On May 23, 5:38*am, John Byrns wrote:
In article ,
*Patrick Turner wrote:
Quad-II used a kind of floating paraphase. Nobody much understands
exactly how it works but fact is that the output of one EF86 is
reduced
by resistance divider and fed to the grid of another EF86 and both
thus have a stage which effectively has 6dB of local positive FB. Both
EF86 have to make up to about 40Vrms for each KT66 grid, ( depending
on the load ).
It's not really that hard to figure out how it works, a lot of the quirkiness is
due to the acrobatics needed to apply negative feedback to the common cathode of
the voltage amplifier and phase splitter. *This was probably a bean counter
thing to eliminate a couple of electrolytic capacitors.
The solution to the problem of low open loop gain suited bean counters
at Quad. But if you make the two EF86 behave as a real LTP with common
cathode R to -400Vdc, or a CCS, and without the mild 6dB PFB boost of
the paraphase inverter principle used, then the differential gain is
halved and if the same amount of NFB was applied globally, you'd need
2.8Vrms for clipping power instead of only 1.4Vrms. One solution is to
use 6BX6 as input tubes with Ia in each at nearly 2mA and then the gm
rises well above what it is in EF86 and by carefully juggling the load
values and biasing R to OP tubes which would better be EL34 with more
gain than KT66, then you get increased open loop gain without the
paraphase connection and better stability etc, and the input pair may
be set up as a true LTP.
I've also used KT88 and KT90 in Quad-II.
Sure anyone may work out ther Quad circuit workings. But I've rarely
ever met that anyone.
Patrick Turner.
--
Regards,
John Byrns
Surf my web pages at, *http://fmamradios.com/