moogus wrote on Sat, 24 July 2004 15:20 |
I cant help myself from responding here.
We just finished 2 mastering pultecs the other week, that are pretty darn close to what youre describing.
snipped
Using the JLM instrument opamp from the mastering console to give fully balanced i/o and provide the makeup gain, which runs on +/-24V, has +36 headroom, bandwidth from DC to above 10meg and plenty of zeros in the THD%.
M
|
Hi Moogus,
Is this with the JLM99V or as your instrument opamp description
suggests, a fully balanced instrumentation amplifier Bal IP ->
SE OP, or just the front end of an instrumentation amp ie;
CM gain = x1 bal in bal out, or lastly, a fully balanced VG
(current) IP such as Nelson Pass's "Supersymmetry"?
There are a few issues which may be relevant WRT BK's and DC's
comments of BW limiting but they don't aply to all topologies
and designs, lets make sure we are all talking the same
language first.
Some of the topologies are in effect current feedback, ie
feedback to the IP emitters and as such you need to be careful
using FB BW compensation (as BK suggested). However there are
ways around all this.
You mention BW of 10meg, just for the record, with devices these
days having Ft's in the 100's of megs that is pretty common.
Joe seems to favour very simple topologies (which I like too)
and high voltage rails (which I also like), as such the high
BW will certainly be a given.
There were valid questions of the necessity of such high rails
(DC). I think one big advantage is that simple circuits can
be more linear without the need for cascoding to linearise
capacitances and early effect etc. The headroom issue as DC
suggests is a red herring IMO the benefits are elsewhere.
Check some BJT data sheets and look where the miller C gets
very linear WRT voltage across the device. You can really
measure this especially on simple low FB circuits and at higher
frequencies.
Another related issue is if the given stage is fed from a
higher impedance such as a passive vol control. In this case
there will be increased distortion due to IP device modulated
miller C and as stated before it will manifest itself at higher
freq.
Cheers,
Terry