RIAA tweaking

I decided to do some serious tweaking of my RIAA amplifier. Armed with PSPICE this is a piece of cake. First a schematic of my previously tweaked version of my old RIAA amplifier:

 

 

I have removed the RIAA network and created a base for simulation of different passive RIAA network topologies. I looked around the net to see other designs and found a few interesting.

The circuit is intended to drive my bootstrapped cathode follower input in my line amplifier without cable. The output stage can be changed to just about anything with the same or more gain like SRPP, mu-follower or a cathode follower compound. Some adjustment of the low frequency pole may be needed.

Tweaks has been run in PSPICE with the enhanced triode model. I used 12AX7 with mu lowered to 70 to emulate the 5751.

 

This is the original net I used in my amplifier. It’s designed by the late Jan Lodström many years ago.

 

It’s not far off the RIAA curve, around +0.3dB at 20 Hz worst case. Simple values, the 30nF is just 2 15nF in parallel.

 

It has two advantages, there is no need for a grid leak on the second tube and C3 with its large value slugs the miller capacitance of it.

 

The disadvantage is the missing 50 kHz zero.

Gain 48dB @ 1 kHz.

I just love this topology.

 

This is my final tweak of the above net. It’s now only +0.2dB off at 20 Hz worst case.

 

I also added the 50 kHz zero. Still no unusual values, all resistors are within E24 and the 30nF is still 2 15nF in parallel.

 

The miller capacitance plays only a minor role, the lift at 20 kHz is there.

 

If you don’t want the 50 kHz zero, remove R8 and ground C6.

Gain 50dB @ 1 kHz.

This is the one I’m going to implement.

 

This is Tweety by Maarten (link).

 

It’s his tweaked version for his amplifier but it works very well in my amplifier above. It’s no more than 0.2dB off anywhere.

 

The disadvantage is the need for a grid leak, a large one which could cause troubles if the tube is a little gassy. It also has an odd capacitor value.

 

I have added a 50kHz zero to it, R12.

 

Gain 53dB @ 1kHz in my circuit.

 

This is Loekie which I found on Maarten’s site and tweaked by him. It’s also very good.

 

This one also needs the grid leak resistor and has a few odd resistor values.

 

Gain 53dB @ 1 kHz in my circuit.

 

Finally the net of Scratch by Triode Dick (link). This one is also good but needs some tweaking at 20 Hz in my circuit. The topology is the same as Loekie but without the 50 kHz zero. Adding the zero transforms the circuit into Loekie above.

 

Gain 52dB @ 1 kHz in my circuit.