Re: 807 amplifier



Bret Ludwig wrote:

> The 807 is similar to but NOT the same as the 6L6 or 5881 or KT66
> tube. In particular the screens are not rated for the full plate
> voltage. You can cheat a little in ultra-linear (or triode strapped)
> service but the 807 can be run at well over 600 volts on the plate but
> is limited to around 300 on screens.

Correct but the makers of the 807 allowed the screen voltage to be equal to
the anode voltage
when using UL connection with 43% taps or triode connection where the Ea
was less than +450V.
Screen dissipation was deemed to be OK with UL and triode with the higher
Ea.


> The original Williamson specified the interchangeable use (with
> sockets and plate cap changes) of the 5881, KT66 or 807. But the
> Williamson only gave around 20 watts and was a poor amp in a lot of
> other ways by modern standards.

20 watts was avail when Ea = 450V and triode connection was used,
and Williamson or any other schematic won't give much more.
But with UL and 43% taps, about 30 watts is available.

The Williamson triode amp wa designed to be mainly a class AB1 amp,
with the first 10 watts being pure class A and the remaining watts up to 20
watts in class AB1.
The amplifier characteristics were very good even by today's standards
if you like to listen at normal levels and your speakers are moderately
sensitive,
say 90dB/W/M or more.
Using the UL connection the extra power gives you slightly greater
headroom.

The Williamson was designed so that without NFB the amp had 1% thd, and
an Rout approx 3 ohms at 15 watts. This was possible only with triode
connection.
The NFB lowered the thd to 0.1% and Rout to about 0.2 ohms,
which was exceptional in 1947, and all one needs today.
At normal listening levels it usually sounds quite healthy.

Should more triode power be desired, the use of KT90 in triode
will allow to get about 30+ watts in triode and more in UL.


> I think the early VTLs that used this tube, were the best commercially
> built amps using this tube and should be studied. If I were going to
> build around them I would go with a regulated screen supply pure
> pentode design, or ultralinear using a separate screen winding. Sowter
> makes an Acrosound TO-350 and this may well be the best transformer you
> can buy today anywhere.

Separate screen windings are not needed with 807 in UL;
the only tube I know needing the separate UL winding is the 6146;
or perhaps the EL/PL509.

KT88 / 6550 or as i said the KT90 are far better to use because 50 watts in
UL
is easily possible, but if one already has the 807 and power needs are not
high then it
will do fine.

The Williamson loading was to have 10k a-a for the tube load.

Such a load allowed plenty of class A power and reducing the anode to anode
load
to say 5k a-a like many makers did simply increases thd a lot and
slightly increases power.

A schematic that would suit you for an intergrated power amp, ie, one with
a line level preamp already built in is at

http://www.turneraudio.com.au/htmlwebpgs02/schema5050ulabinteg.htm

Some changes that you need to make when using 807 are

1. Lower B+ to 450V.
2. Use 150k for the grid bias R instead of the 220k at the output tube
grids,
3. Use 6SN7 instead of the 6CG7 shown because they look nicer beside the
807;
The 6CG7 is electronically the same as the 6SN7 and the R values in my
schematic
will suit the 6SN7 exactly with no changes.
4. Determine the "critical damping R&C zobel networks" used in the amp
by careful experiment to get unconditional stability and wide bandwidth.
5. Make sure the OPT has 10k : 6 ohm load matching.

Patrick Turner.




.

Quantcast