That's a big question....I started messing with the circuit about 2009. This the third and most complex version yet.
Objectives:
Power supply
- Low output ripple.
- +|- 18VDC
- separate 48v rail with its own transformer.
Input Stage
- High CMRR
- Phantom
- Polarity reverse
- Pad circuit
Amplifier stage
- High slew rate
- 70dB of gain
- No coupling caps...
- Low pass and high pass filters...
- Stepped gain control
- Variable trim gain adjust
- Relay switched harmonics emulation circuit with bias adjustment and mix controls.
Output Stage
- Stepped output attenuators
- Variable output trim
- Optional solid state or transformer outputs
Extras
- LED Vu meter the output
With these goals in mind I worked through circuit options.
The power supply is a product of many years of messing around with different designs. This one is very simple in the end but deploys toroidal power transformers, several levels of surge protection, emi filters with common mode chokes, 4th order LC Pi input filters, and finally linear regulators with additional output filtering. Each amplifier has significant decoupling and storage caps. It's total over kill, but the ripple is so low I had borrow my friends $50,000 scope to measure it.
I wanted a high CMRR, and since this for me and a production model, I went for the bling of a Jensen input transformer. I still left pads on the PCB to populate if In case I ever wanted to go solid Stage on the front end. But once I heard the Jensen .....I never went back....
The phantom power switch caused some internal debate, and lot of messing about with concept circuits. I did not want put a dc potential across the input winding of the transformer. So played around for sometime with automatic impedance balancing on the phantom injection, but in the end opted for a manually matching two 6.80k resistors. I played around with a ramp up circuit on the phantom power but did not find much benefit...more can be learned there I am sure.
The polarity reverse was simple to wire but caused havoc with my amplifier. In the end I add relay based circuit that would momentarily ground the input while the polarity was physically switched, the. Reconnect the input. It helped the high gain amplifier from oscillating.
The pad circuit I used matched resistors, simple.
For the amplifier I messed with a few discreet opamp designs using JFET inputs, but could not get a decent slew rate...I wanted ....at least 20uV/s. I ended going a high end opamp....opa627 as the main gain Stage. I used SP12T switch and 12 turned pairs metal film resistors for the course gain and add a pot in series for the trim gain. The opa627 is a bit fussy to keep stable, PCB layout was critical. I got a gain just shy of 68dB in the end, with a BW of slightly more then 200KHz.
I added a wide range HP active filter, standard Butterworth Salen Key circuit using polypropylene capacitors. I did the same style circuit but only for LP. I wanted the LP to experiment with telephone effects or subkick on the way in....it works well but I am not sure how useful it is.
The harmonics emulator is built around a JFET just being driven into saturation. I added a pot to allow bias adjustments. The circuit is a Similar design to "colour" circuit I helped Peterson with... I included a mix knob to blend in the distortion in. This circuit goes from subtle to extreme, but I rarely use it in extreme.
The output has an stepped attenuator with a SP12T switch and resistors. Plus a trim pot just in case I want to match level precisely.
The vu meter uses a dSPIC microprocessor to control a 10 segment LED bar graph. I programmed in A VU mode and DBFS mode. I read the analog output voltage thru a resistive divider into the chips ADC. There is slight delay, as I am only using a 20MHz clock...
IMG_1230.JPG
IMG_1231.JPG
