Thread: 6CA7 in AB2.
View Single Post
  #10   Report Post  
Patrick Turner
 
Posts: n/a
Default



Fred Nachbaur wrote:

Patrick Turner wrote:

[...]

Pentode operation allows for up to 800v for the plate supply, and say 400v
on the screens, and a 12k a-a RL, and grid current never occurs.
But if we reduced the screen volts to say 300v, then
we would maybe run out of grid drive headroom.

UL and trioded tubes cannot be run at such high voltages, because the grid 1
would have to be biased at impossibly low -ve bias voltages.
[...]


Well, I wouldn't say it's impossibly low. My favourite hi-fi amp uses
6L6GC's at about 620 volts B+ in UL. Grid bias for a reasonable plate +
screen dissipation (cathode current 35 mA) is about -78 volts. I don't
think that it would be very much different for EL34 or 6CA7, but haven't
tried it. (If anyone would care to donate a set, I'd be happy to try
them out and report back!)


True, some folks do use really high voltages for UL or triode.
Its sudden dearth territory, and many older tubes just can't be biased
down to be stable, 6CM5 is a classic.
I don't like such high voltages for trioded or UL amps,
because it means the idle current has to be low, to get the tubes
to be biased a LOT below the rated max.
This means the tubes are not working in the most linear region
for the first few watts, which are the most important, and the amp mainly class B.

+500v is as high as I like to go.

Acoustical goes really well at all voltages, and is a better way to
do UL, and the screens are given a reprieve from high screen currents,
since the screen voltage can be a lot lower than the anode supply.



This means, though, that the driver needs lots of headroom; about 160
volts peak-to-peak.


That's only around 55 vrms, per each input, and that's easy to do,
and a heck of a lot easier than a McIntosh, where the drive to each output tube is

maybe 140 vrms.
Same for EAR509.


The usual concertinas are therefore rather
impractical. But what a great application for a long-tail pair with
pentode (or transistor) CCS! Lots of output swing -- and gain as a bonus.


Well exactly.



The net result is lots of power, good clean sound even at full power due
to the UL topology, using relatively common-place tubes. I've found AB2
to be a real nuisance for hi-fi use, but it's great for guitar amps
because it give a more graduated distortion characteristic, passing from
class A to class AB1 to class AB2 to full clipping. It's absolutely
*wonderful* for that, the louder you play the "harder" you play, without
suddenly crunching into the ceiling like you do with SS amps.


One might have to have some limiting R between a CF driver and the output grids,
say 1k, maybe 2k, to give a nice squishy overdrive, ie, a natural compressor
effect.
Also, grid current needs to be limited somewhat in amps
subjected to continual overdrive all night.





But for hi-fi, unless you're a glutton for punishment and want a real
challenge to get a low-impedance driver, and then maybe still have to
try to iron out the added distortion by using NFB tricks, I wouldn't
recommend AB2 using "conventional" valves like the 807, 6L6, EL34, etc.
I think you'd be better off with a lower AB1, even approaching B1, than
to cope with the difficulty of trying to stay clean whilst drawing grid
current.


I would still have the same highish value load I normally use,
so the first 25 watts are still the same low thd as the AB1 amp.
If the load is halved, the power is lower, due to grid voltage excursions
needing to exceed the bias voltage, to get the preferable extra load voltage
swing.
The extra thd above 25 watts isn't going to be huge,
and is for the occasional drumbeats.



I too am interested in AB2 for hi-fi, but would use a pair of
high-voltage TX triodes such as 809 or 811A, in a "darlington" kind of
configuration (direct-coupled cathode followers on the control grids).
From reports I've seen, and from a very crude mock-up I did years ago,
this has the potential of sounding very nice, and have lots of available
power.


!/2 a 6SN7 is a fine CF driver for 6L6 in AB2, and I have a Brimar circuit
in an old tube manual of theirs with this set up, using 807, with +600v
for B+, and +300v for EG2, and it makes 80 watts, and that's what folks used for
PA, back in 1953, complete with hardly any FB, and 10% thd max.
At 10 watts, they weren't too bad.
The 807 meant you didn't get arcing on the tube socket, and the Eg2
voltage meant the screens didn't fry every time some galoot ran a shorted speaker
lead,
and turned up the volume.

Patrick Turner.


Cheers,
Fred
--
+--------------------------------------------+
| Music: http://www3.telus.net/dogstarmusic/ |
| Projects, Vacuum Tubes & other stuff: |
| http://www.dogstar.dantimax.dk |
+--------------------------------------------+