drive 1500 pf ? Thanks.
A 12AU7 with a 20K plate load is ~5K driving 1500pF. That'll be down a couple of dB at 20kHz. Given the amount of feedback necessary to make an amp like that work, it must be heavily compensated to maintain stability.
I thought that with followers, the reverse transfer capacitance (Crss) was the one to look at. Which for the BUZ80 is some 10's of pF, not 750 pF.
If the Crss is the one to look, why they use one mosfet follower to drive two mosfet followers that goes to the two output tubes instead of using only two mosfet followers after the 12AU7 to drive the output tubes ?
I don't know, novelty value? I have not heard of anyone else doing this.
Discussion regarding good devices to use here:
Source follower IRF720 or DN2540
The BUZ80 is not a great choice.
Discussion regarding good devices to use here:
Source follower IRF720 or DN2540
The BUZ80 is not a great choice.
I found a thread in your discussion about The IXTY08N100D2. It seems it could be the solution. Why the BUZ80 is not a great choice ? Is it only because it has a Cin of 750 pf ?
I have been using mosfet followers between the driver and output stages in my amp designs for over 15 years. The well regarded Tubelab TSE and TSE-II amps use a CCS load on the driver triode that feeds into a mosfet follower which directly drives the grid of a 300B, 2A3 or 45 tube. This allows the maximum gain at low distortion from the driver triode since it sees a near infinite plate load. The coupling cap is now working into a very high load resistance, so expensive boutique caps are not needed, and blocking distortion is totally eliminated.....far from a novelty value.
See the schematic in post #1 of this thread for details:
After a 14 year run, the TSE must DIE!
It is, and there are plenty of good mosfets to choose from with Crss under 10 pF.
I can't find a schematic for that amp, so I have no idea what the designer was thinking.
Hi Tubelab, I was referring to a mosfet follower driving other mosfet followers which then drive the tubes. Of course your work is well known amongst most readers here 😀
Z = 1/(2πFC) … rearranges to
F = 1/(2πZC) … subbing in
Z = 10,000 Ω
C = 750×10⁻¹² F (750 pF) gives
F = 21,200 Hz.
This would be the –3 dB point, if I remember. If it were R instead of Z, it'd be –6 dB, but the Z impedance along with the imaginary plane solution to the equation yields only –3 dB at 'crossover' point. Classic Butterworth filter. F = 1/(2πZC) … subbing in
Z = 10,000 Ω
C = 750×10⁻¹² F (750 pF) gives
F = 21,200 Hz.
Anyway, the take-away is, "use a low enough load resistor" to the source-follower … even if it sucks up appreciable power juice. Ensures the highs are “crispy”, hopefully without being sibilant.
⋅-⋅-⋅ Just saying, ⋅-⋅-⋅
⋅-=≡ GoatGuy ✓ ≡=-⋅
Look at a spec sheet when considering gate capacitance. The numbers go up somewhat when you get to less then ten volts drain to source. Best thing is to make sure your follower never gets there or at least account for it.
What do you think about this one to replace the BUZ80 ?
IXTP01N100D
https://www.mouser.ca/datasheet/2/240/Littelfuse_Discrete_MOSFETs_N-Channel_Depletion_Mo-1623211.pdf
IXTP01N100D
https://www.mouser.ca/datasheet/2/240/Littelfuse_Discrete_MOSFETs_N-Channel_Depletion_Mo-1623211.pdf
Yes but this rating is not linear at low drain source voltages.
It requires a minimum of drain to source voltage of 100v for linear operation which in turn will need raised drain voltage and increased dissipation.
For example if you are going to use it in a 6SN7 circuit which needs a plate voltage of 250v @ 8ma and assuming you need 50 Vrms output = 50*1,414 = 70,7 peak, adding those 100 volts means that the supply voltage should be no less than 250+70,7+100 =420,7v DC and your mosfet dissipation
will rise by 100v * 8ma = 0.8w for a total of 1,366 W (59% increased dissipation) which means that you will need a heat sink in a mosfet that is designed to be soldered on the pcb.
I suggest FQPF1N60 or FQPF2N60 instead.
Chris
Last edited: