A tubular 'Blameless'

[GK]

Here is preview of an amplifier that I've been collecting parts for, for a while now. The design is finalised, I just have to wait for the remaining bits to arrive and assemble the thing. The semiconductor complement is greater than the tube complement, so I guess this is the correct forum

I've taken the standard 'Blameless' front-end a substituted all the icky-poo NPN's for loverly warm, glowing triodes. I've also added a cathode follower to buffer the VAS from the class-B double emitter follower output stage. This 12AT7 cathode follower is biased with a current source consisting of an identically connected (self-biasing) 12AT7, which effectively cancels most of the distortion of this stage.
I've simulated and optimised this whole design extensively in SPICE (shock! horror!) with accurate triode models and it works rather well.

All tubes in the input stage / VAS are 12AT7, except for the twin triode used for the LTP, which is a high gm 12AX7, as used in my K10A amplifier.

The input stage / VAS has a total of four twin-triode tubes and requires a power supply of +/-200V @ 35mA. These supplies are going to be provided by individual hybrid shunt regulators for each channel.
The four large tubes in the mock chassis layout picture are 807's, which are the shunt regulator elements, each of which cascode an NPN transistor control element. Four MTP2P50E P-ch MOSFETs mounted on a back-panel heatsink will be the constant current sources for the shunt regulators.
Shielded Chassis sockets will be used for the twin triodes.

The heater supplies for the twin triodes are regulated too. The 807 heaters are (sensibly) fed with AC.

Power up sequencing (heater warm-up/ HT timing /speaker relays, etc) and DC fault protection will be controlled by a PIC16F876.

I'll start etching the PCBs tomorrow evening…….

[jam]

The yellow grid on the schematics is driving me nuts!.....................

Jam

[Wavebourn]

I would suggest you to think on such things as:
1. Orientation of fins according to convection flow
2. Pentodes in CCS?
3. P-type triodes?

I used on labs analog computers, they had similar schematics of opamps in them. Without transistors, of course, but with gas discharge VR tubes for voltage shifts.
They had a couple of potentiometers for null adjusting: to balance a diffpair, and to adjust a null in an output stage.

[GK]

The heatsink is big enough for the power to be dissipated, even in the vertical orientation.
The triodes (with heavy cathode degeneration) work quite fine in this application as constant current sources.
Show me a P-type tube!

Another PCB (containing the DC fault detection comparators and balanced input circuitry) has a the DC servo's to steer the offset imbalance. That reminds me that I left out the servo return terminal on the input stage schematic & PCB. Oops.

[destroyer X]

can you, please, remove it and post the schematic once again.

thanks in advance by your kindness,

regards,

Carlos

[Lumba Ogir]

Spare our eyes per favor!

[GK]

Originally Posted by destroyer X

can you, please, remove it and post the schematic once again.

thanks in advance by your kindness,

regards,

Carlos

I'll post the full schematic in high res. B&W after it's built and tested; hopefully before the week is out.

[Wavebourn]

Originally Posted by G.K.

The heatsink is big enough for the power to be dissipated, even in the vertical orientation.

Ok, it's yours, you may waste it for your pleasure.
But I am getting kind of a nervous reaction when see incompetently oriented heatsinks, like a tiger when somebody pets him in an opposite direction of his fur.

The triodes (with heavy cathode degeneration) work quite fine in this application as constant current sources.

Plain resistors will give as well blameless results.

Show me a P-type tube!

AFAIR, Demian (1audio) posted a picture of one.

Another PCB (containing the DC fault detection comparators and balanced input circuitry) has a the DC servo's to steer the offset imbalance. That reminds me that I left out the servo return terminal on the input stage schematic & PCB. Oops.

Oops, no "blameless" tag anymore.

[GK]

Originally Posted by Wavebourn

Ok, it's yours, you may waste it for your pleasure.
But I am getting kind of a nervous reaction when see incompetently oriented heatsinks, like a tiger when somebody pets him in an opposite direction of his fur.

Originally Posted by Wavebourn

Plain resistors will give as well blameless results.

With a -200V powersupply, no, plain resistors "long tail" won't perform as well.

The LTP and VAS triode current sources are biased with a 20V zener diode and are cathode degenerated accordingly.

The 2.2mA LTP current source has a plate impedance of 580k.
The 5.1mA VAS current source has a plate impedance of 230k.

So, with a -200V supply, ~6 times higher impedance than fixed resistors to sink the same current.


The self-biasing current source for the cathode follower has a much lower plate impedance, of course, but this is entirely deliberate, as then gm variation of the CF CCS opposes the gm variation of the CF, cancelling (to a degree) the distortion (the CF has equal value cathode degeneration).

[Bigun]

I like the concept, seeing as there aren't p-type tubes, just use silicon. In other words, if the mountain won't come to Mohammad....

The mock-up looks really nice, so much nicer than a plain box where it's all under the hood.