The genesis of this is a bit complicated... to make it short, I need a balanced line receiver with high impedance inputs (high like 1meg, not the 25k or so of an INA134) and able to directly drive all kinds of heaphones. A pair of ds1802 (50k digital pot) will be at its inputs.
On one hand, JCX suggested here a way to proceed (other than with opamps). Jam suggested to make it fully complementary.
On the other hand, I've got this amp available. It would easily be modded into the attached schematic... a few transistors to add, some resistors value to change.
It seems to work decently enough in simulations but I'd love some comments before starting to hack the existing preamp.
On one hand, JCX suggested here a way to proceed (other than with opamps). Jam suggested to make it fully complementary.
On the other hand, I've got this amp available. It would easily be modded into the attached schematic... a few transistors to add, some resistors value to change.
It seems to work decently enough in simulations but I'd love some comments before starting to hack the existing preamp.
Attachments
Your idea of fully complementary zero negative feedback is the right way to do it.... Personally, I'll recommend a folded cascode and using a single set of high current high hfe output transistor like the Zetex or Japanese parts, those small BC547/557 hfe drops just to quickly when delivering current.
My suggestion is attached, it should world class....
You could loose the cascode on the input fets, but they're good for the linearity. Adjust fet current with source resistors, output current with their emitter resistors (small signal diodes and high current bipolar result in about 0.1V over each), with bias+ and bias- shorted then current limit will be about 6 times bias. For lowest distortion (<0.01% at high out level and high frequencies) ensure that p-p fet signal current is lower than 25% of bias, adjusting R12,R13 & R6,R20 as needed. And maybe set the output transistors for high bias class A operation....
Soren
My suggestion is attached, it should world class....
You could loose the cascode on the input fets, but they're good for the linearity. Adjust fet current with source resistors, output current with their emitter resistors (small signal diodes and high current bipolar result in about 0.1V over each), with bias+ and bias- shorted then current limit will be about 6 times bias. For lowest distortion (<0.01% at high out level and high frequencies) ensure that p-p fet signal current is lower than 25% of bias, adjusting R12,R13 & R6,R20 as needed. And maybe set the output transistors for high bias class A operation....
Soren
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.