Super preamp.

[GK]

G’day.

Attached below is a preliminary simulation of a fully differential preamplifier I’ve designed to give 100mVrms sensitivity to the 512+512W rms class A power amplifier I’m building.
The input stage is comprised of two cascaded, cascoded differential amplifiers using precision matched U404 JFETs running low drain load resistance (2k) to keep the bandwidth high and relatively high tail currents (5mA), to keep gm high so as to get a decent amount of voltage gain out of each stage.
An emitter follower buffered and cascoded differential VAS follows. Although not simulated yet, the common mode output voltage of the VAS outputs will be steered to 0V with a servo (as is necessary with the differential op-amp topology). For simulation purposes, I just tweaked the value of the VAS current sink bias battery to get the common mode voltage to 0V. The differential VAS is heavily emitter-degenerated for the best linearity and a pair of “diamond buffers” form the complementary output stages.

Simulation results so far look promising:
Single ended open loop gain is 72.9dB with the –3dB roll off point at 1.2MHz.
Open loop THD driving 8Vp-p (approx 2mVp-p single ended input) into 1000 ohms at 20kHz is just 0.034%
This thing is going to run with a closed loop gain of about 20.

I think I’ve pretty much optimised the design the best I can for the lowest THD. Before finalising the design and starting to layout the PCB I just though I’d just post the design here to see if anyone out there spot any obvious improvements.

Cheers,
Glen

[lineup]

Interesting G.Klein

I know you are good with transistors (from suzyj guitar preamp topic)

Could you just attach a bit bigger image.
Then I am sure several will be able to tell,
what they think is good or less good.

Regards
lineup

[GK]

OK, here it is. With differential feedback applied and the common mode voltage servo added.
One thing I should have mentioned in the first post is that I’m using a differential preamplifier because my power amplifier is bridged design, which has differential inputs. Also note that this is a single ended input to differential output design. Due to the differential feedback, it necessarily has low input impedance (2k).
This preamp is to be preceded with a discrete fet input voltage follower configured op-amp to give high input impedance.



Cheers,
Glen

[kubeek]

My opinion is that such design is way too complex for me to even be able to look at the schematic and grasp what it should do. You are using 25 transistor, and still don´t have the current sources "unpacked", and there are many of them...

I wonder who would design a PCB for this

[Upupa Epops]

Twony five transistors cost cca up to one Euro and it isn´t " big money ", Kubeek....

[bikehorn]

512 watts class A?

are you going to park it inside a freezer to manage the heat?

[bogdan_borko]

Man, if you`re not joking then you must be crazy

Do you have the extra room for that amp?
The idea with the freezer is also very good!

Why all that electronics for a preamp??

SENSIBLE MINIMALISM IS THE WAY TO GO!

[Wavebourn]

To screw a bulb into a socket old engineers take a stool, stand on it, and screw the bulb into the socket.

Young engineer needs additionally a team of 8 younger engineers to do the job: 4 to rotate the stool, 4 to go around to prevent dizziness.

Originally posted by kubeek


I wonder who would design a PCB for this

Computer.

[kubeek]

Originally posted by Upupa Epops
Twony five transistors cost cca up to one Euro and it isn´t " big money ", Kubeek....

I am not talking about money.
In DIY, someone has to make the boards, solder the components and with 25 transistors a chance of some error is much higher than with less devices. Don´t forget that with increasing nuber of active devices, there are also more passives, traces and joints.
I understand that for production of huge series made by machines it is not a problem, but I do not want to complicate my life by using a preamp which in this case will have a very high chance of failure because of its complexity.

Originally posted by Wavebourn

Computer.

Show me a program which will take this schematic and produce a more than good board, while taking in account all rules and tinkering you would do when you design board for a superb amp or preamp.

There has to be someone who sits few hours or days by the computer and plays with the design.

[kubeek]

And a freezer won´t help, the heat extracted from inside is produced on the rear side of the freezer through the cooling fins.

G.Kleinschmidt, you live in Australia, I guess it is not cold there like on the south pole?

1024W in class A, that´s something like 3-4kW heat, isn´t?
You would have to have a) heatsink on the wall of the house sized like 1x1m
b) Really big and efficient air-conditioner.

Both possibilites are expensive and are not worth the trouble.