i've been working on this for a while, i wanted an input
stage that performs well both for unbalanced and balanced inputs
without feedback, fully dc coupled etc. I came up with this which
sounds really good and has a thd of <.01%
http://gilmore.chem.northwestern.edu/ubaltobald.pdf
buildable with standardly available parts from mouser.
performs even better with jfets as the inputs.
comments?
stage that performs well both for unbalanced and balanced inputs
without feedback, fully dc coupled etc. I came up with this which
sounds really good and has a thd of <.01%
http://gilmore.chem.northwestern.edu/ubaltobald.pdf
buildable with standardly available parts from mouser.
performs even better with jfets as the inputs.
comments?
without spending time analysing the detailed circuit I can't see how you are controlling the overall gain.
The input diff pairs are driving current mirrors - good. The current mirror drives into the common base cascade, which between the two pairs you develop the outputs in push pull.
The diff pair gain should be quite high given its load. How is this set? Similarly I note the output would normally have very high gain, those 430 odd ohm resistors on the output would "pull this down" but I suspect that in the absence of overall feedback the gain and more general performance would be quite load dependant.
Have you tried the heresy of upping the gain and adding overall feedback?
The input diff pairs are driving current mirrors - good. The current mirror drives into the common base cascade, which between the two pairs you develop the outputs in push pull.
The diff pair gain should be quite high given its load. How is this set? Similarly I note the output would normally have very high gain, those 430 odd ohm resistors on the output would "pull this down" but I suspect that in the absence of overall feedback the gain and more general performance would be quite load dependant.
Have you tried the heresy of upping the gain and adding overall feedback?
thermal stability seems rock solid with the output stage servo's.
the gain is set by the 475 ohm output resistors. current mirror into
a resistor is fixed gain.
Handles about 5 volts of common mode. That is what the 1k tail
resistors are for.
adding feedback would just lower the input resistance a bunch
so not sure that is a good idea.
the gain is set by the 475 ohm output resistors. current mirror into
a resistor is fixed gain.
Handles about 5 volts of common mode. That is what the 1k tail
resistors are for.
adding feedback would just lower the input resistance a bunch
so not sure that is a good idea.
Last edited:
Kevin,
Reminds me a bit of the enhanced cascode gain circuit in fig.9 of the below linked Malcolm Hawksford paper. You've added a nice little local feedback-loop via the current-mirror implementation. I like it.
http://www.essex.ac.uk/csee/research/audio_lab/malcolmspubdocs/J10 Enhanced cascode.pdf
For me, the most interesting and, perhaps, challenging area of solid-state gain stage design involves low-ish distortion global open-loop circuits. Especially, those which do not greatly depend mostly on very precise device matching to achieve low distortion.
Reminds me a bit of the enhanced cascode gain circuit in fig.9 of the below linked Malcolm Hawksford paper. You've added a nice little local feedback-loop via the current-mirror implementation. I like it.
http://www.essex.ac.uk/csee/research/audio_lab/malcolmspubdocs/J10 Enhanced cascode.pdf
For me, the most interesting and, perhaps, challenging area of solid-state gain stage design involves low-ish distortion global open-loop circuits. Especially, those which do not greatly depend mostly on very precise device matching to achieve low distortion.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.